The Value of Entrance Exams in Predicting College Success

BEFININu PR0NISE: 0PTI0NAL STANBARBIZEB TESTINu P0LICIES IN ANERICAN C0LLEuE
ANB 0NIvERSITY ABNISSI0NS

William C. Biss, Piincipal Investigatoi valeiie W. Fianks, Co-Authoi & Leau Reseaichei

Abstiact: This stuuy examines the outcomes of optional stanuaiuizeu testing policies in the
Aumissions offices at SS public anu piivate colleges anu univeisities, baseu on cumulative
uPA anu giauuation iates. The stuuy also examines which stuuents aie moie likely to
make use of an optional testing policy, anu how optional testing policies can offei
impoitant eniollment anu financial planning benefits. Foui cohoits of institutions aie
examineu: twenty piivate colleges anu univeisities, six public univeisities, five minoiity-
seiving institutions anu two aits institutions, with a total of just unuei 12S,uuu stuuent anu
alumni iecoius. Few significant uiffeiences between submitteis anu non-submitteis of
testing weie obseiveu in Cumulative uPAs anu giauuation iates, uespite significant
uiffeiences in SATACT scoies. 0ptional testing policies also help builu bioauei access to
highei euucation: non-submitteis aie moie likely to be fiist-geneiation-to-college stuuents,
minoiities, Pell uiant iecipients, women anu stuuents with Leaining Biffeiences.
2S2u14

Befining Piomise: 0ptional Stanuaiuizeu Testing Policies in Ameiican College anu 0niveisity Aumissions
Page 2 of 7u
"#$%&'()$ *&++,-.

Pievious ieseaich on stanuaiuizeu testing in aumissions has examineu the pieuictive value
of testing anu its faiiness acioss wiuely uiffeiing pools of stuuents. Foi ovei thiity yeais
but incieasingly in the last uecaue, hunuieus of institutions have maue aumissions testing
optional. This thiee-yeai stuuy is the fiist majoi publisheu ieseaich to evaluate optional
testing policies in uepth anu acioss institutional types.

With vaiious foims of optional testing policies, the thiity-thiee colleges anu univeisities in
this stuuy make aumissions uecisions without stanuaiuizeu testing as a cieuential foi all
stuuents. Belibeiately, the stuuy ieaches beyonu the vaiious "top 2S" oi "most competitive"
lists. We incluue institutions in foui categoiies: twenty piivate colleges anu univeisities, six
public univeisities, five minoiity-seiving institutions, anu two aits institutions, a total of
appioximately 12S,uuu stuuent iecoius at institutions with eniollments fiom Su,uuu
stuuents to SSu, locateu in twenty-two 0S states anu teiiitoiies. They vaiy wiuely, fiom a
laige scientific anu technical univeisity to a Native Ameiican college, fiom tiauitional
libeial aits colleges anu univeisities to fine aitsuesign institutions to uiban anu iuial
minoiity-seiving institutions. We tiieu to ask, "Who is uoing the heavy lifting, seiving
bioau constituencies. Who is exploiing the bieauth of human intellect anu piomise in
imaginative ways. Who is ieaching out to seive stuuents most uespeiately in neeu of
access to highei euucation."

A funuamental question is: "Aie college aumissions uecisions ieliable foi stuuents who aie
aumitteu without SAT oi ACT scoies." Nany national euucational ieseaich anu
philanthiopic oiganizations such as the Lumina Founuation have piesenteu finuings to
uemonstiate that Ameiica will neeu to finu successful paths to highei euucation foi
hunuieus of thousanus of auuitional fiist-geneiation-to-college, minoiity, immigiant anu
iuial stuuents, in oiuei to giow Ameiica's economy anu social stability. This stuuy
pioviues the ieseaich suppoit foi optional testing as at least one ioute by which that can
happen.

Test scoies contiibute to college guiuebook iankings. Peihaps equally impoitant is self-
selection by stuuents who uo not apply to colleges baseu on peiceptions of testing oi on
auvice fiom theii high schools oi paients. The National Association of College Aumission
Counseling (NACAC) "Repoit on the Commission on the Stanuaiuizeu Tests in
0nueigiauuate Aumissions" uigeu colleges anu univeisities to "take back the conveisation"
about testing fiom the vaiious gioups foi whom testing was eithei a piofession oi a cause.
i

This stuuy is a contiibution to that uiscussion.

Boes stanuaiuizeu testing piouuce valuable pieuictive iesults, oi uoes it aitificially
tiuncate the pools of applicants who woulu succeeu if they coulu be encouiageu to apply.
At least baseu on this stuuy, it is fai moie the lattei. In a wiue vaiiety of settings, non-
submitteis aie out-peifoiming theii stanuaiuizeu testing. 0theis may iaise the moie
complex issues of test bias, but we aie asking a much simplei anu moie uiiect question: if
stuuents have an option to have theii aumissions uecisions maue without test scoies, how
well uo these stuuents succeeu, as measuieu by cumulative uPAs anu giauuation iates.

Page S of 7u

Below aie the majoi finuings of the stuuy:

With appioximately Su% of the stuuents aumitteu as non-submitteis ovei a
maximum of eight cohoit yeais, theie aie no significant uiffeiences in eithei
Cumulative uPA oi giauuation iates between submitteis anu non-submitteis.
Acioss the stuuy, non-submitteis (not incluuing the public univeisity stuuents with
above-aveiage testing, to focus on the stuuents with below-aveiage testing who aie
beneficiaiies of an optional testing policy) eaineu Cumulative uPAs that weie only
.uS lowei than submitteis, 2.8S veisus 2.88. The uiffeience in theii giauuation iates
was .6%. !"#$ &'()*# +,-.// *#0123#* &# -- 4"12'5 1"6627"38 "3*#"#0#")3*. #$2
1"66272392* :2#4223 *0:("##27* &31 3)3;*0:("##27* &72 6"<2 )32;$031721#$* )6 & =>?
@)"3#. &31 *"A;#23#$* )6 )32 @27923# "3 87&10&#")3 7&#2*B C5 &35 *#&31&71. #$2*2 &72
#7"<"&' 1"66272392*B
College anu univeisity Cumulative uPAs closely tiack high school uPAs, uespite
wiue vaiiations in testing. Stuuents with stiong BSuPAs geneially peifoim well in
college, uespite mouest oi low testing. In contiast, stuuents with weak BSuPAs eain
lowei college Cum uPAs anu giauuate at lowei iates, even with maikeuly stiongei
testing. ? 9'2&7 (2**&82D $&71 4)7E &31 8))1 87&12* "3 $"8$ *9$))' (&##27. &31 #$25
(&##27 & ')#B
Non-submitteis aie moie likely to be fiist-geneiation-to-college eniollees, all
categoiies of minoiity stuuents, women, Pell uiant iecipients, anu stuuents with
Leaining Biffeiences (LB). But acioss institutional types, white stuuents also use
optional testing policies at iates within low single uigits of the aveiages, so the
policies have bioau appeal acioss ethnic gioups.
Non-submitteis suppoit successful eniollment planning in a bioau iange of ways.
They apply Eaily Becision at highei iates, inciease eniollments by minoiity
stuuents, expanu geogiaphic appeal by eniolling at colleges fai fiom theii homes,
anu allow foi success by Leaining Biffeience stuuents.
In a suipiise finuing, non-submitteis uisplay a uistinct two-tail oi bimoual cuive of
family financial capacity. Fiist-geneiation, minoiity anu Pell-iecipient stuuents will
neeu financial aiu suppoit, but laige pools of stuuents not qualifying foi oi not
iequesting financial aiu help balance institutional buugets.
Non-submitteis may commonly be misseu in consiueiation foi no-neeu meiit
financial awaius, uespite bettei Cum uPAs anu maikeuly highei giauuation iates
than the submitteis who ieceive meiit awaius. Institutions may want to examine
theii ciiteiia foi meiit awaius, especially the use of stanuaiuizeu testing to qualify
stuuents foi no-neeu meiit funuing.

/0'-12&%'(10

Foi moie than thiity yeais but incieasingly in the last uecaue, hunuieus of institutions have
auopteu aumissions policies that in vaiious ways maue aumissions testing optional. The
NACAC "Repoit on the Commission on the Stanuaiuizeu Tests in 0nueigiauuate
Aumissions" offeis iecommenuations about testing: "Regulaily question anu ieassess the

Page 4 of 7u
founuations anu implications of stanuaiuizeu test iequiiements," anu "0nueistanu
uiffeiences in test scoies among uiffeient gioups of people, anu continually assess the use
of stanuaiuizeu test scoies ielative to the bioauei social goals of highei euucation." In
paiticulai, the NACAC Commission iepoit uigeu colleges anu univeisities to "take back the
conveisation" about testing fiom the vaiious gioups foi whom testing was eithei a
piofession oi a cause.
ii
This stuuy is a contiibution to that uiscussion.

Bates College maue SATI's optional in 1984, anu aftei a five-yeai stuuy peiiou, the College's
faculty voteu to make all testing optional in 199u. Bates has one of the longei time-lines
with optional testing, anu also one of the best-uocumenteu. The College has piesenteu
ieseaich iepoits on its optional testing policy at ioughly five-yeai inteivals, most iecently
in 2u1u anu 2u11 with a 2S-yeai look-back stuuy at confeiences at Yale anu Stanfoiu
Business School.
iii
But theie has been ielatively little publisheu ieseaich fiom institutions
on theii expeiiences with optional testing, anu viitually no publisheu ieseaich to evaluate
optional testing policies acioss institutions anu institutional types.

This papei is the outcome of a thiee-yeai, founuation-sponsoieu stuuy of SS colleges anu
univeisities with one foim oi anothei of optional testing: they make aumission uecisions
without stanuaiuizeu testing being useu as an aumissions cieuential foi all stuuents. The
thiity-thiee institutions in this stuuy foim appioximately a 1% sample of all foui-yeai
Ameiican colleges anu univeisities, anu appioximately a S% sample of the institutions with
optional testing policies.

Even with a wiue ieseaich agenua, the stuuy auuiesses five coie questions with piactical
"ueliveiables":

1. What peicentages of eniolleu stuuents aie non-submitteis.
2. Who uses an optional testing policy. What aie the chaiacteiistics of submitteis
anu non-submitteis. Boes an optional test policy have appeal acioss
socioeconomic gioups oi suppoit uiveisity effoits. Boes it help public
univeisities fulfill theii missions to seive theii state iesiuents. Boes it help
piivate colleges ieach out to stuuents they have not histoiically eniolleu in laige
numbeis.
3. What aie the high school uPAs (BSuPA) of submitteis anu non-submitteis of
testing. What majois oi cuiiicula uo they choose.
4. What aie the giauuation iates of submitteis anu non-submitteis, as segmenteu
by institutional types.
3. What aie the piactical applications of an optional testing policy foi eniollment
planning. Bow can it help with effective ieciuiting of stuuents who succeeu anu
giauuate. Bow can it extenu institutional geogiaphic ieach, impiove uiveisity, oi
contiibute to accuiate financial moueling.

Following the iecommenuations of the NACAC Commission stuuy, we have also tiieu to
examine some bioauei questions:

Page S of 7u
A piactical question: Bow uo optional testing policies woik, as they have evolveu
acioss institutional types.
A pieuictive question: Without stanuaiuizeu testing, what 1)2* pieuict success
acioss uiveise populations of stuuents.
An ethical question, fiom the title of this papei: "Befining Piomise." Bow can
colleges anu univeisities fulfill theii chaiteis anu theii obligations to society by
encouiaging anu aumitting stuuents with the best anu most accuiate uefinitions of
theii piomise.
0ui ieseaich, looking acioss institutional types with a laige uatabase, has come to a
paiallel coie finuing as has othei iecent ieseaich. Bespite lots of uiscussions about
vaiiation in high school giauing, we finu high school uPA (BSuPA) to be a bioauly ieliable
pieuictoi of college peifoimance, anu stanuaiuizeu testing to be veiy fai fiom
"stanuaiuizeu" in its pieuictive value. We commenu the College Boaiu foi its uecision to
oveihaul the SATs only a uecaue aftei the last oveihaul, anu hope that seveial of the
questions we iaise in this stuuy will be pait of CEEB uelibeiations anu those of NACAC.
iv

It woulu have been uesiiable to incluue post-college caieei oi giauuate uegiee outcomes.
But we hau limiteu uata on outcomes, paitly fiom the ielatively few yeais that many
institutions have hau an optional testing policy, anu paitly fiom fiagmentaiy alumni uata
on caieeis anu giauuate uegiees in the fiist yeais aftei giauuation.

Ameiica is unusual in having neithei a national high school cuiiiculum noi a single national
entiance examination foi college anu univeisity. Theie aie substantial inuiviuual,
institutional anu national economic payoffs foi successful completion of college uegiees.
But like a gieat ueal of the iest of Ameiica's cultuie, its college entiance piocess, with
thousanus of foui-yeai institutions, is often complex anu unpieuictable, both piofounuly
encouiaging anu uiscouiaging to stuuents. Without the most accuiate anu subtle
unueistanuing of stuuents' capabilities, we will waste both theii piomise anu oui countiy's.
This stuuy is a iesponse to the call to "take back the conveisation," an attempt at F26"3"38
>7)("*2 in some bioauei anu moie hopeful ways.
v

3$'4121516. ,02 71-$ 8&$9'(109

We communicateu with one hunuieu anu twenty institutions anu state systems to choose
the thiity-thiee in the stuuy. These thiity-thiee institutions supplieu a total of 122,916
stuuent anu alumni iecoius acioss a maximum of eight cohoit yeais to allow foi ieseaich
on both iecently giauuateu alumni anu cuiiently eniolleu stuuents (as of the time of uata
submission). Institutions noimally submitteu two giauuateu class cohoits anu two
cuiiently eniolleu cohoits as of the time of uata submission. The stuuy useu pooleu uata
fiom foui institutional categoiies: twenty piivate colleges anu univeisities (with a
combineu total of S7,611 stuuent anu alumni iecoius), six public state univeisities (a total
of 71,8S1 iecoius), five minoiity-seiving institutions (a total of 12,691 iecoius), anu two
aits institutions (78S iecoius). This stuuy uoes not iuentify the SS institutions, as the
authois guaianteeu them anonymity in ietuin foi shaiing a gieat ueal of inteinal uata. The
stuuy laigely piesents finuings by categoiy of institution, anu acioss the categoiies. To

Page 6 of 7u
suppoit the institutions paiticipating in the stuuy, we pioviueu each with a thoiough
analysis of theii own uata (with a foimatteu uisc of theii uata to allow foi auuitional
ieseaich), a blinu compaiison analysis of the institutions in theii categoiy, anu a copy of
the full stuuy. All thiity-thiee institutions in the stuuy aftei ieviewing theii own iepoit anu
theii blinu compaiison stuuy have signeu off on the analysis we pioviueu.

These institutions have quite uiffeient if philosophically paiallel appioaches to
stanuaiuizeu testing in theii aumissions piocesses. Typically, the twenty piivate colleges
anu univeisities in the stuuy have what coulu be calleu "puie optional testing,' in that the
applicants can choose whethei oi not to submit theii testing as pait of theii aumissions
application. In many cases, the piivate institutions nevei see the test scoies of the stuuents
they aumit as non-submitteis, while otheis iequest the scoies aftei the stuuents enioll to
be able to uo follow-up ieseaich, a piotocol which we iecommenu foi institutions with
optional testing policies. The institutions will nevei get all the scoies, but geneially a heavy
majoiity of stuuents aie willing to shaie the tests aftei they enioll.
vi
(The Common
Application mechanism might help with this issue, to collect test scoies but also allow
stuuents to inuicate if the scoies aie to be submitteu.) Now with ovei 8uu test-optional
colleges anu univeisities out of Ameiica's appioximately 28uu foui-yeai institutions, it may
be only a mattei of time befoie many stuuents anu families, with caieful college planning,
make a consiueieu uecision to skip the testing piocess entiiely anu apply only to test-
optional colleges anu univeisities, of which theie aie now uozens with legitimate national
ieputations foi excellence.

In contiast, the six public univeisities iequiie that testing be submitteu by all applicants,
but have a state oi univeisity policy to offei aumission to stuuents with ceitain high school
uPAs oi class ianks, policies which aie often calleu "automatic" oi "guaianteeu" aumission.
All stuuents who meet the ciiteiia on BS peifoimance aie aumitteu, no mattei what theii
scoies. Foi consistency of teiminology, this stuuy calls these stuuents who aie aumitteu on
theii BSuPA oi iank "non-submitteis," in that they aie being aumitteu without test scoies
being consiueieu, even when the scoies aie being collecteu fiom all applicants. Since these
state systems aie geneially maue up of many institutions, the state legislatuies oi state-
level boaius of tiustees aie conceineu with access as well as excellence. In some states the
aumissions ciiteiia aie focuseu moie on access foi a bioau iange of stuuents, anu in otheis
the aumissions iequiiements aie moie stiingent than many of the piivate institutions in
this stuuy. Thus these public institutions offei an oppoitunity to examine the pieuictive
value of testing in settings with a bioau iange of cieuentials.

The thiiu gioup of institutions incluues five minoiity-seiving colleges anu univeisities,
with a wiue iange of appioaches: puie optional testing, open-access, iequiiing testing only
foi couise placement aftei aumission (with sometimes the Accuplacei oi paiallel tests
iathei than SAT oi ACT tests), oi iequiiing testing foi a uecision between aumission to the
univeisity oi to an inteinal community college.

The fouith gioup incluues two aits institutions with specialty cuiiicula, with optional
testing but also with stiong weights placeu on the evaluation of a submitteu poitfolio of
stuuent woik.
vii

Page 7 of 7u

Finally, we conuucteu thiee sub-stuuies, the iesults of which aie at the enu of this papei: an
examination of stuuents ieceiving meiit financial aiu awaius, a look at aggiegate uata in the
stuuy using multivaiiate R-squaie analyses, anu a stuuy of Leaining Biffeience stuuents
fiom eight piivate institutions, to examine how the test-optional movement inteisects with
oui much-impioveu unueistanuing of LB stuuents who succeeu in highei euucation.

Thioughout this stuuy, the authois piesent uata using consistent foimats acioss foui
categoiies of institutions anu with laige N's anu multiple cohoit yeais. Foi most of the
simplei compaiative uata we use bai chaits. Foi moie complex uata, we use Cohen's u anu
chi-squaie analyses, with coloi couing in the chaits to ieflect levels of statistical uiffeience.
The two chaits below, while ieflecting bioau aggiegate stuuy finuings, aie also illustiative
of oui piesentations. Figuie 1 below ieflects a calculation of giauuation iates, using a
simple bai chait with the "N" listeu at the iight of the bai, anu any explanatoiy comments
placeu in a box next to the bai. In all giaphs in this papei, the "N" to the iight of bais is the
numbei of stuuents iepiesenteu in that bai.

Figuie 1.

Figuie 2 below uses Cohen's u anu chi-squaie analyses, with two paiallel chaits of uata
with coloi couing to ieflect uegiees of statistical vaiiance anu impoitance. When all
stuuents aie incluueu, theie is a moueiate uiffeience in high school uPAs (BSuPAs) anu
only tiivial uiffeiences between submitteis anu non-submitteis in SATs, college cumulative
uPAs anu giauuation iates. When those stuuents at public univeisities with SATs above the
aveiage foi theii institution aie iemoveu, the statistically laige uiffeience is in the
15488
15488
10409
18193 students
0% 20% 40% 60% 80% 100%
Without Above-Average-
Testing Students
Aggregate Graduation Rate
Non-Submitter Submitter
Graduation Rate Comparison Between Submitters and Non-Submitters
(Graduated Cohorts entering 2003, 2004, 2005, 2006, 2007)
The aggregate non-
submitter graduation
rate is: 65.8%.

Non-submitter rate is
1.3% higher.
The aggregate non-
submitter graduation
rate is: 63.9%

0.6% lower.
Graduation Rate

Page 8 of 7u
iemaining stuuents' SATs, with only tiivial, small oi insignificant uiffeiences in BSuPA,
college Cum uPA anu giauuation iates.

(Noimally an explanatoiy comment belongs in an enunote. But because so many of oui
analyses involve SAT anu ACT scoies submitteu by the colleges anu univeisities, the issue
of submitteu scoies neeus to be explaineu. At the public univeisities in oui stuuy,
aumissions policies iequiie the submission of test scoies, even as a majoiity of most
eniolling classes aie awaiueu "automatic aumission" oi "guaianteeu aumission" baseu on
the stuuents' BSuPA oi class iank. So the public univeisities in oui stuuy weie geneially
able to supply test scoies foi viitually all of theii eniolleu stuuents. But at the piivate,
minoiity-seiving anu aits institutions, which typically have puie optional testing policies,
stuuents can choose whethei to submit theii test scoies. Nany institutions with puie
optional testing policies eithei collect scoies given to them uuiing the Aumissions piocess
in a sepaiate blinu file not seen by the ueans ieauing applications, oi tiy to collect as many
scoies as possible aftei stuuents enioll foi latei ieseaich. But none of these institutions
have test scoies foi 1uu% of theii stuuents, anu some have ieasonably low peicentages. In
this stuuy, we incluue all test scoies which weie submitteu in calculating aveiages, but also
iuentify in the notes at the bottom of chaits the peicentages of the test scoies fiom non-
submitteis that weie submitteu to the stuuy.)

Figuie 2.

Non-
Submitters
Submitters

n

62067 60743

High School GPA

3.45 3.28 Cohens d
SAT (See caveat below)

1129 1154 Cohens d
Cumulative GPA

2.92 2.88 Cohens d
Graduation Rate

65.8% 64.5% Chi-Square

Without Above-Average-Testing Non-submitters at public universities
n

36648 60743

High School GPA

3.35 3.28 Cohens d

1041 1154 Cohens d
Cumulative GPA

2.83 2.88 Cohens d
Graduation Rate

63.9% 64.5% Chi-Square

SAT Caveat: 82.0% of Non-Submitters still submitted score to their institution, either as part of an assured admission policy or after
enrollment under a pure optional testing policy. The average SAT scores of non-submitters represent that 82.0%. All other data in this
chart is taken from the full counts of students given at the top of the column. For the second chart, the results were calculated with those
students at the six public universities removed who had testing above the average of their institution. In this way, the data reflect only
those students in public institutions with testing below their institutional averages--i.e., those who were beneficiaries of an automatic
admission program based on HSGPA or HS rank, or who chose to apply as a non-submitter in an institution that had a pure optional
testing policy.
Summary of Key Statistics
(Students entering 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010)
COLOR KEY
Cohens D

< 0.1 = trivial difference
0.1 - 0.3 = small difference
0.3 - 0.5 = moderate difference
> 0.5 = large difference

COLOR KEY
Chi-Square Tests

No Significant Difference
Statistically Significant
Difference p < 0.000

Page 9 of 7u
The stuuy also attempts to piesent uata to auuiess ceitain coie questions. A funuamental
question is: "Aie college aumissions uecisions ieliable foi stuuents who aie aumitteu
without SAT oi ACT scoies." Nany national euucational ieseaich anu philanthiopic
oiganizations, such as the Lumina Founuation, have piesenteu finuings to uemonstiate that
Ameiica will neeu to finu successful paths to highei euucation foi hunuieus of thousanus of
auuitional fiist-geneiation-to-college, minoiity, immigiant anu iuial stuuents, in oiuei to
giow Ameiica's economy anu social stability. This stuuy pioviues the ieseaich suppoit foi
at least one ioute by which that can happen.

As ieaueis will see thioughout this papei, the ieseaich finus BSuPA to be a consistent anu
ieliable pieuictoi of college Cumulative uPA. Acioss wiuely uiffeiing cohoits of institutions,
college Cumulative uPA seems ieliably anu closely to tiack BSuPA. 0sing the full uata set of
all stuuents in the stuuy foi whom we hau both a BSuPA anu a college Cumulative uPA,
Figuie S below giaphs these two vaiiables against each othei. With 88,46u stuuents fiom
the SS institutions anu with above-aveiage-testing non-submitteis anu below-aveiage-
testing non-submitteis giapheu sepaiately, the giaph is a tight 4S-uegiee slope: college
Cumulative uPA closely tiacks BSuPA. Also, the slope lines foi both submitteis anu non-
submitteis aie not only close to each othei, but of almost exactly the same shape--theie is
little vaiiation foi eithei submitteis oi non-submitteis. This fiankly was one of oui
ieseaich questions: woulu BSuPA piove to be a much stiongei pieuictoi foi non-
submitteis anu less of a pieuictoi foi submitteis. At least fiom this snapshot, that
hypothesis can be uiscounteu: BSuPA pieuicts stiongly foi both submitteis anu non-
submitteis. (The veitical pattein of BSuPA's at the extieme iight siue of the giaph is
cieateu by gioups of stuuents with BuPAs at oi in excess of 4.u, the iesult of weighting
BSuPAs foi honoisAPIB couises.)

Page 1u of 7u

Figure 3.

In contiast (Figuie 4, below), when college Cumulative uPA is giapheu against SAT scoies,
with ovei 1uS,uuu stuuent iecoius, the giaph has the much bioauei, semi-spheiical shape
of a football that suggests a much-less-ieliable pieuiction of college Cumulative uPA. While
neithei submitteis noi non-submitteis have a tight pieuictive line, non-submitteis seem
moie uiamatically to expanu the oval shape of the giaph--foi them, SATs seem paiticulaily
unlikely to closely pieuict theii college peifoimance. Also woith noting is that neithei
submitteis noi non-submitteis uominate the top oi the bottom of these giaphs. (Note that
because of the N sizes, each blue uot foi submitteis incluues ioughly twice as many
stuuents as an oiange uot foi eithei of the two gioups of non-submitteis.)

1.6
2.0
2.4
2.8
3.2
3.6
4.0
1.6 2.0 2.4 2.8 3.2 3.6 4.0
Submitter (N:41608)
Above-Average-Testing Non-Submitters (N:22232)
W/O Above-Average-Testing Non-Submitters (N:24610)
High School GPA
Academic Overlay: College Cumulative GPA and High School GPA
(students entering 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010)

Page 11 of 7u

Figuie 4.

Reaueis shoulu not attach too much piecision to these two giaphs.
viii
With tens of
thousanus of stuuents in each giaph, each uot (Su uots foi submitteis, Su foi above-
aveiage-testing non-submitteis anu Su foi below-aveiage-testing non-submitteis)
iepiesents hunuieus of stuuents. As with all ieseaich, it will be impoitant foi othei
scholais to examine anu ietest oui finuings. 0ne skilleu statistician on oui Auvisoiy
Committee commenteu in his peei ieview of the uiaft stuuy that the BSuPA tight 4S uegiee
slope was a "leptokuitic" scatteiplot, with the uata veiy tightly buncheu, while the football-
shapeu SAT scatteiplot was a moie typical, bioauly spieau set of uata. Nany of us who have
spent oui caieeis as seconuaiy oi univeisity faculty anu auministiatois finu compelling the
aigument that "what stuuents uo ovei foui yeais in high school is moie impoitant than
what they uo on a Satuiuay moining." Anu to state the obvious, iepeateuly seen in this
stuuy with vaiious cohoits, the tightly paiallel lines of the BSuPAcollege Cum uPA chait
ieflect #4) foui-yeai patteins of commitment anu self-uiscipline. The bioau illustiations of
these giaphs aie consistent with the finuings in this papei.

Foi just the value of seeing stuuent peifoimance ovei a long timeline, we also have chosen
to base oui piincipal analyses on the G0(0'&#"<2 college oi univeisity uPA, iathei than
Fiist-yeai uPA. All college auministiatois aie well awaie of the multiple anu sometimes
piofounu aujustments that stuuents must manage in theii fiist yeai of college. In choosing
the colleges anu univeisities foi this stuuy, we tiieu to incluue some with ieputations foi
woiking with stuuents who hau pioven themselves to eveiyone except peihaps the testing
1.6
2.0
2.4
2.8
3.2
3.6
4.0
600 800 1,000 1,200 1,400 1,600
Submitter (N:55673)
Above-Average-Testing Non-Submitters (N:25002)
W/O Above-Average-Testing Non-Submitters (N:24824)
SAT
C
o
l
l
e
g
e

C
u
m
u
l
a
t
i
v
e

G
P
A

Academic Overlay: College Cumulative GPA and SAT

Page 12 of 7u
agencies. Almost by uefinition, test-optional colleges anu univeisities will attiact anu enioll
laigei numbeis of these committeu anu uisciplineu stuuents. The ieliance on FYuPA seems
a systemic weakness of much valiuity ieseaich, as by uefinition it ignoies the coie mission
of oui colleges anu univeisities: to giauuate stuuents who have leaineu how to succeeu,
often in the face of obstacles. Nuch of that leaining, anu the visible success that follows
fiom it, occuis ovei the longei couise of a college euucation. We hope that extenuing the
lenses to tiack Cumulative uPA anu giauuation iates will be one of the key stiengths of this
stuuy.

It is sometimes aigueu that the ieason that stanuaiuizeu testing piouuces only veiy
mouest anu uneven pieuictive iesults is iestiiction of iange. Since stuuents with lowei
scoies aie not aumitteu, so the aigument iuns, most of the eniolling stuuents have stiongei
(anu closely clusteieu) test scoies, anu theiefoie the pieuictive value of the tests is
uiminisheu. uiven the uata in this stuuy, we can examine whethei this aigument foi
iestiiction of iange is tiue. Because of the optional testing policies, we have tens of
thousanus of stuuents in the stuuy with mouest scoies, most of them with veiy sounu
BSuPAs. We have just the wiue iange of scoies that supposeuly woulu be neeueu to
uemonstiate whethei iestiiction of iange is the cause foi the mouest pieuictive value of
testing. As will be seen, the veiy stiong coiielation of BSuPA with college Cumulative uPA
veisus the uneven anu mouest coiielation with test scoies is moie than a scailet thieau in
this stuuy--it is the biight coloieu fabiic of the entiie stuuy.

*&++,-. 1: ;-(0%(<,5 =(02(069

>4$ 9'&2. ,22-$99$9 :()$ %1-$ ?&$9'(109@ ;-(0%(<,5 :(02(069 :1- '4$9$ :()$ ?&$9'(109
,-$ 5(9'$2 A$51B.

1. C4,' <$-%$0',6$ 1: $0-155$2 9'&2$0'9 ,-$ 010D9&A+(''$-9E

F%-199 '4$ 9'&2.G HIJ 1: $0-155$2 9'&2$0'9 B$-$ ,2+(''$2 &02$- ,0 1<'(10,5
'$9'(06 <15(%.@ K&' , +1-$ -$:5$%'()$ 9','(9'(% (9 LMJ 1: $0-155$2 9'&2$0'9@ >4(9
LMJ (0%5&2$9 'B1 6-1&<9@ >4$ :(-9' (9 ,55 010D9&A+(''$-9 :-1+ '4$ 'B$0'.D9$)$0
<-(),'$G +(01-('.D9$-)(06 ,02 ,-'9 (09'('&'(109 B('4 <&-$ 1<'(10,5 '$9'(06 <15(%($9@
>4$ 9$%102 (9 105. '419$ 9'&2$0'9 ,' '4$ 9(# <&A5(% &0()$-9('($9 B('4 '$9' ,)$-,6$9
A$51B '4$(- (09'('&'(10,5 ,)$-,6$9G (@$@G '419$ B41 A$0$:(''$2 :-1+ ,0 1<'(10,5
'$9'(06 <15(%.@ >1 A$ 9&-$G '4$ 9'&2$0'9 ,' <&A5(% &0()$-9('($9 B('4 '$9' 9%1-$9
,A1)$ '4$(- (09'('&'(10,5 ,)$-,6$ B$-$ ,2+(''$2 915$5. 10 '4$ A,9(9 1: '4$(-
N*O;F 1- %5,99 -,0PG ,02 '4&9 ,-$ '$%40(%,55. 010D9&A+(''$-9@ K&' %1++10 9$09$
'$559 &9 '4,' '419$ 9'&2$0'9 B1&52 ,5+19' %$-',(05. 4,)$ A$$0 ,2+(''$2 B('4 1-
B('41&' '4$(- (09'('&'(10Q9 ,&'1+,'(% 1- 6&,-,0'$$2 ,2+(99(109 <15(%.@ >4&9 +&%4
1: 1&- -$9$,-%4 ,' '4$ <&A5(% &0()$-9('($9 2()(2$9 '4$ 010D9&A+(''$-9 (0'1 '419$
B('4 9%1-$9 ,A1)$ ,02 A$51B '4$(- (09'('&'(10,5 ,)$-,6$9G ,02 $#,+(0$9 '4$
<$-:1-+,0%$ 1: A1'4 6-1&<9@

At the twenty piivate colleges anu univeisities, non-submitteis compiiseu SS% of
all eniollees, iising fiom 26% to SS% of enteiing classes ovei the yeais of the stuuy.

Page 1S of 7u
At the six public univeisities, 6S% of all eniollees weie aumitteu on the basis of
theii BSuPA oi class iank without testing being taken into account. Thiity-five
peicent weie stuuents with testing above theii institution's aveiages who
piesumably woulu be aumitteu if testing weie consiueieu, but 28% hau test scoies
below class aveiages who weie beneficiaiies of the automatic aumissions policy. We
also examineu a smallei sub-gioup of stuuents (7.S% of all public univeisity
eniollees, a sub-set incluueu in the 28% above) who met the BSuPA ciiteiia but
with uiamatically lowei SATACT scoies.
At the five minoiity seiving institutions, 28% of eniollees weie non-submitteis. A
wiue iange of policies iesulteu in non-submittei iates fiom 1S% to 64%.
At the two aits institutions, non-submitteis iepiesenteu S4% of eniollees.

2. C4,' ,-$ '4$ %1+<,-,'()$ N*O;F9 ,02 %155$6$ 7&+&5,'()$ O;F9 1: 9&A+(''$-9
,02 010D9&A+(''$-9 1: '$9'(06E

F' ,55 %,'$61-($9 1: (09'('&'(109G %155$6$ ,02 &0()$-9('. O;F9 %519$5. '-,%P N*O;FG
$)$0 B('4 B(2$ ),-(,'(109 (0 '$9'(06@ F%-199 '4$ 9'&2.G 010D9&A+(''$-9 R01'
(0%5&2(06 <&A5(% &0()$-9('. 9'&2$0'9 B('4 ,A1)$D,)$-,6$ '$9'(06S $,-0$2 7&+
O;F9 '4,' B$-$ 105. @MH 51B$- '4,0 9&A+(''$-9G T@UL )$-9&9 T@UU@

At the piivate institutions, submitteis anu non-submitteis eain almost iuentical
uPAs in both high school anu college, with submitteis only .u4 highei in high school.
At giauuation, theii college cumulative uPAs weie still veiy close, only .1u apait,
with submitteis highei. Submitteis hau SATs (incluuing conveiteu ACTs) that weie
149 points highei than non-submitteis.
At the public univeisities, non-submitteis hau high school uPAs that weie
significantly highei than submitteis, both in high school anu univeisity, by .4 of a
uPA point. Consistently, stuuents who met the BSuPA iequiiement but with below-
aveiage testing eaineu highei univeisity uPAs than stuuents with lowei BSuPAs but
highei testing, by about .2S of a uPA point. These below-aveiage-testing non-
submitteis have aveiage SAT testing that is 9S points lowei than the submitteis.
At the minoiity-seiving institutions, submitteis peifoimeu bettei in both high
school anu college, by about .S of a uPA point. Foi both submitteis anu non-
submitteis, college uPAs tiack BSuPA, anu non-submitteis hau maikeuly lowei
BSuPAs. Both submitteis anu non-submitteis at these institutions have lowei
BSuPAs than at othei institutions in the stuuy, though theie aie hopeful signs of
impiovement in BSuPAs.
In the aggiegate of the two aits institutions, submitteis anu non-submitteis hau
laigely similai BSuPAs, poitfolio iatings anu college outcomes on uPA. We hau
hopeu to see some pieuictive value in the poitfolio iatings, but the N's weie too
small anu uata aveiages too close foi a ieliable analysis.

L@ C4,' ,-$ '4$ 6-,2&,'(10 -,'$9 ,02 '(+$D'1D%1+<5$'(10 -,'$9 1: 9&A+(''$-9 ,02
010D9&A+(''$-9E

Page 14 of 7u
O-,2&,'(10 -,'$9 1: 9&A+(''$-9 ,02 010D9&A+(''$-9 ,' '4$ <-(),'$ ,02 <&A5(%
(09'('&'(109 (0 '4(9 9'&2. ,-$ $#'-$+$5. %519$G 2(::$-(06 A. 105. @VJ ,%-199 '4$ 9'&2.G
,6,(0 B('4 ,A1)$D,)$-,6$D'$9'(06 9'&2$0'9 ,' <&A5(% &0()$-9('($9 01' (0%5&2$2@ W10D
9&A+(''$-9 B('4 9'-106 N*O;F9 01-+,55. 6-,2&,'$ ,' 4(64$- -,'$9 '4,0 9'&2$0'9 B('4
B$,P$- N*O;F9 A&' +&%4 9'-106$- '$9'(06@ =1- '4$ <-(),'$G <&A5(% ,02 ,-'9
(09'('&'(109 (0 '4$ 9'&2.G '(+$D'1D%1+<5$'(10 -,'$9 :1- 9&A+(''$-9 ,02 010D
9&A+(''$-9 ,-$ 5$99 '4,0 IJ ,<,-'@ F' '4$ +(01-('.D9$-)(06 (09'('&'(109G '(+$D'1D
%1+<5$'(10 -,'$9 :1- 9&A+(''$-9 B$-$ LJ 4(64$- '4,0 :1- 010D9&A+(''$-9G B('4 O;F9
,02 '(+$D'1D%1+<5$'(10 -,'$9 ,' '4$9$ +(01-('.D9$-)(06 (09'('&'(109 5,-6$5.
+(--1-(06 N*O;F9@ R>4$ 9'&2. 2(2 01' '-. '1 A-$,P 1&' :1&-G :()$ 1- 9(# .$,-
6-,2&,'(10 -,'$9X (: , 9'&2$0' 4,2 -$%$()$2 '4$(- 2$6-$$ ,' '4$ '(+$ '4$ 2,', B,9
%155$%'$2G 4$Y94$ B,9 %1&0'$2 (0 '4$ 9'&2. ,9 , 6-,2&,'$@S

At the twenty piivate institutions, the oveiall aggiegate giauuation iate is 76.9%.
Bepenuing on the methou of calculation (ieacheu by aveiaging one aggiegate pool
of stuuents oi by aveiaging the institutional giauuation iates), non-submitteis
giauuate at iates eithei 1.1% highei oi 1.9% lowei than submitteis, but in eithei
case, with veiy similai giauuation outcomes.
In the public institutions, with an oveiall giauuation iate of 66%, non-submitteis
giauuate at iates S% above submitteis, 68% veisus 6S%. Non-submitteis with
below-aveiage testing giauuate at iates 4% above submitteis.
In the minoiity-seiving institutions, submitteis giauuate at iates 1S% highei than
non-submitteis, S7% veisus 24%. At the minoiity-seiving institutions, lowei
aveiage giauuation iates foi non-submitteis aie uiiven by significantly highei
peicentages of non-submitteis with maikeuly lowei BSuPAs.
4. At the only aits institution with giauuateu cohoits, submitteis giauuate at iates
11% highei than non-submitteis, 64% veisus SS%, but this seems laigely unielateu
to acauemic cieuentials, which have no significant uiffeiences.

Z@ C41 &9$9 ,0 1<'(10,5 '$9'(06 <15(%.E

W10D9&A+(''$-9 ,-$ +1-$ 5(P$5. '1 A$ B1+$0G :(-9'D6$0$-,'(10D'1D%155$6$G ,55
%,'$61-($9 1: +(01-('. 9'&2$0'9G ;$55 O-,0' -$%(<($0'9G ,02 9'&2$0'9 B('4 [$,-0(06
\(::$-$0%$9 R[\S@ K&' ,%-199 (09'('&'(10,5 '.<$9G B4('$ 9'&2$0'9 ,591 &9$ 1<'(10,5
'$9'(06 <15(%($9 ,' -,'$9 B('4(0 9(065$ 2(6('9 1: '4$ ,)$-,6$9 :1- '4$(- (09'('&'(10G 91
'4$ <15(%($9 9$$+ '1 4,)$ A-1,2 ,<<$,5 ,%-199 $'40(% 6-1&<9@

Theie aie not uiamatic uiffeiences in eniollments by high school type. Public,
paiochial anu piivate high school stuuents use optional testing policies at iates
within single uigits of each othei. In piivate colleges anu univeisities, a slightly
highei use of optional testing policies by stuuents fiom paiochial anu piivate
schools may ieflect focuseu college counselois with lowei stuuent loaus.
Theie is mouest eviuence at piivate institutions of athletes who have been iateu by
coaches in the aumissions piocess using an optional testing policy at highei iates.

Page 1S of 7u
LB stuuents, fiom a sample of 8 piivate institutions, weie much moie likely to apply
as non-submitteis.

H@ C4,' ,-$ '4$ <-,%'(%,5 ,<<5(%,'(109 1: ,0 1<'(10,5 '$9'(06 <15(%. :1- $0-155+$0'
<5,00(06E

W10D9&A+(''$-9 9&<<1-' 9&%%$99:&5 $0-155+$0' <5,00(06 (0 +&5'(<5$ B,.9G :-1+
,<<5.(06 ",-5. \$%(9(10 ,' 4(64$- -,'$9G '1 (0%-$,9$2 $0-155+$0'9 A. +(01-('.
9'&2$0'9G '1 A-1,2 6$16-,<4(% ,<<$,5G '1 , 'B1D',(5 %&-)$ 1: :,+(5. :(0,0%(,5
%,<,%('.G '1 9&%%$99 A. [$,-0(06 \(::$-$0%$ 9'&2$0'9@ N1B$)$-G 010D9&A+(''$-9
+,. %1++105. A$ +(99$2 (0 %109(2$-,'(10 :1- +$-(' ,B,-29G 2$9<('$ A$''$- 7&+
O;F9 ,02 +,-P$25. 4(64$- 6-,2&,'(10 -,'$9 '4,0 '4$ 9&A+(''$-9 B41 -$%$()$
+$-(' ,B,-29@

Non-submitteis in piivate institutions aie 1u% moie likely to use an Eaily Becision
policy than apply foi iegulai aumission, not a tiivial uiffeience in competition foi
yielu.
In piivate institutions, optional testing has bioau geogiaphic ieach, with non-
submitting stuuents eniolling fiom outsiue an institution's noimal iegional
teiiitoiies. In the public institutions, foui of the six have maue theii guaianteeu
aumissions policies available to out-of-state stuuents, iunning countei to the
pattein of having guaianteeu aumissions be an auvantage foi in-state stuuents.
Albeit with small numbeis, theie is some eviuence of optional testing helping these
public institutions inciease theii geogiaphic ieach.
In a suipiise finuing, both piivate anu public institutions ieflect a bioau pattein of
both low-income anu high-income stuuents using optional testing policies at highei
iates. This cieates a "two-tail cuive" with impoitant financial planning implications.
0n the one hanu, non-submitteis weie moie likely to be fiist-geneiation-to-college,
minoiity, oi Pell iecipients, anu thus cieate neeus foi financial aiu. 0n the othei
hanu, a significant pool of non-submitteis not asking foi financial aiu helps balance
institutional buugets.
In the fiist of thiee sub-stuuies, we examineu almost 27,uuu stuuents who ieceiveu
meiit financial awaius. When all meiit iecipients weie incluueu, non-submitteis
anu submitteis ieceiveu meiit awaius at almost exactly the same iates, but non-
submitteis eaineu maikeuly highei Cum uPAs of .2S ovei submitteis anu giauuateu
at iates that weie 11% highei, 81% veisus 7u%. When the above-aveiage-testing
non-submitteis weie iemoveu fiom the compaiisons, the numbei of meiit awaius
to the iemaining non-submitteis uioppeu uiamatically fiom 1S,7u8 to S,u64, but
these below-aveiage-testing non-submitteis still eaineu slightly highei Cum uPAs
anu giauuateu at iates 6% highei than the submitteis. These non-submitteis aie
also at statistically significant levels moie likely to be unueiiepiesenteu minoiities,
fiist-geneiation-to-college, women, Pell iecipients, anu STEN majois. While the
iationales foi meiit awaius vaiy wiuely, if a coie assumption is that they shoulu
iewaiu stuuents with no-neeu funuing who aie moie likely to uo well anu giauuate,
that assumption seems not to be boine out by the uata. Institutions with optional

Page 16 of 7u
testing policies may want to examine theii ciiteiia foi meiit awaius, especially the
use of stanuaiuizeu testing to qualify stuuents foi no-neeu meiit funuing.
In the seconu sub-stuuy, a basic R-squaie founu that BSuPA accounts for mosL of Lhe
varlaLlon ln Cum CA, with testing pioviuing an auuitional few peicentage points. At
public institutions, testing seemeu to auu especially low peicentages of coiielation
to BSuPA, often only 1% to S%, while it auueu some auuitional coiielation at
piivate institutions. We uiu not obseive that testing hau maikeuly lowei R-squaie
values foi non-submitteis, anu R-squaie analyses in geneial seem to piouuce mixeu
coiielations foi non-submitteis. 0ui stuuy chose to use othei piincipal statistical
tools, Cohen's u anu chi-squaie, in oui analyses, but since R-squaie is commonly
useu, we wanteu to pioviue at least a baseline analysis.
The thiiu sub-stuuy examineu the iecoius of 1uS9 stuuents at eight piivate
institutions with uiagnoseu Leaining Biffeiences. We founu that LB stuuents use
optional testing policies at significantly highei iates, Su% veisus SS% foi non-
submitteis geneially at the piivate institutions. The LB stuuents, both submitteis
anu non-submitteis, eain uPAs anu giauuate at iates that aie only slightly below
non-LB stuuents. A sepaiate stuuy at Bates founu that given basic accommouations
to which they aie legally entitleu, LB stuuents uiamatically impioveu theii
Cumulative uPAs anu giauuation iates ovei two uecaues fiom Su% to 9u%, Bates'
oveiall giauuation iate.
ix
0vei this time, Bates stiengtheneu its auministiation of LB
suppoit, anu LB stuuents weie moie knowleugeable anu open about theii LBs.
These changes in the success of LB stuuents hau the effect of iaising Bates' oveiall
giauuation iate by seveial peicentage points, not a tiivial change foi faculties,
piesiuents anu tiustees. While acknowleuging the issues of meuical piivacy, how
colleges anu univeisities can effectively seive LB stuuents is woithy of much moie
attention anu cooiuinateu ieseaich than it seems to have gotten to uate.

Beie then aie oui iesults as oiganizeu by the foui categoiies of institutions we stuuieu,
followeu by concluuing iemaiks.

/@ >B$0'. ;-(),'$ 7155$6$9 ,02 ]0()$-9('($9X ;-(0%(<,5 =(02(069
The twenty piivate college anu univeisities in the stuuy incluue fifteen colleges anu five
univeisities, ianging in size fiom SSu to 7Suu stuuents. These finuings aie baseu on the
aggiegate uata fiom S6,8S9 iecoius at the twenty institutions.
x

What Peicentage of Eniolleu Stuuents 0se an 0ptional Testing Policy.
Fiom 2uuS to 2u1u, non-submittei eniollees iose giauually fiom 26% to SS% in this
aggiegate cohoit of the stuuy. The 2u colleges anu univeisities have a wiue iange of
eniollee non-submittei iates, fiom 9% to 6S%. (Figuie S.)

Page 17 of 7u

Figuie S.
While the numbei of piivate institutions with optional test policies continues to expanu
mouestly, the shaie of stuuents within these institutions choosing to be non-submitteis is
also climbing ovei time. (Figuie 6.) We uiu not gathei uata to analyze aumissions funnels,
so uo not know whethei this incieaseu shaie of non-submitteis is uue to laigei pools of
non-submittei applicants fiom which to choose stable eniollments, highei yielu fiom
aumissions offeis to non-submitteis, ieshapeu aumissions piioiities by aumissions staffs,
oi colleges using non-submittei applications to inciease oveiall eniollments. As with
seveial othei facets of this stuuy, the aumissions funnel uata is a piomising topic foi
fuithei stuuy.
140
183
237
493
294
316
373
649
662
578
554
632
1432
109
835
1064
1379
959
782
333 students
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
PR3
PR10
PR16
PR8
PR14
PR20
PR6
PR18
PR2
PR7
PR11
PR12
PR5
PR17
PR1
PR15
PR9
PR13
PR19
PR4
Policy Use by Private Institutions, by Policy Type
(non-submitters entering 2003, 2004, 2006, 2007, 2008, 2009, 2010)
Percent Non-Submitter

Pure Optional Testing: all
applicants can choose to be
Non-Submitters
Optional Plus: Non-
Submitters are required to
submit additional materials,
such as an additional essay
or portfolio materials
Academic Threshold: Non-
Submitters must meet a
minimum threshold,
normally a certain HSGPA
or class rank.

Page 18 of 7u

Figuie 6.
Who 0ses an 0ptional Testing Policy.
New Englanu stuuents have highei non-submittei iates than othei geogiaphic iegions,
peihaps because moie colleges anu univeisities in the stuuy aie locateu in New Englanu oi
the Niuule Atlantic states, anu because these New Englanu institutions have hau optional
testing policies foi longei peiious of time. But the policy has bioau geogiaphic appeal:
peicentage use within the national geogiaphic iegions ianges fiom 24% to 49%. At many
colleges, we noticeu a pattein of the institutions uiawing highei peicentages of non-
submitteis at long uistances fiom theii campus, so non-submittei policies seem to be
helping colleges to iaise theii national appeal. Non-submittei iates vaiy mouestly by high
school type: Su% aie fiom public schools, S7% fiom paiochial, anu S8% fiom piivate
schools. (Figuie 7.)
26%
35%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2003 2004 2006 2007 2008 2009 2010
Non-Submitter Growth at Private Institutions from 2003 to 2010
Cohort Year
P
e
r
c
e
n
t

N
o
n
-
S
u
b
m
i
t
t
e
r

Page 19 of 7u

Figuie 7.
It may seem a bit counteiintuitive to finu even mouestly highei iates of non-submitteis
fiom paiochial anu piivate schools, given that many non-submitteis aie minoiity, fiist-
geneiation-to-college oi Pell iecipients. But keep in minu the stuuent-to-guiuance-
counseloi iatios: at public schools acioss Ameiica, the iatio is ovei 47u-1, while at most
piivate schools theie aie college counselois with focuseu uuties anu geneially with iatios
well unuei 1uu-1. The coie thesis of }onathan Kozol's Savage Inequalities is that Ameiica
spenus substantially moie pei heau on the public K-12 euucation of its wealthy stuuents
than on the public K-12 pooi stuuents. In each of the geogiaphic aieas Kozol stuuieu, the
top-funueu subuiban school system spent about twice pei stuuent what the neaiby innei
city school system spent, anu Kozol's ieseaich ueals only with public schools.
xi
In this
focuseu stuuy of optional testing, as in so many othei facets of the aumissions woilu, the
painful uispaiities of seivices aie visible at unexpecteu moments.
0ptional testing also seems to suppoit effoits to uiaw uiveise stuuent bouies. Women,
fiist-geneiation-to-college stuuents, Pell uiant iecipients, all categoiies of minoiity
stuuents, athletes, anu LB stuuents all choose optional testing at highei iates. Non-
submitteis also choose to be eaily uecision applicants at highei iates, anothei finuing with
impoitant eniollment implications. (Figuie 8.)
144
161
182
666
1061
311
844
3898
4321
2
20
126
5833
77
1673
2540 students
0% 20% 40% 60% 80% 100%
Rocky Mtns
Plains
Southwest
Great Lakes
Far West
Outside US
Southeast
Mid East
New England
Military Base
Outlying
Homeschooled
Public
International
Parochial
Private
G
e
o
g
r
a
p
h
i
c

R
e
g
i
o
n

H
i
g
h
-
S
c
h
o
o
l

T
y
p
e

Percentage of Non-Submitters at Private Institutions by Geography and by High School Type

Page 2u of 7u

Figuie 8.
The optional testing policies seem to be suppoiting incieases in unueiiepiesenteu
minoiity stuuents at these twenty colleges anu univeisities. But 29% of white stuuents
also chose to be non-submitteis, only low single uigits below the oveiall aveiage of S2%.
So while optional testing seems to woik as an affiimative action uevice, the policy also has
wiue appeal acioss all ethnic categoiies, incluuing white stuuents, who veiy substantially
outnumbei all categoiies of minoiity anu inteinational stuuents combineu. To those who
see ieuuctions in emphasis on test scoies as opening a back uooi foi affiimative action, we
woulu point out that no optional testing policy is baseu on iace, ethnicity, citizenship,
genuei oi any of the othei categoiy aiguments that have beuevileu highei euucation. As we
have been telling stuuents applying to Bates foi the last Su yeais, "Nake youi own uecision
about the tests. If you think they accuiately ieflect youi piomise, submit them, anu they
will be consiueieu as pait of the uecision. If you think they uo not ieflect youi piomise,
uon't submit them, anu theii absence will not be helu against you." (Figuie 9.)
Male: 4184
Not First Generation: 7229
Not Pell: 8561
Regular: 4988
Female: 7820
First Generation: 1418
Pell: 2545
Early: 3317 students
0% 20% 40% 60% 80% 100%
Gender
First Generation
Pell
Admit Type
Percentage of Non-Submitters at Private Institutions by Admit Type, Pell, First Generation, and
Gender Status

Page 21 of 7u

Figuie 9.
0ne unexpecteu finuing iuns consistently acioss the institutional gioupings in the stuuy:
both low-income anu high-income stuuents (incluuing as high income the "No FAFSA"
stuuents not applying foi Feueial financial aiu) use the policy at highei iates, cieating a
"two-tail" oi "banana cuive" financial pattein. While non-submittei policies suppoit
stuuent uiveisity anu thus cieate financial aiu obligations, a ioughly equal pool of stuuents
who aie not financial aiu iecipients help maintain balanceu institutional buugets. To put it
miluly, theie aie extiemely impoitant eniollment planning anu financial moueling
implications in these finuings, though institutions will neeu to examine theii own
compaiisons between submitteis anu non-submitteis. (Figuie 1u.)
7525
578
577
65
1411
834
1014 students
0% 20% 40% 60% 80% 100%
White, non-Hispanic
Nonresident Alien
Asian/Pacific Islander
American Indian/Alaskan Native
Unknown
Black, non-Hispanic
Hispanic
Policy Use at Private Institutions by Ethnicity and Citizenship
Note: This is based on the former IPEDS categories. Students from cohorts 2009 and 2010 who selected multiple backgrounds were re-categorized as "unknown":

Page 22 of 7u

Figuie 1u.
What aie the Acauemic 0utcomes of Non-Submittei Policies.
Submitteis anu non-submitteis entei college with veiy closely paiallel BSuPAs (.u4 apait)
but ieasonably laige uiffeiences in SAT scoies of 149 points. In college, non-submitteis
stait with a fiist-yeai uPA uiffeience of .1S below submitteis, but at giauuation the uPA
gap is .1u. So uespite the test scoie uiffeiences, submitteis anu non-submitteis in both high
school anu college aie only sepaiateu by & 624 $031721#$* )6 & =>? @)"3#B Nay we pause in
this mountain of statistics to unueiline a common-sense obseivation. If almost S7,uuu
stuuents at twenty piivate institutions stait with BSuPAs that aie .u4 apait anu enu with
college Cum uPAs that aie .1u apait, & *$"6# )6 B/H )6 & =>? @)"3# is tiivial. (Figuie 11.)
415
3489
20
7
49
306
257
608
1280
3337
1198
0% 20% 40% 60% 80% 100%
Non-Resident Alien
No FAFSA
$70,000 or more
$60,000 - $69,999
$50,000 - $59,999
$40,000 - $49,999
$30,000 - $39,999
$20,000 - $29,999
$10,000 - $19,000
$1 - $9,999
$0
Policy Use at Private Institutions by Expected Family Contribution
E
x
p
e
c
t
e
d

F
a
m
i
l
y

C
o
n
t
r
i
b
u
t
i
o
n

B
a
n
d
s
*

Note: EFC values were adjusted for inflation.

Page 2S of 7u

Figuie 11.
The giauuation iates of submitteis anu non-submitteis aie quite close, within a few
peicentage points. The oveiall giauuation iate foi this aggiegate sample is 76.9%. The non-
submittei giauuation iate is 1.1% highei when calculateu as one aggiegate pool; when
calculateu as an aveiage of the 2u institutional giauuation iates, the non-submittei iate is
1.9% lowei. In eithei case, with almost S7,uuu iecoius at twenty institutions, theie is an
insignificant uiffeience. Aggiegate time-to-completion iates, calculateu by counting foui-
yeai giauuations as 1uu% of a time line, aie almost iuentical, uiffeiing by .8%. (Figuie 12.)

3.53
3.14 3.17
3.47
3.10
3.18
3.49
2.99
3.04
3.43
2.96 3.08
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
High School
GPA
First Year
GPA
Cumulative
GPA
High School
GPA
First Year
GPA
Cumulative
GPA
Submitter Non-Submitter
ENROLLED COHORTS GRADUATED COHORTS
Policy Use by GPA at Private Institutions - High School, College First Year, College Cumulative
Note: The focus of this chart is on the difference between non-submitter and submitter means.

Page 24 of 7u

Figuie 12.
Thioughout this stuuy, coloi couing in these compaiative chaits ieflects statistical
uiffeiences, with a Cohen's u analysis laigely useu in the top chait anu a Chi-Squaie
analysis laigely in the bottom chait. In the top chait, the SAT uiffeiences of 149 points
between submitteis anu non-submitteis ieflect a laige statistical uiffeience, anu all othei
seven outcome chaiacteiistics being measuieuincluuing BSuPA, college uPA anu
giauuation iateaie small oi tiivial uiffeiences. In contiast, in the lowei chait, theie aie
statistically significant uiffeiences in five of seven fielus.
Non-submitteis aie less likely to be STEN majois, anu to eain Latin honois with vaiious
"Cum Lauue" awaius. But as in so many uata facets of this stuuy, the peicentage
uiffeiences aie small: submitteis favoi BiologyLife Sciences anu non-submitteis favoi
Psychology by uiffeiences of about 2%. The iatios aie paiallel in Physical Sciences, Nath,
Computing anu Engineeiing. The ieal issue foi these Ameiican piivate colleges anu
univeisities may be that low peicentages of stuuents, both submitteis anu non-submitteis,
aie choosing STEN fielus. Nany stuuies have pointeu out this issue of choice of majois, anu
the paiallel issue of uispiopoitionate peicentages of STEN majois being inteinational
stuuents. (Figuie 1S.)

Non-
Submitters
Submitters

High School GPA 3.47 3.51 Cohens d
Academic Rating 6.53 6.76 Cohens d

1096 1245 Cohens d
First Year GPA

2.98 3.13 Cohens d
Cum GPA (enrolled cohorts)

3.04 3.17 Cohens d
Cum GPA (graduated cohorts) 3.08 3.18 Cohens d
Graduation Rate** 77.7% 76.6% Chi-Square
Completion Rate** 101.4% 102.2% Cohens d

Underrepresented Minority 16% 9% Chi-Square
First Generation

16% 10% Chi-Square
Gender (Female) 65% 59% Chi-Square
Pell 23% 17% Chi-Square
EFC

$21,790 $26,303 Cohens d
EFC Adjusted for Inflation $10,570 $12,817 Cohens d
STEM Major

24% 32% Chi-Square

Academic Rating: All institutions submitted their respective scales, but for comparison purposes we converted all of them to a 10 point
scale, where 10 is the highest rating.
SAT Caveat: Only 41% of Non-Submitters still submitted scores. This data on the average scores of non-submitters only represents that
41%. All other data in this chart represents the full set of non-submitter and submitters, with all the data that was submitted.
** Graduated Cohorts Only
Summary of Key Statistics at Private Institutions
COLOR KEY
Cohens D


COLOR KEY
Chi-Square Tests


Page 2S of 7u

Figuie 1S.
With only iecent giauuateu cohoits, we coulu offei no impoitant conclusions about caieei
outcomes oi giauuate stuuy in any of the institutional gioupings in the stuuy.
The summaiy below, with key uata fiom each of the 2u piivate institutions, ieflects what
ieaueis will finu thioughout this stuuy: 4$23 3)3;*0:("##27* &31 *0:("##27* &72 9)(@&721
&97)** (0'#"@'2 97"#27"&. #$2 $2&<"2*# *#&#"*#"9&' 1"66272392* &72 6)031 "3 #2*#"38 *9)72*. :0#
82327&''5 &72 3)# 6)031 "3 IJ=>?*. 6"7*# 52&7 =>?*. G0( =>?* )7 87&10&#")3 7&#2*B In othei
woius, most of the uiffeiences between the two pools of submitteis anu non-submitteis aie
in theii SAT scoies. All othei peifoimance anu outcome measuies aie essentially equal.
(Figuie 14.)

0% 10% 20% 30%
Engineering
Industrial Arts & Consumer
Law & Public Policy
Communications and Journalism
Agriculture and Natural Resources
Health
Education
Computers & Mathematics
Physical Sciences
Arts
Psychology & Social Work
Business
Biology & Life Science
Withdrew Undecided
Social Science
Undecided
Humanities & Liberal Arts
Percent Major Selection at Private Institutions by Submitter Status
Non-Submitters Submitters
Percent Major Selection Percent Major Selection
0% 10% 20% 30%
Engineering
Industrial Arts &.
Law & Public Policy
Agriculture and Natural.
Communications and.
Health
Computers & Mathematics
Education
Physical Sciences
Arts
Business
Biology & Life Science
Psychology & Social Work
Withdrew Undecided
Social Science
Undecided
Humanities & Liberal Arts

Page 26 of 7u

Figuie 14.
//@ *(# ;&A5(% ]0()$-9('($9X ;-(0%(<,5 =(02(069
This peei compaiison is baseu on aggiegate cohoits fiom six public univeisities, a
combineu total of 71,8S1 iecoius. Two univeisities aie public flagships, two aie the
piincipal state univeisities in states with a flagship, one is a piincipal scientific anu
technical univeisity, anu one is an uiban minoiity-seiving campus of a state univeisity
system (though counteu in this stuuy only as a public univeisity, anu not in the minoiity-
seiving section).
Each of the public univeisities has an aumissions policy allowing foi automatic aumission
with a ceitain BSuPA anuoi class iank. So while they noimally collect ACT oi SAT testing
fiom all applicants, stuuents meeting the uPA oi iank ciiteiia aie aumitteu iegaiuless of
testing scoies, anu thus aie stiuctuially test-optional, oi as we have teimeu them foi
consistency with the piivate colleges anu univeisities in this national stuuy, "Non-
Submitteis." The public univeisity poition of this stuuy compaies the outcomes of these
stuuents meeting the uPAiank ciiteiia with those with lowei uPAsianks foi whom
testing was evaluateu as pait of the aumissions uecision ("Submitteis").
Private Institutions--Academic Comparison Summary

High School GPA SAT (CR + Math) First Year GPA
Cumulative GPA
(Graduated Cohorts)
Graduation Rate

Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub
PR1 3.38 3.43 1146 1332 3.02 3.27 3.15 3.34 89.2% 89.4%
PR2 3.57 3.62 1112 1294 3.27 3.43 3.34 3.44 82.1% 82.6%
PR3 3.14 3.11 No Data 1098 2.83 2.73 2.92 2.74 56.0% 59.8%
PR4 3.86 3.54 1307 1265 3.44 3.22 3.45* 3.24* No Data No Data
PR5 3.38 3.45 1068 1213 2.90 3.10 3.18 3.29 82.6% 83.3%
PR6 3.49 3.63 1068 1283 2.83 3.13 3.03 3.26 69.2% 74.1%
PR7 3.22 3.39 1086 1283 2.89 3.18 2.91* 3.19* No Data No Data
PR8 3.63 3.49 1109 1276 2.96 3.07 3.24 3.06 88.3% 86.4%
PR9 3.71 3.73 1276 1275 3.03 3.19 3.16 3.30 90.6% 91.3%
PR10 3.34 3.44 No Data 1257 No Data No Data No Data No Data 58.0% 62.8%
PR11 3.35 3.39 950 1107 2.67 2.85 2.72 2.87 55.1% 56.0%
PR12 3.25 3.37 981 1177 2.79 3.07 2.98 3.12 66.6% 66.5%
PR13 3.26 3.41 1045 1232 2.82 3.06 2.93 3.14 75.0% 77.1%
PR14 3.76 3.79 No Data 1297 3.19 3.37 3.22* 3.39* No Data No Data
PR15 3.55 3.62 No Data 1207 2.93 3.12 3.02 3.18 76.4% 79.8%
PR16 3.61 3.68 1233 1271 2.88 3.14 3.06 3.21 76.3% 77.9%
PR17 3.44 3.54 1061 1197 No Data No Data No Data No Data 62.2% 69.0%
PR18 3.83 3.80 1176 1311 3.04 3.15 3.06* 3.17* No Data No Data
PR19 3.58 3.53 No Data 1242 3.41 3.36 3.37 3.41 76.9% 77.1%
PR20 3.38 3.50 1058 1266 2.77 3.09 2.77* 3.07* No Data No Data

*No graduated cohorts . This number represents enrolled cohorts.

Page 27 of 7u
To evaluate moie piecisely the outcomes in Cum uPAs anu giauuation iates of stuuents
with vaiious levels of testing, we uiviueu both non-submitteis anu submitteis into sub-
gioups, especially to be able to evaluate stuuents aumitteu on the basis of the uPAiank
ciiteiia with testing both above anu below the class aveiages at theii univeisity. Fiom the
non-submitteis, we cieateu thiee gioups. "Above-Aveiage-Testing Non-Submitteis" have
both the iequiieu BSuPAiank anu testing above theii institution's aveiage. "Below-
Aveiage-Testing Non-Submitteis" have the iequiieu BSuPAiank but testing below theii
institution's aveiage testing. "Low-Testing Non-Submitteis" have the iequiieu BSuPAiank
but testing well below theii institution's aveiage. The "Low-Testing Non-Submitteis" aie a
subset of the "Below-Aveiage-Testing Non-Submitteis," so aie incluueu in the analyses of
the "Below-Aveiage-Testing" stuuents. "Submitteis", those who uo not meet the BSuPA oi
iank ciiteiia, weie also uiviueu into subgioups: those who met a lowei stanuaiu of
BSuPAiank with testing ("Scoie Submitteis"), anu those who uiu not meet the
BSuPAiank oi testing stanuaius but who weie aumitteu on the basis of inuiviuual
ieauings ("Categoiy aumits").
Theie is consiueiable vaiiation in state stanuaius foi automatic aumission without testing
being consiueieu. 0ne of the institutions in oui stuuy iequiies a top 1u% BS iank, one
iequiies a S.S BSuPA, one iequiies a S.2, two iequiie a 2.S, anu one a 2.u. Two of the public
institutions have moie stiingent stanuaius foi out-of-state stuuents, but inteiestingly, foui
allow out-of-state stuuents to be aumitteu as non-submitteis. (0ne of these allows out-of-
state stuuents to be aumitteu without testing but with a highei uPAiank iequiiement than
foi in-state.) In the chait below aie the paiticulai aumissions ciiteiia foi each univeisity,
with the "N's" foi each sub-categoiy foi each institution. As the outcomes analysis follows
fiom this chait, a moment to look it ovei caiefully will pay uiviuenus. Figuie 1S below
pioviues the uata on how each of the six univeisities was analyzeu.

Page 28 of 7u
Public University Student Segments: Defining the Subgroups of Admit Category
(non-submitters entering 2003, 2004, 2009, 2010)

Policy Non-Submitters
Above-Average-Testing Non-
Submitters
Below-Average-Testing Non-
Submitters
Low-Testing
Non-Submitters
PU1
In-state GPA!3.2
35.8% (3643 of 10183)
SAT!1050
23.5% (2395 of 10183)
SAT<1050 or no SAT
12.2% (1248 of 10183)
SAT<950
4.5% (455 of 10183)
PU2
High School GPA!3.5
44.4% (2799 of 6302)
SAT!1095
25.9% (1634 of 6302)
SAT<1095 or no SAT
17.4% (1165 of 6302)
SAT<995
8.3% (524 of 6302)
PU3
In-state top 10%
51.3% (14239 of 27687)
SAT!1300
12.7% (3507 of 27687)
SAT<1300 or no SAT
38.8% (10732 of 27687)
SAT<1050
9.1% (2506 of 27687)
PU4
HS GPA!2.5 or top 50%
84.8% (7727 of 9115)
SAT!1030
58.0% (5284 of 9115)
SAT<1030 or no SAT
26.8% (2443 of 9115)
SAT<900
7.7% (699 of 9115)
PU5
In-state GPA!2.0 or top 50%
Out-state GPA!2.0 or top 33%
95.3% (14812 of 15550)
In-state SAT!980
Out-state SAT!1090
73.5% (11436 of 15550)
In-state SAT<980 or no SAT
Out-state SAT<1090 or no SAT
21.7% (3376 of 15550)
SAT<900
5.8% (899 of 15550)
PU6
High School GPA!2.5
75.3% (2175 of 2888)
SAT!990
40.3% (1163 of 2888)
SAT<990 or no SAT
35.0% (1012 of 2888)
SAT<800
11.1% (321 of 2888)
Total
63.3%
(45395 of 71725)
35.4%
(25419 of 71725)
27.9%
(19976 of 71725)
7.5%
(5404 of 71725)

Figuie 1S.
"Policy Non-Submittei": The institutional policies on stanuaiuizeu testing alloweu
these stuuents to be aumitteu baseu only on BSuPA oi Rank, iegaiuless of SATACT
scoie. This iepiesents a uiiect tianslation of each institution's policy. It incluues all
stuuents aumitteu unuei the policy. Foi each institution, listeu is the ciiteiia foi
aumission unuei that state's policy, the peicentage shaie of the submitteu iecoius
aumitteu unuei the policy, the N's of the numbeis of non-submitteis anu the total
size of the cohoits submitteu, noimally foui class cohoits. This layout of infoimation
is continueu acioss all foui veitical columns.
"Above-Aveiage-Testing Non-Submitteis": This gioup is a subset of the Policy Non-
Submitteis: stuuents who weie aumitteu baseu on BSuPAiank, anu whose
stanuaiuizeu test scoies weie above that institution's aveiage scoies. These
stuuents piobably woulu have gaineu entiy to the univeisity 2<23 "6 the cuiient
policy on testing uiu not exist. Listeu foi each univeisity is the SAT scoie, noimally
eithei the SAT iequiieu foi Aumission if the uPA oi Rank stanuaius aie not met, oi
the class aveiage SAT scoie, that uefines this gioup.
! ! ! !

Page 29 of 7u
"Below-Aveiage-Testing Non-Submitteis": This gioup is a subset of the "Policy Non-
Submitteis," stuuents who weie aumitteu baseu on BSuPA, but whose stanuaiuizeu
test scoies might not have gaineu them entiy to the univeisity "6 the cuiient policy
on testing uiu not exist. These stuuents aie the beneficiaiies of the automatic oi
guaianteeu aumissions policy that uoes not consiuei stanuaiuizeu tests. They will
be the piincipal focus of the uiscussion below.
"Low-Testing Non-Submitteis": This gioup is a subset of the "Below-Aveiage-
Testing Non-Submitteis", also aumitteu on theii BSuPAiank, but with significantly
lowei testing than theii classmates.
What Peicentage of Eniolleu Stuuents 0se an 0ptional Testing Policy.
Fiom 2uuS to 2u1u, non-submittei eniollees ueclineu slightly fiom 66% to 62% of this
aggiegate cohoit. The univeisities have a wiue iange of eniollee non-submittei iates, fiom
9S% uown to S9%, with these peicentages shapeu paitly by the size anu stiength of theii
applicant pool, anu paitly by that institution's BSuPA oi iank ciiteiia. (Figuie 16.)

Figuie 16.
Who 0ses an 0ptional Testing Policy.
As at the piivate colleges anu univeisities, at the public univeisities we see that highei
peicentages of eniolleu stuuents who aie fiist-geneiation-to-college stuuents, Pell uiant
3643
2799
14239
2175
7727
14812 students
0% 20% 40% 60% 80% 100%
PU1
PU2
PU3
PU6
PU4
PU5
Policy Use by Public Institutions, by Percent of Use

Page Su of 7u
iecipients, all categoiies of minoiity stuuents anu women eniolleu unuei the public
univeisities' "guaianteeu aumission" optional testing policies tieu to BSuPA oi high school
iank. The policy seems to be suppoiting incieases in unueiiepiesenteu minoiity stuuents,
but all ethnic categoiies of stuuents aie eniolling unuei these policies, at iates fiom S9% to
71%. So while optional testing seems to woik as an affiimative action uevice, the policy
also has wiue appeal acioss all ethnic categoiies, incluuing white stuuents. As a policy
baseu puiely on BSuPA oi iank, it favois those stuuents who have peifoimeu well in theii
high schools. (Figuie 17.)

Figuie 17.
Since these public univeisities uo not have Eaily Becision policies anu we uiu not collect
aumissions funnel uata, we weie not able to calculate likelihoou of non-submitteis veisus
submitteis applying anu eniolling unuei the policy. But the uata pioviueu back to each
univeisity in the stuuy shoulu allow foi those yielu statistics to be calculateu fiom
Aumissions office uata.
Also paiallel to the finuings in the piivate colleges anu univeisities, both low-income anu
high-income stuuents use the policy at highei iates, cieating a "two-tail" financial pattein.
While a non-submittei policy suppoits stuuent uiveisity anu thus cieates financial aiu
obligations, a laige pool of stuuents who aie not financial aiu canuiuates helps maintain
balanceu institutional buugets. The peicentages of non-submitteis aie ieasonably equal in
the vaiious bais, but the "N's" at the enu of the bais ieflect the cleai two-tail cuive. As in
902
1733
2607
3639
36079
435 students
0% 20% 40% 60% 80% 100%
Unknown
Black, non-Hispanic
Hispanic
White, non-Hispanic
Policy Use at Public Universities by Ethnicity and Citizenship

Page S1 of 7u
piivate institutions, theie woulu seem to be impoitant eniollment planning anu financial
planning implications with these automatic aumission policies. (Figuie 18.)

Figuie 18.
What aie the Acauemic 0utcomes of Automatic Aumissions Policies Baseu on BSuPA anu
Rank.
Acauemic outcomes of submitteis anu non-submitteis aie maikeuly uiffeient. Non-
submitteis aumitteu unuei the BSuPA anu Rank policies eain highei Cum uPAs anu
giauuate at highei iates than stuuents aumitteu unuei the othei aumissions piotocols, with
the exception of a small pool of inteinational stuuents. 0vei the entiie aggiegate pool, non-
submitteis giauuate at iates fiom 4.8% to 17.6% highei than submitteis, uepenuing on
how the aveiages aie calculateu. (Figuie 19.)

11660
6
1
193
740
939
2178
6522
18990
4166 students
0% 20% 40% 60% 80% 100%
No FAFSA
$70,000 or more
$60,000 - $69,999
$50,000 - $59,999
$40,000 - $49,999
$30,000 - $39,999
$20,000 - $29,999
$10,000 - $19,000
$1 - $9,999
$0
Policy Use at Public Universities by Expected Family Contribution

Page S2 of 7u

Figuie 19.
uiauuation iates also vaiy wiuely between the public univeisities. At each of the six public
univeisities, non-submitteis giauuateu at highei iates than submitteis with lowei BSuPAs,
ianging fiom 9.8% highei to 28.9% highei. (It is woith noting that the aveiages aie skeweu
to be closei by a laige anu highly competitive public univeisity wheie both non-submitteis
anu submitteis giauuate at high iates.) (Figuie 2u.)
6834
6834
13924
13924 students
0% 20% 40% 60% 80% 100%
Institutional Average
Aggregate
Graduation Rate Comparison at Public Universities Between Submitters and Non-Submitters
(graduated cohorts entering 2003, 2004)
The aggregate
graduation rate for the
sample is: 66.2%.

4.8% higher.
The average graduation
rate for the sample is:
51.4%.

17.6% higher
Graduation Rate

Page SS of 7u

Figuie 2u.
As with giauuation iates, non-submitteis eain maikeuly bettei univeisity uPAs, both while
eniolleu (by .18) anu in theii Cum uPAs at giauuation (by .29). (Figuie 21.)

39
174
148
1605
4868
271
1927
5044
1124
5558 students
0% 20% 40% 60% 80% 100%
PU6
PU4
PU5
PU1
PU3
Graduation Rate
Graduation Rate Comparison at Public Universites Between Submitters and Non-Submitters
(graduated cohorts entering 2003, 2004)

Page S4 of 7u

Figuie 21.
The chait below pioviues a compaiison of key uata at each of the six public univeisities in
the stuuy, ieasonably consistent patteins that will help with fuithei analysis. Because of
the institutional policies foi automatic aumission with ceitain BSuPAs anu ianks, non-
submitteis enioll with consistently stiongei BSuPAs. SATs (incluuing conveiteu ACTs)
vaiy: S of 6 have highei SATs fiom non-submitteis, but 2 of the S aie by small uiffeiences,
anu the most competitive institution has highei SATs fiom the submitteis by a small
uiffeience. The only laige uiffeience in SATs is at the institution with compaiatively low
oveiall acauemic ciiteiia, incluuing testing. All six institutions have non-submitteis eaining
consistently highei fiist-yeai anu cumulative uPAs. 0f the five institutions with giauuateu
cohoits, all five have highei giauuation iates foi non-submitteis. (Figuie 22.)

3.22
2.71 2.74
2.82
2.63
2.62
3.56
2.93
2.92
3.45
2.91 2.91
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
High School
GPA
First Year
GPA
Cumulative
GPA
High School
GPA
First Year
GPA
Cumulative
GPA
Policy Use at Public Universities by GPA - High School, College First Year, College Cumulative

Page SS of 7u
Public Universities--Institutional Level Academic Comparison Summary

High School GPA
SAT
(CR + Math)
First Year GPA
Cumulative GPA
(Students entering
in Graduated
Cohorts)
Graduation GPA
(Graduates Only)
Graduation Rate

Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub
PU1 3.68 3.02 1135 1045 3.08 2.61 3.00 2.46 3.23 2.86 70% 54%
PU2 3.74 3.18 1124 1069 3.16 2.58 3.19* 2.66* N/A N/A N/A N/A
PU3 3.90 3.74 1189 1208 3.03 2.81 3.10 2.76 3.24 3.02 86% 76%
PU4 3.40 2.19 1111 1009 2.94 2.39 2.87 2.33 3.25 3.13 54% 25%
PU5 3.46 2.24 1127 1113 2.74 2.04 2.79 2.41 3.20 3.09 64% 50%
PU6 3.20 2.15 998 868 2.85 2.27 2.67 1.96 3.28 3.32 27% 9%
* No graduated cohorts. This number represents enrolled cohorts.
Figuie 22.
In the pool of ovei 71,uuu stuuents, non-submitteis have much bettei BSuPAs, anu almost
exactly the same aveiage testing scoies. As we will see iepeateuly in this stuuy, college anu
univeisity Cum uPA's tiack BSuPAs: non-submitteis entei with BSuPAs that aie .S bettei
than submitteis, anu giauuate with aveiage Cum uPA's that aie .S highei than submitteis.
uiauuation iates of non-submitteis aie S% bettei than submitteis. (Figuie 2S.)

Page S6 of 7u

Figuie 2S.
Bowevei, this oveiall compaiison of non-submittei anu submitteis, while useful, is not the
impoitant snapshot, as it meiges all stuuents aumitteu unuei the BSuPAiank ciiteiia,
which incluue a gieat many stuuents who have both the iequiieu uPA anu testing above
the aveiage foi each institution.
In public univeisities with "acauemic thiesholu" aumissions policies baseu on BSuPA oi BS
iank, it is impoitant to compaie sub-gioups of non-submitteis with submitteis, to isolate in
the laige pools of non-submitteis those stuuents with testing below class aveiages who
might not have been aumitteu if the policy uiu not exist, anu to compaie theii outcomes
with those of submitteis who uiu not meet the BSuPA anu BS iank ciiteiia, anu so weie
aumitteu with theii testing being consiueieu. That analysis follows on the next two chaits.
Non-submitteis in :)#$ the "Above-Aveiage-Testing" anu "Below-Aveiage-Testing"
subgioups eain highei uPAs anu giauuate at highei iates than any of the submittei gioups.
The key finuing is the peifoimance of the "Below-Aveiage-Testing Non-submitteis" who
met the BSuPA ciiteiia but hau lowei-than-aveiage testing foi theii institution. Both in the
aggiegate uata below (Figuie 24.), anu in each of the six public univeisities (Figuie 2S.),
these "Below Aveiage-Testing Non-submitteis" hau highei Cum uPAs anu giauuateu at
highei iates than uiu the submitteis with lowei BSuPAs but highei testing. With the
above-aveiage-testing stuuents iemoveu, the patteins of the iemaining non-submitteis aie

Non-
Submitters
Submitters

n

45395 26330

SAT (See caveat below) 1138 1130 Cohens d
First Year GPA

2.92 2.68 Cohens d

2.92 2.74 Cohens d
Graduation Rate**

68% 63% Chi-Square

First Generation

24% 22% Chi-Square
EFC

$17,547 $17,271 Cohens d
STEM Major

47% 53% Chi-Square

SAT Caveat: 99.3% of Non-Submitters still submitted scores. This data represents that 99.3%.
Public Universities--Summary of Key Statistics
(Students entering 2003, 2004, 2005, 2009, 2010)
COLOR KEY
Cohens D


COLOR KEY
Chi-Square Tests


Page S7 of 7u
in stiong contiast with the submitteis. The non-submitteis entei with stiongei BSuPAs,
weakei testing by ovei 9u points, anu giauuate with highei Cum uPAs anu highei
giauuation iates. As we have founu consistently in the stuuy, anu heavily in this snapshot,
they aie moie likely to be unueiiepiesenteu minoiities, fiist-geneiation-to-college, women
anu Pell iecipients.

Figuie 24.

Non-
Submitters
Submitters

n

19976 26330

First Year GPA

2.76 2.68 Cohens d

2.74 2.74 Cohens d
Graduation Rate** 67% 63% Chi-Square

First Generation

32% 22% Chi-Square
Pell

27% 15% Chi-Square
EFC

$14,825 $17,271 Cohens d
STEM Major

51% 53% Chi-Square

SAT Caveat: 98.9% of Non-Submitters still submitted scores. This data represents that 98.9%.
Public Universities--Summary of Key Statistics without Above-Average-Testing Students
(Students entering 2003, 2004, 2009, 2010)
COLOR KEY
Cohens D

COLOR KEY
Chi-Square Tests

Page S8 of 7u
Public Universities--Summary of Student Segment Statistics
(students entering 2003, 2004, 2009, 2010)

Ns HS GPA
SAT
(CR+Math)
FY GPA
Cum GPA
(Graduates
Only)
Cum GPA
(Students
entering in
Graduated
Cohorts)
Graduation
Rate
First Gen Minority
All Students 71725 3.40 1135 2.83 3.15 2.81 66% 23% 15%

Non-submitters 45395 3.52 1138 2.92 3.23 2.91 68% 24% 15%
Above-Average-
Testing
25419 3.58 1217 3.05 3.33 3.03 69% 18% 7%
Below-Average-
Testing
19976 3.40 1037 2.76 3.09 2.78 67% 32% 24%
Low-Testing 5404 3.33 895 2.49 2.90 2.47 56% 45% 41%

Submitters 26330 3.12 1130 2.68 2.99 2.62 63% 22% 15%
Score Submitters 11526 3.10 1199 2.74 3.04 2.65 61% 16% 9%
Category Admits 14031 3.12 1070 2.63 2.94 2.57 65% 26% 20%
International
Submitter*
773 3.44 1140 2.63 3.28 2.93 72% 8% N/A

Figuie 2S.
Stuuents in the "Low-Testing Non-Submittei" gioup, with SAT aveiages that aie 24u points
below the oveiall cohoit aveiage, giauuate at a S6% iate, S% below Scoie Submitteis with
lowei BSuPAs but SATs that aie Suu points highei. These Low-Testing Non-Submitteis aie
4S% fiist-geneiation anu 41% minoiity, while the Scoie Submitteis aie 16% fiist-
geneiation anu 9% minoiity. Theie is consiueiable vaiiation in the peifoimance of the
Low-Testing Non-Submittei gioups in this stuuy: at foui of the six univeisities, they eain
eithei Cum uPAs oi giauuate at iates equal to oi above the Scoie Submitteis.
In the thiee sub-categoiies above of Non-Submittei gioups (Above-Aveiage-Testing,
Below-Aveiage-Testing, Low-Testing), the peicentages of Fiist-ueneiation-to-College anu
minoiity stuuents aie shaiply inveise to the aveiage SAT scoies in the thiee gioups. But to
iepeat the key compaiison, uespite the majoi uiffeiences in SAT scoies in this chait, the
Below-Aveiage-Testing stuuent s giauuate at highei iates (67%) anu with bettei uPAs
(2.78) than eithei the Scoie Submitteis oi the Categoiy Aumits.
The same inveise pattein is SAT scoies is obseiveu in the Submitteis, but not at the
uiamatic levels of uiffeience. Within the Submittei populations, theie aie only mouest
uiffeiences. Scoie Submitteis who uiu not meet the BSuPARank ciiteiia but can meet the

Page S9 of 7u
testing ciiteiia giauuate with slightly highei Cum uPAs (2.6S) anu lowei giauuation iates
(61%) than the "Categoiy Aumits" who met neithei the uPA noi testing ciiteiia but aie
aumitteu on a holistic ieauing. (Theie is wiue vaiiation by institution, laigely cieateu by
the one laige institution in the stuuy with a veiy stiong stuuent bouy pulling up the
aveiages of this aggiegate cohoit.) But as with othei aspects of this stuuy, it is woith noting
the "N's" in vaiious categoiies. The "Submitteis", setting asiue a ielatively small pool of
inteinational stuuents who peifoim well, iepiesent ovei 2S,uuu stuuents, oi SS.6% of all
the public univeisity stuuents in the stuuy.
In sum, the automatic aumission policies seem to be woiking well, with stiong iesults fiom
stuuents with testing below theii institutional aveiages, anu ieasonable iesults even foi
the stuuents with the lowest SATs. Theie aie wiue institutional vaiiations, but the patteins
seem consistent acioss this quite uiveise gioup of public univeisities. 0n campuses, theie
might be useful policy uiscussions to examine who is in these vaiious categoiies, how they
can be effectively taught, anu whethei aumissions policies ought to be examineu.
///@ =()$ 3(01-('.D*$-)(06 7155$6$9 ,02 ]0()$-9('($9X ;-(0%(<,5 =(02(069
This peei compaiison is baseu on aggiegate cohoits fiom five minoiity-seiving colleges
anu univeisities, a combineu total of 12,691 iecoius. Two univeisities aie public, one uiban
anu the othei in a 0S teiiitoiy. Thiee colleges aie piivate, one uiban anu laige, one uiban
anu small, anu the thiiu a Native-Ameiican college with multiple campuses foi theii wiuely
scatteieu iuial populations. At thiee of the campuses, minoiity stuuents iepiesent ovei
9u% of theii eniollment. But also incluueu in this gioup aie two ielatively laigei uiban
institutions with ioughly 4S% minoiity eniollments, one with a focuseu cuiiiculai
stiength. This college compiises about two-thiius of the iecoius foi this gioup, a statistical
imbalance anu also a cuiiiculai imbalance, with most of its stuuents eniolling foi its
focuseu offeiings. It is faii to say that the five minoiity-seiving institutions in oui stuuy
uiffei fai moie fiom each othei than uo the institutions in the piivate anu public gioupings
pieviously examineu, anu theiefoie we must use caution in uiawing conclusions.
These minoiity-seiving institutions seive a wiue iange of stuuents, fiom those with veiy
stiong acauemic iecoius to otheis whom few woulu have expecteu to puisue a college
uegiee. 0ften the stuuents aie managing multiple issues: ielatively weak high school
backgiounus, thin college counseling, low oi non-existent stanuaiuizeu testing, finances,
ietuining to college as an oluei stuuent (the aveiage enteiing age of these five institutions
is 22.S), jobs oi theii own chiluien. At one institution, the aveiage EFC (Expecteu Family
Contiibution, fiom the FAFSA analysis) is $22S, since viitually all stuuents aie of non-
tiauitional ages, many of them suppoiting families of theii own.
0ne must give substantial cieuit to both these stuuents anu to the institutions which have
focuseu theii mission on seiving them. Theii uPAs anu giauuation iates aie not as high as

Page 4u of 7u
many of the othei institutions in this stuuy, but these minoiity-seiving institutions aie
uoing some of the heaviest of the heavy lifting in highei euucation, if one measuies the
accomplishments of an institution paitly by how substantial aie the changes anu giowth
that stuuents expeiience while eniolleu. Too often, colleges aie juugeu by the cieuentials of
theii enteiing stuuents, with little attention oi unueistanuing given to the to the stuuents'
uevelopeu skills, imagination, uiscipline anu self-confiuence upon giauuation. It is
impoitant that national stuuies like this one incluue some minoiity-seiving institutions in
theii finuings, paitly foi theii ciucially impoitant woik, anu paitly because they aie fai too
often not noticeu in the piestige anu financial compaiisons that uominate most of the
college iankings anu evaluations.
What Shaie of Eniolleu Stuuents 0se an 0ptional Testing Policy.
At these five institutions in aggiegate, 27.S% of eniollees aie non-submitteis.
Who 0ses an 0ptional Testing Policy at these Ninoiity Seiving Institutions.
Foi many stuuents, not submitting testing is moie often a piactical ieality iathei than a
stiategic uecision (the common ieason at the 2u piivate institutions in oui stuuy), oi
eniolling unuei a policy of automatic aumission foi ceitain BSuPA oi class iank (as at the 6
public univeisities). 0nlike the piivate anu public univeisities in the stuuy with ieasonably
consistent appioaches to optional testing, these five minoiity-seiving institutions have
wiuely vaiying appioaches to testing. 0ne institution has a puie optional testing policy, two
collect veiy few SAT oi ACT scoies but use the Accuplacei tests, one uses the SAT to help
ueciue whethei to aumit stuuents to the univeisity oi an inteinal community college, anu
one uses the SAT not foi the aumissions uecision but as a tiiggei to iequiie auuitional
Accuplacei testing. (Figuie 26.)

Page 41 of 7u

Figuie 26.
At these institutions, fiist-geneiation-to-college stuuents, Pell uiant iecipients, all
categoiies of minoiity stuuents anu men geneially enioll at highei iates unuei an optional
testing policy. (Figuie 27.)
201
790
2107
60
336 students
0% 20% 40% 60% 80% 100%
MS3
MS2
MS1
MS5
MS4
Policy Use at Minority-Serving Institutions by Percent of Use
(non-submitters entering 2003, 2004, 2006, 2007, 2009, 2010)

Page 42 of 7u

Figuie 27.
As with the piivate anu public gioups of institutions alieauy piesenteu, at these minoiity-
seiving institutions both low-income anu highei-income stuuents use the policy at highei
iates, cieating a "two-tail" financial pattein. While a non-submittei policy suppoits stuuent
uiveisity anu thus cieates financial aiu obligations, stuuents who aie not ieceiving financial
aiu help maintain balanceu institutional buugets. (Figuie 28.)

1376
65
248
71
206
1449
79 students
0% 20% 40% 60% 80% 100%
White, non-Hispanic
Hispanic
Unknown
Black, non-Hispanic
Non-Resident Alien
Policy Use at Minority-Serving Institutions by Ethnicity and Citizenship

Page 4S of 7u

Figuie 28.
What aie the Acauemic 0utcomes foi Submitteis anu Non-Submitteis.
Submitteis anu non-submitteis entei these institutions with quite uiffeient BSuPAs (S.u
foi submitteis veisus 2.61 foi non-submitteis), anu also a significant uiffeience in SATs
(incluuing conveiteu ACT scoies) of 18S points, though this iepiesents non-submittei
scoies fiom only one institution out of the five. The uiffeiences in BSuPA aie veiy closely
miiioieu in the college anu univeisity uPAs, with submitteis aveiaging 2.66 anu non-
submitteis at 2.S1. (Figuie 29, Figuie Su.)
141
72
876
12
26
33
70
206
1203
855 students
0% 20% 40% 60% 80% 100%
Unknown
Non-Resident Alien
No FAFSA
$50,000 - $59,999
$40,000 - $49,999
$30,000 - $39,999
$20,000 - $29,999
$10,000 - $19,000
$1 - $9,999
$0
Policy Use at Minority-Serving Institutions by Expected Family Contribution

Page 44 of 7u

Figuie 29.

Figuie Su.
3.00
2.76 2.69
2.99
2.77
2.66
2.73
2.52
2.43
2.52
2.51
2.31
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
High School
GPA
First Year
GPA
Cumulative
GPA
High School
GPA
First Year
GPA
Cumulative
GPA
Policy Use at Minority-Serving Institutions by GPA - High School, College First Year, College
Cumulative

Non-
Submitters
Submitters

n

3494 9197

First Year GPA

2.52 2.76 Cohens d

2.43 2.69 Cohens d
Completion Rate** 114% 117% Cohens d

First Generation

42% 40% Chi-Square
EFC

$8,966 $13,634 Cohens d
STEM Major

5% 11% Chi-Square

SAT Caveat: The average for non-submitters represents only one institution that had scores for non-submitters, so it is not an accurate
comparison with submitters across the institutions.

Summary of Key Statistics at Minority-Serving Institutions
(Students entering 2003, 2004, 2006, 2007, 2009, 2010)

COLOR KEY
Cohens D


COLOR KEY
Chi-Square Tests

Page 4S of 7u
Non-submitteis giauuate at iates 1S% below submitteis, 24% veisus S7%. (Figuie S1,
Figuie S2. )

Figuie S1.
1320
1320
515
515 students
0% 20% 40% 60% 80% 100%
Aggregate
Graduation Rate Comparison at Minority-Serving Institutions Between Submitters and Non-
Submitters
(graduated cohorts entering 2003, 2004, 2006, 2007)
The aggregate
sample is: 31.9%.

Submitter rate is 13.2%
higher.
24.9%.

Submitter rate is 11.7%
higher.
Graduation Rate

Page 46 of 7u

Figuie S2.

College uPAs laigely tiack BSuPAs, but also ieflect a bioau iange of iisk factois: weakei
high schools, last-minute applications, less unueistanuing of fielus of stuuy, low EFC's,
oluei stuuents, anu those with jobs oi families of theii own. (Figuie SS.)
Minority-Serving Institutional Level Academic Comparison Summary

High School GPA
SAT
(CR + Math)
First Year GPA
Cumulative GPA
(Students entering
in Graduated
Cohorts)
Graduation GPA
(Graduates Only)
Graduation Rate

Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub Non-Sub Sub
MS1 2.70 2.94
No data
available
991 2.61 2.87 2.55 2.80 3.23 3.29 33% 43%
MS2 2.38 3.17 791 1028 2.33 2.85 1.93 2.70 3.12 3.31 8% 29%
MS3 3.00 3.12
No data
available
792 2.15 2.12 2.39 2.15 3.31 2.95 19% 31%
MS4
No data
available
No data
available
No data
available
No data
available
2.74 2.65 1.77 1.22 3.18 3.14 14% 17%
MS5 1.16 0.94
No data
available
No data
available
1.98 2.41 2.07 2.64 3.43 3.06 17% 29%

272
1
6
178
863
38
37
6
13
421 students
0% 20% 40% 60% 80% 100%
MS2
MS4
MS5
MS3
MS1
Graduation Rate
Graduation Rate Comparison at Minority-Serving Institutions Between Submitters and Non-
Submitters
(graduated cohorts entering 2003, 2004, 2006, 2007)

Page 47 of 7u
Figuie SS.
Baving a wiue vaiiation in both BSuPAs anu testing ovei time can pioviue veiy helpful
infoimation, so in futuie stuuies, we woulu encouiage a focus to gathei moie complete
uata, anu a laigei pool of institutions. As is visible in the chait above, only one institution of
the five coulu submit testing uata. We tiieu to enioll in the stuuy some of the stiongest anu
best-known of the minoiity-seiving colleges anu univeisities, but founu to oui suipiise that
viitually all of them iequiieu testing anu uiu not seem to have consiueieu an optional
testing policy. The bioau finuings in this stuuy anu otheis show minoiity stuuents aie moie
likely to use an optional testing policy if offeieu, anu also that giauuation iates foi minoiity
stuuents iise with the oveiall quality of the institution--that is, theie is little eviuence of
"ovei-matching." So it may be sensible foi moie of the minoiity-seiving institutions to
examine caiefully with inuepenuently conuucteu ieseaich the pieuictive value of testing
veisus othei ciiteiia foi aumission.
Examining the uata in the chaits below, it may not be submittei status that cieates the
uiffeiences in uPAs anu giauuation iates, but the numbeis of stuuents with uiamatically
uiffeient banus of BSuPAs. (Figuie S4.)
Summary of Student Performance at Minority-Serving Institutions by High School GPA and
Testing Score Categories
(students entering 2003, 2004, 2006, 2007, 2009, 2010)
High School GPA
SAT
(CR+Math)
NS Cumulative GPA First-year GPA Graduation Rate
>3.5
>1000 1291 3.44 3.50 56%
800-1000 465 3.15 3.20 47%
<800 58 2.77 2.76 42%

>3.0 & "3.5
>1000 1188 3.08 3.16 46%
800-1000 1140 2.71 2.81 43%
<800 545 2.09 2.17 24%

>2.5 & "3.0
>1000 751 2.69 2.75 35%
800-1000 1145 2.53 2.62 32%
<800 556 2.18 2.33 21%

"2.5
>1000 383 2.27 2.32 29%
800-1000 940 2.16 2.26 21%
<800 730 1.80 1.90 10%

Page 48 of 7u

Figuie S4.
Foi stuuents who peisist to giauuation, Cum uPAs aie geneially soliu: both submitteis anu
non-submitteis at all levels of BSuPA above 2.S achieve bettei than S.u univeisity uPAs.
The issue is giauuation iates, with iates that iange fiom 6S% (eaineu by non-submitteis
with BSuPAs ovei S.S) uown to 1S% (by non-submitteis with BSuPAs unuei 2.S). At least
baseu on this eviuence, the issue is that theie aie so many stuuents at these five
institutions with BSuPAs unuei 2.S: 288S stuuents of the 12,691 total iecoius, oi 2S%,
have high school uPAs unuei 2.S.
In all eight categoiies of BSuPA banus4 foi submitteis anu 4 foi non-submitteisboth
FYuPA anu Cum uPA tiack the BSuPA.
At the high enu of this chait, theie is a cleai pattein with no suipiise: stuuents who entei
with stiong uPAs anu stiong testing will eain stiongei univeisity uPAs anu giauuation
iates. Stuuents with BSuPAs ovei S.S anu also those ovei S.u, anu with SATs in the 8uu-
1uuu iange, get bettei univeisity uPAs anu giauuate at highei iates than stuuents enteiing
with lowei BSuPAs but highei scoies. Stuuents with gieatei than S.S BSuPAs anu scoies
unuei 8uu have moueiate giaues but giauuation iates of 42%, highei than almost all the
gioups on the lauuei below them. Theie aie ieasonably pooi uPAs anu giauuation iates foi
stuuents enteiing with less than a 2.S BSuPA, even with a wiue vaiiation in SATs.
As with othei institutional gioups, foi these minoiity-seiving institutions BSuPA is a
ieasonably consistent pieuictoi of college oi univeisity uPA anu giauuation iates, while
testing seems to be less pieuictive. At least baseu on these quite limiteu uata, stuuents with
BSuPAs ovei S.u anu scoies ovei 8uu seem to have ieasonable chances of success. But
theie is simply not enough uata fiom these institutions to be confiuent about pieuicting
success, except foi the impoitance of BSuPA.
The highei giauuation iates foi submitteis seem to be causeu by a uiamatic imbalance in
the BSuPAs of the two gioups. In fact, both non-submitteis with BSuPAs ovei S.S anu those
ovei S.u giauuate at highei iates than submitteis with paiallel BSuPAs. The highei
giauuation iates foi submitteis aie causeu by much highei numbeis of submitteis in the
top two uPA categoiies, while non-submitteis have much laigei numbeis with BSuPAs
below S.u anu especially below 2.S. Fifty-seven peicent of submitteis have BSuPAs ovei
S.u, while only 29% of non-submitteis aie ovei S.u. In contiast, only 21% of submitteis
have a BSuPA below 2.S, while 4S% of non-submitteis aie below 2.S. (Figuie SS.)

Page 49 of 7u
Summary of Student Segment Statistics at Minority-Serving Institutions
(students entering 2003, 2004, 2006, 2007, 2009, 2010)

Ns
SAT
(CR+Math)
FY GPA
Cum GPA
(Graduates
Only)
Cum GPA
(Students
entering in
Graduated
Cohorts)
Graduation
Rate
First Gen Minority
All Students 12691 960 2.70 3.24 2.53 32% 45% 44%
Submitters 9197 976 2.76 3.25 2.66 37% 45% 41%
>3.5 1808 1101 3.40 3.51 3.27 53% 46% 24%
>3.0 & "3.5 2827 971 2.84 3.22 2.70 40% 45% 42%
>2.5 & "3.0 2324 944 2.61 3.15 2.54 32% 44% 40%
"2.5 1741 893 2.17 2.96 2.20 24% 44% 50%
No HSGPA record 497 903 2.58 3.20 2.41 34% 55% 72%
Non-Submitters 3494 790 2.52 3.23 2.31 24% 44% 5%
>3.5 224 658 3.25 3.44 3.31 65% 21% 23%
>3.0 & "3.5 513 713 2.84 3.38 2.95 43% 32% 39%
>2.5 & "3.0 651 697 2.56 3.17 2.48 30% 41% 45%
"2.5 1142 774 2.10 2.97 1.97 15% 48% 55%
No HSGPA record 964 955 2.58 3.27 2.10 15% 51% 61%

Figuie SS.
Some encouiaging news comes fiom compaiing the shaies of these uPA gioups in the
eniolleu anu giauuateu cohoits. In the giauuateu cohoits, stuuents with BSuPAs unuei 2.S
maue up 2S% of the stuuents, while stuuents ovei S.S maue up only 17%. In the eniolleu
cohoits, about five yeais latei, stuuents with BSuPAs below 2.S hau uioppeu to 2u% of the
class cohoits, anu stuuents ovei S.S hau incieaseu to 21% of the cohoits. So at least at
these institutions, theie is visible piogiess at eniolling stuuents with stiongei high school
iecoius.
A woiking piemisebaseu paitly on these uata anu paitly fiom the comments in
inteiviews with Beans of Aumission anu Biiectois of Institutional Reseaich at these
institutionsis that iaising the shaie of eniolling classes who have stiong high school
iecoius may be the quickest anu most effective way to stiengthen the minoiity-seiving
institutions. A majoi goal foi these colleges anu univeisities, which they aie all woiking to
achieve, is to iaise theii oveiall giauuation iates, anu the cleaiest path to that goal is to
encouiage moie stuuents to achieve in high school.
xii

/^@ >B1 F-'9 /09'('&'(109X ;-(0%(<,5 =(02(069
With only two ait institutes in oui stuuy anu with 78S stuuent iecoius combineu, a full
Aggiegate iepoit was not sensible. We pioviueu foi the two aits institutions a copy of the

Page Su of 7u
Aggiegate Repoit foi the 2u piivate colleges anu univeisities in the stuuy, to give them a
laige (S7,611 stuuent anu alumni iecoius) set of finuings foi institutions that, like the
institutes, aie piivate anu often small.

As one might guess in an ait cuiiiculum with ielatively small eniollments, statistical
analysis of stuuent outcomes uoes not show gieat pieuictive value foi incoming measuies,
incluuing testing. We obseiveu veiy little uiffeience in the iespective uPAs of submitteis
anu non-submitteis of testing at both Institutes. Peihaps a set of Piincipal Finuings ought
to say, "No glaiing finuings." viitually alone among the SS institutions in the stuuy, the ait
institutes have almost no gieen oi uaik gieen shauings inuicating patteins of significant
statistical uiffeiences. The obvious finuing, at one level, is that both submitteis anu non-
submitteis aie succeeuing at paiallel levels, anu that is goou news.

We hau hopeu to finu some pieuictive value (in oui fonuest hopes, a pieuictive holy giail)
in the submitteu poitfolio iatings, but the poitfolio iatings of submitteis anu non-
submitteis weie veiy close to each othei at each institution, as weie the eniolleu Cum
uPAs of stuuents. Peihaps with a laigei uata sample, poitfolio iatings coulu pieuict goou
peifoimance at an ait institute, but we saw no majoi uiffeiences. At one of the institutes,
BSuPAs of submitteis anu non-submitteis weie veiy paiallel, as weie theii Institute uPAs.
At the othei institute, the BSuPAs of non-submitteis weie quite low, but the non-
submitteis eaineu mouestly highei uPAs than the submitteis. We encouiageu the
Aumissions staff at this seconu institute to look thiough the files of non-submitteis, to see if
peihaps these non-submitteis weie the "gambles" the aumissions staff wanteu to take,
feeling the stuuents hau significant piomise uespite theii low BSuPAs. If so, those
"gambles" seemeu to pay off.

With the small N's anu a goou ueal of missing uata (one institute coulu submit eniolleu
cohoits but not giauuateu cohoits), it is haiu to be teiiibly confiuent about the finuings.
But we have listeu in each Institute's iepoit a set of oui tentative finuings, anu suggestions
foi futuie ieseaich. Foi example, one Institute submitteu uata on theii stuuents with
Leaining Biffeiences, anu theie was no coiielation of LB with non-submittei status, a
finuing uiffeient than at the othei piivate colleges anu univeisities which submitteu LB
uata. 0ui hypothesis hau been that peihaps gifteu LB stuuents with spectaculai aitistic
talents woulu use the non-submittei policies to enioll at Ait Institutes. In this micio-
sample fiom one Institute, the eviuence uoes not suppoit the hypothesis. But we encouiage
oui colleagues to continue to exploie LB stuuents who aie gifteu aitists, anu to shaie the
finuings.

At one institute, theie aie some mouest uiffeiences in giauuation iates anu the time-to-
completion iates, with non-submitteis lowei on both. But the ieasons uo not seem to be
the acauemic iecoius of the stuuents, so a topic foi exploiation. Aumissions colleagues at
the ait institutes tolu us that often peisistence to giauuation came fiom the stuuents'
juugments that they weie in the kinu of ait cuiiiculum they wanteu. The compaiatively low
giauuation iates ieflecteu stuuents tiansfeiiing, foi example, fiom a fine aits cuiiiculum to
a uesign cuiiiculum, oi to a laigei univeisity with bioauei offeiings.

Page S1 of 7u
It is woith noting that even with veiy small N's, highei peicentages of non-submitteis came
fiom othei paits of the countiy, a paiallel pattein to othei institutions in oui stuuy, that
optional testing can have a wiue geogiaphic uiaw. It may be that an optional testing policy
is pioving to be an eniollment planning tool, helping these institutions to spieau out
geogiaphically. We also encouiageu the institutes to consiuei inteinational stuuents as an
eniollment oppoitunity, as talenteu inteinational aitists may be in countiies without
suitable ait tiaining. Nany of the piivate colleges have built extiaoiuinaiily fiuitful
connections with inteinational schools, anu the stuuents fiom abioau aie some of the veiy
stiongest on oui campuses.
xiii

>4-$$ *&AD*'&2($9X 3$-(' FB,-29G _D*?&,-$ F0,5.9$9 ,02 [\ *'&2$0'9

F@ =(0,0%(,5 F(2 3$-(' FB,-29 ,02 `<'(10,5 >$9'(06

Nany colleges anu univeisities offei both neeu-baseu financial aiu anu meiit awaius, the
lattei uefineu as financial awaius given in excess of uemonstiateu neeu oi given to
stuuents not iequesting financial aiu. While many meiit awaius go to stuuents who cleaily
woulu not qualify foi neeu-baseu awaius, some go to stuuents who uemonstiate neeu, anu
the meiit awaiu exists on top of the neeu-baseu awaiu, helping to ease family finances.
The iationales behinu meiit awaius aie numeious anu vaiieu: attiacting moie oi stiongei
stuuents with a lowei net piice, iewaiuing vaiious foims of acauemic oi non-acauemic
achievement, oi encouiaging in-state stuuents to attenu a state institution. Theie aie also a
iange of awaius foi paiticulai extiacuiiiculai skills, oi even paiticulai acauemic fielus.

Neiit aiu, while geneially tieu to acauemic piomise, also suppoits eniollment goals by
attiacting stuuents with uesiiable financial piofiles. This use of meiit awaius oi tuition
uiscounting is commonly useu in complex financial matiix systems to inciease tuition
ievenue, by giving a stuuent with a ceitain piofile just enough meiit money to cause them
to enioll, biinging a highei iate of tuition ievenue than the institution woulu otheiwise
ieceive. At the extieme, the college meiit awaiu matiix chait can look iathei like the seat
piicing matiix foi commeicial aiilines, anu function inuepenuently of neeu analysis
altogethei.
xiv

0se of meiit aiu is wiuespieau in oui stuuy: 14 of the 2u piivate colleges anu univeisities
iepoiteu meiit aiu uata, as uiu S of the 6 public univeisities, while 2 of the S minoiity-
seiving institutions iepoiteu meiit aiu uata. 0sage ianges fiom the 12 institutions which
iepoiteu no meiit awaius up to a high of meiit awaius given to 9u% of stuuents in one
piivate institution anu 6S% of stuuents in a public univeisity.

This sub-stuuy will exploie whethei non-submitteis aie moie oi less likely to get meiit
awaius than theii submittei classmates, whethei those meiit awaius weie ieflecteu in
impioveu acauemic peifoimance in teims of Cum uPAs anu giauuation iates, anu whethei
stuuents getting meiit awaius have the same backgiounus as othei stuuents. That is to say,
aie the institutions spenuing theii funus wisely, especially given that the funus by
uefinition aie in excess of stuuents' uemonstiateu neeu.

Page S2 of 7u

0veiall, 27,S11 of the 1u2,76S stuuents in oui stuuy foi whom financial uata was
submitteu ieceiveu a meiit awaiu, oi 26.6%. In geneial, counting all submitteis anu non-
submitteis, the peicentages ieceiving meiit awaius aie quite close. In the giaph below, the
peicentages ieceiving meiit awaius aie eithei 1.4% highei oi S% lowei, uepenuing on
whethei the aveiages aie calculateu fiom meigeu aggiegates oi by aveiaging the
institutional iates, but in eithei case, ieasonably close. (Figuie S6.)

Figuie S6.

But these calculations above incluue all the non-submitteis at public univeisities whose
testing is above theii institutional aveiages. When we iemove the above-aveiage-testing
non-submitteis, as in othei sections of this stuuy, a quite uiffeient pictuie emeiges. The
numbei of non-submitteis with a meiit awaiu uiops piecipitously, fiom 1S,7u8 to Su64.
The shaie of those ieceiving meiit awaius tips stiongly to submitteis, with non-submitteis
ieceiving meiit awaius at iates that aie 9.6% oi 6.S% lowei than submitteis, uepenuing on
how they aie calculateu. (Figuie S7.)

13603
13603
13708
13708 students
0% 20% 40% 60% 80% 100%
Aggregate
All Institutions: Merit Aid Percentage Comparison Between Submitters and Non-Submitters
(merit aid students entering 2003, 2004, 2006, 2007, 2008, 2009, 2010)
The aggregate Non-
Submitter merit aid rate
for the sample is: 27.3%.

1.4% higher.
The average Non-

3.0% lower.
% Receiving Merit Aid

Page SS of 7u

Figuie S7.

The aigument might be maue that the below-aveiage-testing non-submitteis aie less likely
to uo stiong acauemic woik oi giauuate at high iates, anu theiefoie the much highei
numbeis of meiit awaius to the submitteis aie justifieu. We have seen in the othei
poitions of this stuuy that the below-aveiage-testing non-submitteis eaineu Cum uPAs anu
giauuateu at iates that weie geneially as stiong as the submitteis. But those eaining meiit
awaius, about a quaitei of the oveiall pool of stuuents in the stuuy, aie piesumably a
highei calibei of stuuent, anu peihaps in this smallei pool the submitteis will outpeifoim
the below-aveiage-testing non-submitteis.

That is not what the uata seems to tell us, both with all the non-submitteis incluueu in the
fiist analysis below, anu also in a seconu analysis with the above-aveiage-testing non-
submitteis iemoveu. With all the non-submitteis incluueu, the non-submitteis entei the
institutions with viitually iuentical SATACT scoies (12uS foi non-submitteis veisus 1197
foi submitteis), anu BSuPAs that aie .2S highei than the submitteis. The non-submitteis
giauuate with Cum uPAs that aie .22 highei, anu with a giauuation iate that is 11% highei
than the submitteis, 81% veisus 7u%.

Why, one might ask, uo submitteis get as many meiit awaius as non-submitteis. That,
piesumably, involves the iequiiements in many states anu in many piivate institutions foi
ceitain levels of testing to qualify foi a meiit awaiu. But at least fiom these cleai outcomes,
the institutions might want to question whethei the testing iequiiements aie piouucing
13603
13603
5064
5064 students
0% 20% 40% 60% 80% 100%
Aggregate
Merit Aid Percentage Comparison Between Submitters and Non-Submitters
without Above-Average-Testing Non-Submitters
(merit aid students entering 2003, 2004, 2006, 2007, 2008, 2009, 2010)
The aggregate Non-

9.1% lower.
The average Non-

6.5% lower.
% Receiving Merit Aid

Page S4 of 7u
the acauemic iesults that the meiit awaius aie uesigneu foi, given the tight finances of
most institutions. (Figuie S8.)

Figuie S8.

We also examineu the iesults of non-submitteis with the above-aveiage-testing non-
submitteis iemoveu. This smallei set of non-submitteis still has stiongei BSuPAs than the
submitteis, albeit by a statistically tiivial .1, with S.S4 foi non-submitteis veisus S.44 foi
submitteis. Now the laige statistical uiffeience is in SAT scoies, with the submitteis highei
by 14S points. These non-submitteis anu the submitteis giauuate with viitually iuentical
Cum uPAs, but the non-submitteis have giauuation iates that aie 6% highei than the
submitteis, 76% veisus 7u%. So the only statistically laige uiffeiences aie in SAT scoies, in
favoi of the submitteis, anu in giauuation iates, in favoi of the non-submitteis.

It is also woith noting, in the seconu poition of the chait, that theie aie statistically
significant uiffeiences in minoiity status, fiist-geneiation-to-college, genuei, Pell
iecipients, anu STEN majois, with non-submitteis highei in all categoiies. (Figuie S9.)

Non-
Submitters
Submitters

n

13,708 13,603

High School GPA

3.69 3.44 Cohens d
First Year GPA

3.21 3.07 Cohens d

3.19 3.13 Cohens d
Completion Rate** 109% 107% Cohens d

First Generation

20% 18% Chi-Square
Pell

16% 15% Chi-Square
EFC

$21,607 $24,364 Cohens d
STEM Major

51% 44% Chi-Square

SAT Caveat: 90.7% of Non-Submitter submitted scores. The average SAT scores for non-submitters represent that 90.7%. All other data
in this chart is drawn from the full number of students in the chart.

Summary of Key Statistics for Students Receiving Merit Awards
(Students entering 2003, 2004, 2006, 2007, 2008, 2009, 2010)
COLOR KEY
Cohens D

COLOR KEY
Chi-Square Tests


Page SS of 7u

Figuie S9.

The same question must be askeu: in times of tight iesouices, why aie meiit awaius given
in uiamatically highei numbeis to submitteis of testing, uespite slightly lowei Cum uPAs
anu significantly lowei giauuation iates. We woulu hazaiu thiee guesses. Fiist, as a
ciiteiia foi meiit awaius, testing may be an unexamineu anu unfounueu assumption. The
awaius aie given to enteiing fiist yeai stuuents anu ieneweu each yeai with little
examination of outcomes. Seconu, the awaius may be uesigneu to influence guiuebook
iatings that aie paitly uiiven by test scoies. But since the guiuebooks also incluue
giauuation iates in theii iatings, this stiategy may be inteinally woiking against itself, at
gieat expense. Thiiu, meiit awaius may be uesigneu to inciease tuition ievenue by
eniolling stuuents who have the capacity to pay full costs, but will only enioll if the piice is
ieuuceu. It is a complex analysis: uo meiit awaius ieuuce opeiating funus that coulu
otheiwise impiove the quality of eveiything fiom faculty to facilities, oi by eniolling
auuitional stuuents, inciease opeiating funus. In this as in othei facets of this stuuy, we
hope that oui analysis will spui examination of assumptions, at multiple levels fiom
institutions to piofessional associations to testing agencies to state uepaitments of
euucation.

K@ _D*?&,-$ F0,5.9$9

Because many stuuies of college peisistence have useu R-squaie analyses, we also uiu a
baseline R-squaie analysis of the piincipal compaiisons in oui stuuy. 0sing the college

Non-
Submitters
Submitters

n

5064 13,603

First Year GPA

2.99 3.07 Cohens d

2.97 3.13 Cohens d
Graduation Rate**

76% 70% Chi-Square

First Generation

24% 18% Chi-Square
EFC

$22,702 $24,364 Cohens d
STEM Major

49% 44% Chi-Square

SAT Caveat: 74.9% of Non-Submitters submitted scores. The average SAT scores represent that 74.9%. All other data in this chart is
drawn from the full number of students represented in the chart.

Summary of Key Statistics without Above-Average-Testing Non-Submitters
(Students entering 2003, 2004, 2006, 2007, 2008, 2009, 2010)
COLOR KEY
Cohens D


COLOR KEY
Chi-Square Tests

Page S6 of 7u
Cumulative uPA as the uepenuent vaiiable, we conducLed lineai iegiession analyses to
compute R-squaie values. Since we chose to use othei statistical techniques as oui
piincipal tools, we offei these R-squaie analyses as only a supplemental anu quite basic
snapshot, albeit one with a laige "N" of just unuei12S,uuu iecoius acioss a wiue vaiiety of
institutions. 0ui finuings in these R-squaie analyses seem paiallel to foui coie finuings in
Ciossing the Finish Line: Completing College at Ameiica's Public 0niveisties (Bowen,
Chingos & NcPheison): (1.) BSuPA seems to be a much bettei pieuictoi of college
cumulative uPA anu giauuation iates than SATACT scoies, accounting foi all but a few
peicentage points of R-squaie value; (2.) favoiing BSuPA ovei testing will benefit laigei
numbeis of low SES, fiist-geneiation-to-college anu minoiity stuuents, anu Pell iecipients;
(S.) compaiing Cumulative uPAs anu giauuation iates is a moie piouuctive measuie of
college success than fiist-yeai uPAs; anu (4.) theie is little eviuence of "ovei-matching" foi
low SES stuuents, who giauuate at highei iates in piopoition as they enioll at stiongei
institutions.
xv

0ui R-squaie finuings also suppoit the conclusions in Ciossing the Finish Line that testing
seems to have paiticulaily little pieuictive value at public univeisities.
xvi
In oui R-squaie
analyses, BSuPA piouuces the laige shaie of pieuictive value acioss the entiie aggiegate
stuuy, with testing auuing less than S% of R-squaie to BSuPA. Foi ieasons we uo not
pietenu to unueistanu, at the public univeisities in the stuuy, testing auus less than 1.S% of
R-squaie to BSuPA, a statistic that is iemaikably consistent acioss both the aggiegate
gioup anu five sub-gioups. At the piivate colleges anu univeisities, testing auus about 4.S%
of R-squaie to BSuPA, anu with highei R-squaie foi non-submitteis, though this uoes not
line up with oui finuings of closely paiallel giauuation iates anu uPAs of submitteis anu
non-submitteis. At the minoiity-seiving institutions, we founu that testing auus about S%
of R-squaie to BSuPA. (Figuie 4u.)

Summary of R-Square Analysis - Log Variables
Aggregate Data (PR+PU+MSI+AR)

SAT Only HSGPA Only SAT+HSGPA
Aggregate 0.1214 0.2066 0.2358
Aggregate without Above-Average-Testing Students

0.1270 0.2017 0.2435
Non-Submitters 0.1166 0.1993 0.2313
Non-Submitter without Above-Average-Testing Students

0.1353 0.1684 0.2063
Submitters

0.1282 0.2267 0.2553

Private Institutions

Aggregate 0.1262 0.1765 0.2317
Non-Submitters 0.1688 0.1594 0.2598
Submitters

0.1216 0.1830 0.2289

Public Institutions

Aggregate

0.0897 0.2108 0.2247
Aggregate without Above-Average-Testing Students 0.0673 0.1606 0.1751
Non-Submitters

0.1069 0.2149 0.2296

0.1045 0.1379 0.1507
Non-Submitter Low-Testing

0.0633 0.1139 0.1185
Submitters

0.0663 0.1932 0.1997

Minority-Serving Institutions

Aggregate

0.1164 0.1537 0.2065

Page S7 of 7u
Non-Submitters

0.0239 0.1297 0.1151
Submitters

0.1062 0.1461 0.1890

Figuie 4u.

The R-squaie calculations weie peifoimeu in two ways, piouucing mouestly uiffeient
iesults. The giiu above was cieateu using a Log R-squaie analysis, which iemoveu fewei
than 1uuu iecoius out of 12S,uuu with a "zeio" ieauing foi eithei BSuPA oi college Cum
uPA. At the high school level, a zeio BSuPA ieflects home-schooleu stuuents oi those at
non-giauing schools, while at the college level a zeio uPA aie mostly stuuents who
withuiew befoie completing any couises. The giiu below was cieateu using a basic R-
squaie analysis, anu incluues all iecoius. The R-squaie ieauings aie mouestly uiffeient,
but the same patteins aie visible, incluuing lowei pieuictive value foi testing at public
univeisities. (Figuie 41.)

Summary of R-Square Analysis - Standard Variables
Aggregate Data (PR+PU+MSI+AR)

SAT Only HSGPA Only SAT+HSGPA
Aggregate 0.1661 0.2544 0.3042

0.1723 0.2397 0.3088
Non-Submitters 0.1621 0.2534 0.3064

0.1782 0.1959 0.2588
Submitters

0.1743 0.2734 0.3218

Private Institutions

Aggregate 0.1563 0.2086 0.2797
Non-Submitters

0.1920 0.1765 0.2890
Submitters

0.1491 0.2227 0.2776

Public Institutions

Aggregate

0.1332 0.2807 0.3014

0.0987 0.2063 0.2265
Non-Submitters

0.1548 0.2899 0.3126

0.1456 0.1859 0.2008
Non-Submitter Low-Testing

0.091 0.1559 0.1605
Submitters

0.1026 0.2428 0.253

Minority-Serving Institutions

Aggregate

0.1786 0.1847 0.2876
Non-Submitters

0.0456 0.1574 0.1052
Submitters

0.1666 0.1771 0.2699

Figuie 41.

C. [\X *'&2$0'9 B('4 [$,-0(06 \(9,A(5('($9 ,02 `<'(10,5 >$9'(06

This peei compaiison is baseu on an aggiegate cohoit of 1uS9 stuuents anu alumni fiom
eight piivate colleges anu univeisities, seven of them libeial aits institutions anu one aits
institution.

All institutions in the stuuy weie offeieu the oppoitunity to submit uata on LB stuuents
fiom the class cohoits submitteu foi the stuuy. Stuuents hau pioviueu uiagnostic eviuence

Page S8 of 7u
to theii home institutions which alloweu them to be iecognizeu as LB stuuents. We uiu not
ask the institutions foi types of uisabilities, noi weie they shaieu. Baseu on a pievious
stuuy of LB stuuents ovei Su yeais at Bates, it is peihaps ieasonable to assume that the LB
uiagnoses in this sub-stuuy aie a bioau mixtuie of uyslexia, math uisability, ABBB, 0CB,
anxiety, executive oi infoimation piocessing, anu a smallei numbei of physical uisabilities
in mobility, sight oi heaiing.
xvii
The uetails of what kinus of uisabilities the colleges aie
seeing anu effective iemeuiation steps woulu be an excellent follow-up to this small sub-
stuuy.

But again stanuaiuizeu testing becomes an issue: in most yeais, only about 2% of SAT
takeis aie given an accommouation foi any foim of uisability, a figuie that uoes not seem to
have changeu even as schools anu colleges unueistanu LB issues moie thoioughly anu
stuuents aie moie open about seeking accommouations. Among the 2u1S SAT test-takeis,
foi example, S8,184 out of 1,66u,u47 total SAT takeis, oi 2%, weie given
accommouations.
xviii

It is woith noting that, even though the stuuy offeieu to ieimbuise institutions foi any
expenses of gatheiing LB uata, only 8 of the SS institutions in the stuuy submitteu
infoimation, even with the stuuents' iuentities fully piotecteu unuei the same "blinu cioss-
walk" IB numbei as othei uata in the stuuy. In some institutions, LB infoimation is
consiueieu as confiuential meuical infoimation, anu in otheis, the uata is helu by a sepaiate
LB office, ciicumstances that causeu the institution to uecline to shaie it.

In this sub-stuuy, LB stuuents useu optional testing policies at highei iates, appioximately
Su% in this sample veisus the SS% in the aggiegate sample of 2u piivate institutions. So
offeiing an optional testing policy uoes seem to make it moie attiactive foi LB stuuents to
enioll at an institution. Theie was wiue vaiiation in the eight institutions, fiom 76% to u%
of the LB stuuents choosing to be non-submitteis. (Yes, u% is coiiect: at one institution, all
the LB stuuents weie submitteis.)

LB non-submitteis eain uPAs anu giauuate at iates closely paiallel to LB submitteis.
Eniolleu LB non-submitteis eaineu uPAs that weie .u9 highei than the LB submitteis,
while in the giauuateu cohoits the LB submitteis weie .19 highei. LB non-submitteis
eaineu uPAs that weie .19 of a uPA point below othei non-submitteis in the piivate
institutions in the stuuy, anu LB submitteis eaineu uPAs that weie .2S of a uPA point
below othei submitteis. uiauuation iates of LB stuuents weie close to the oveiall
giauuation iate foi the 2u piivate institutions, which is 76.9%. LB non-submitteis
giauuateu at piecisely the same iate, 76.9%, while LB submitteis giauuateu at 7S.6%.
Time-to-completion iates weie iuentical, anu also viitually iuentical with non-LB stuuents,
with less than 1% vaiiance. (Figuies 42 anu 4S.)

Page S9 of 7u

Figuie 42.

Figuie 4S.

Inteiestingly, finuings foi LB stuuents paialleleu those founu moie bioauly in the stuuy.
Theie is the same "two-tail" financial capacity cuive, with if anything, a highei
3.33
2.82
2.84
3.31
2.97
3.08
3.39
2.92
2.95
3.25
2.75
2.89
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
High School
GPA
First Year
GPA
Cumulative
GPA
High School
GPA
First Year
GPA
Cumulative
GPA
Policy Use by LD Students: GPA - High School, College First Year, College Cumulative
(non-submitters entering 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010)
134
134
166
166 students
0% 20% 40% 60% 80% 100%
Aggregate
Graduation Rate Comparison Between LD Submitters and Non-Submitters
(graduated cohorts entering 2003, 2004, 2005, 2006, 2007)
The aggregate
sample is: 75.4%.

3.23% higher.
75.2%.

1.68% higher.
Graduation Rate

Page 6u of 7u
concentiation of LB stuuents not iequesting financial aiu, so that they outnumbei the high-
neeu stuuents by about S-1. Whethei this ieflects a pattein of moie sophisticateu families
unueistanuing LBs, iuentifying colleges wheie LB stuuents succeeu, oi selecting non-
submittei options to inciease chances of aumission, woulu iequiie auuitional ieseaich.

Theie was a mixeu pattein of minoiity stuuents using LB uiagnoses, with non-iesiuent
aliens, Bispanics anu Asians moie likely to be non-submitteis at ioughly 68%, whites (who
heavily uominate numeiically) at 49%, anu Black anu Native Ameiican stuuents much less
likely to be non-submitteis at 2S% anu SS%. 0ne can only speculate on the ieasons foi
these laige gaps, but it is not unieasonable to suggest that uiagnoses of LB issues in
minoiity populations ought to be getting moie attention.
xix

The same statistical tests foi significant uiffeiences in these LB stuuents as we have useu
foi othei cohoits piouuces the same finuing: the statistically laige uiffeience between LB
non-submitteis anu LB submitteis is theii SATACT scoies (anu in this cohoit, genuei, with
moie women). (Figuie 44.)

Figuie 44.

A stuuy at Bates founu that given basic accommouations, LB stuuents uiamatically
impioveu theii uPAs anu giauuation iates ovei a uecaue fiom Su% to 9u%, Bates' oveiall
giauuation iate. In the laige majoiity of cases, the only accommouation neeueu was
auuitional time on couise tests, which is easy to auministei anu inexpensive. As Bates

Non-
Submitters
Submitters

n

531 528

First Year GPA

2.85 2.87 Cohens d

2.95 2.84 Cohens d
Graduation Rate** 76.9% 73.6% Chi-Square
Completion Rate**

102.4% 102.8% Cohens d

First Generation

7% 7% Chi-Square
Gender (Female)

57% 45% Chi-Square
EFC

$26,425 $33,524 Cohens d
STEM Major

18% 26% Chi-Square

SAT Caveat: 33.3% of Non-Submitters submitted scores . This data on average SAT scores of Non-Submitters represents that 33.3%. All
other data in this chart is based on the full number of students in the chart.

Summary of Key Statistics for LD Students
COLOR KEY
Cohens D


COLOR KEY
Chi-Square Tests


Page 61 of 7u
stiengtheneu its auministiation of LB suppoit anu LB stuuents weie moie knowleugeable
anu open about theii LB's, this hau the effect of iaising Bates' oveiall giauuation iate by
seveial peicentage points, not a tiivial change foi faculties, piesiuents anu tiustees.

We believe stiongly that how colleges anu univeisities can effectively seive LB stuuents is
woithy of much moie caieful attention than it has ieceiveu to uate. Acknowleuging the
confiuentiality issues, theie aie significant ieasons to examine the potential foi acauemic
success of LB stuuents acioss institutions.

710%5&9(10X a\$:(0(06 ;-1+(9$b

We began with a question: uoes stanuaiuizeu testing suppoit goou uecision making, oi
uoes it aitificially tiuncate the pools of stuuents who woulu succeeu in college if they weie
given a chance. In this stuuy, we have founu optional testing policies will woik successfully
at bioauly uiffeient kinus of institutions. A cential conclusion: stuuents with stiong
BSuPAs, even without testing, aie likely to succeeu in college, anu stuuents with low
BSuPAs, even with a bioau iange of testing, have much lowei college uPAs anu giauuation
iates.

This might be a miluly inteiesting point of ieseaich, except that the economic health of oui
society will significantly uepenu on how many stuuents uevelop piofessional anu cultuial
skills thiough uemanuing euucations. Theie aie uiamatic choices to be maue: the numbeis
aie quite laige of potential stuuents with stiong BSuPAs who have pioveu themselves to
eveiyone except the testing agencies. Theie is little eviuence of "ovei-matching": even with
lowei scoies anu fiom lowei socio-economic backgiounus, stuuents succeeu at highei
iates at stiongei anu moie competitive institutions wheie they aie given both high
expectations anu suppoit fiom faculty anu staff.

To be blunt, also laige is the shaie of college anu univeisity stuuents who aie cuiiently
aumitteu with low BSuPAs, anu the outcomes foi those eniolleu stuuents aie cleai in lowei
giauuation iates. All stuuents neeu to heai sustaineu, consistent messages fiom theii high
schools anu fiom colleges anu univeisities that the impoitance of a successful high school
iecoiu cannot be oveiestimateu.

We hope that this stuuy will pioviue eviuence of "Befining Piomise," of how institutions
can altei theii aumissions policies to aumit stuuents moie likely to succeeu. We also hope
this stuuy will seive as a piototype foi many stuuies to follow, both within institutions anu
acioss institutional cohoits. As with any caieful scholaily piocess, the ieseaich neeus to be
ietesteu anu uuplicateu in vaiious settings. We hope that colleagues will finu oui woik a
helpful map to examine youi own ioutes to "Befining Piomise."

F%P01B5$26+$0'9

Bill Biss' wise late pieuecessoi as Bean of Aumissions at Bates, Nilton Linuholm, useu to
chuckle about "Linuholm's Nistakes," the stuuents he aumitteu as gambles foi one ieason

Page 62 of 7u
oi anothei. But Linuholm's Nistakes went on to piactice law oi meuicine, founu
companies, leau school systems anu colleges, anu uo useful ieseaich. Peihaps Bill's long
commitment to ieseaich on optional testing is anothei of Linuholm's Nistakes, an acoin
that uiun't fall fai fiom the tiee. We offei a fonu bow to his memoiy anu wisuom, anu to his
example as a woiking uean who was also an in-the-tienches ethicist.

The authois of this stuuy acknowleuge with giatituue the geneious multiyeai financial
suppoit foi the ieseaich fiom a piivate founuation. We also expiess giatituue to oui foui
Bates stuuent ieseaicheis--Sophy Nin '11, Qasim Nahmoou '1S, K.}. Lee '14 anu Awais Raza
'1S, anu to seveial Bates faculty anu Institutional Reseaich officeis who weie theii
mentois, always available to help with questions fiom this stuuy. In paiticulai, we thank
two colleagues who pioviueu invaluable suppoit when co-authoi anu leau ieseaichei
valeiie Fianks was committeu with a sick family membei. Piofessoi Nichael Saigent, who
teaches Psychology couises in statistics anu ieseaich methouologies at Bates, peifoimeu a
veiy thoiough statistical auuit of the final sections of the stuuy. Ellen Peteis, Biiectoi of
Institutional Reseaich anu Retention at the 0niveisity of Puget Sounu anu foimeily
Biiectoi of Institutional Reseaich at Bates, gave the papei a helpful euitoiial ieauing as an
IR uiiectoi who woikeu foi many yeais with oui ieseaich on optional testing.

We also expiess oui thanks to the scholais, institutional ieseaicheis anu aumissions
piofessionals who seiveu as membeis of oui stuuy's Auvisoiy Committee oi auuitional
auvisois foi the stuuy. The Auvisoiy Committee membeis functioneu as the "peei
ievieweis" foi the stuuy, commenting on the ieseaich methouology anu analysis uesign in
the eaily stages, iesponuing to multiple uiafts of sections of the stuuy ovei the yeais of the
stuuy, anu pioviuing veiy helpful comments on the final uiaft.

F2)(91-. 71++(''$$ 3$+A$-9X
Baviu Bawkins, Biiectoi of Reseaich anu Policy, National Association foi College Aumission
Counseling
uaiy Pike, Associate Piofessoi of Bighei Euucation anu Executive Biiectoi of Infoimation
Nanagement anu Institutional Reseaich. Inuiana 0niveisity Puiuue 0niveisity
}ohn Fiaiie, vice Piesiuent foi Stuuent Affaiis & Eniollment, Washington State 0niveisity
Naiy Baiiington, Biiectoi of Institutional Reseaich anu Assessment, 0niveisity of
Nississippi
Cate Rowan, Biiectoi, Institutional Reseaich & Euucational Assessment, Smith College
Alisa Cunningham, foimei vice Piesiuent of Reseaich anu Piogiams, Institute foi Bighei
Euucation Policy
Kate Boiia, Bata Reseaich anu Web Tiaffic Analyst, Amheist College
Kaieem Aziz, Biiectoi of Institutional Reseaich anu Planning, Sojouinei-Bouglass College

F22('(10,5 F2)(9$-9X
Nike }ohnson, Biiectoi, Institutional Reseaich, Bickinson College
ueialu Bizinno, Associate Professor, Department of Educational Leadership and Policy Studies,
and foimei Assoc. vice Piesiuent, Institutional Reseaich, 0niveisity of Texas, San Antonio
Royal Bawson, Asst. vice Piesiuent foi Institutional Effectiveness, Columbia College of
Chicago

Page 6S of 7u
Tom Noitenson, Euitoi anu Publishei, Postseconuaiy Euucation 0ppoitunity anu Senlor
Scholar, 1he ell lnsLlLuLe for Lhe SLudy of CpporLunlLy ln Plgher LducaLlon
}oseph A. Soaies, Piofessoi of Sociology, Wake Foiest 0niveisity
Kevin Rask, 0niveisity Reseaich Piofessoi of Economics, Coloiauo College
Naiylyn Scott, Associate Bean of Aumissions foi Nulticultuial Reciuitment, Bucknell
0niveisity
Lisa Plimpton, Biiectoi of Reseaich, Senatoi ueoige }. Nitchell Scholaiship Reseaich
Institute.

The Authois

William C. Biss, Piincipal Investigatoi, eaineu a B.A. in English at Bates, a N.T.S. in ethics at
Baivaiu Bivinity School, anu a N.A. anu Ph.B. in Ameiican liteiatuie anu ieligion at Tufts.
Bill was the Bean oi vP in Bates' aumissions anu financial aiu offices fiom 1978 to 2uuu,
anu has leu the ieseaich on Bates' optional testing policy foi ovei thiity yeais. Be seiveu
foi six yeais as an appointee fiom the 0. S. Senate to the Feueial Auvisoiy Committee on
Stuuent Financial Assistance, anu leu foi a uecaue the auvisoiy committee of college
aumissions ueans at 0.S.News. Foi seveial yeais he taught a Fiist Yeai Seminai in Asian
Stuuies at Bates, stuuying liteiatuie anu film in five countiies that expeiienceu 2u
th
centuiy
cataclysms.

valeiie Fianks, Co-authoi anu Leau Reseaichei, eaineu hei B.A. at Bates in Psychology, anu
foi seven yeais afteiwaius in Euiope, New Yoik anu Naine peifoimeu maiket ieseaich foi
tiansfoimational business anu stiategy initiatives. She seiveu as an Assistant Bean of
Aumissions at Bates, anu in 2uu8 founueu hei fiim as a ieseaich consultant to college
institutional ieseaich offices, uesigning anu peifoiming qualitative anu quantitative
ieseaich stuuies.

0ui foui stuuent ieseaicheis have peifoimeu analyses unuei the uiiection of the stuuy co-
authois, anu with fiequent auvice anu suppoit fiom theii Bates faculty.

Khin "Sophy" Nin '12 giauuateu with uouble majois in Economics anu Stuuio Ait,
anu a minoi in Nathematics. She seiveu foi 18 months as oui piincipal stuuent
ieseaichei, until she accepteu a position with NcKinsey. She has leu touis fiom hei
native Buima to }apan anu Koiea, anu uone uiban planning ieseaich in 0hio.
Kyoung-}une Lee'14 will giauuate with uouble majois in Economics anu
Nathematics anu a minoi in }apanese. K} eaineu his high school uiploma in
Switzeilanu, anu took two yeais off fiom Bates foi iequiieu Koiean militaiy seivice.
Aftei a summei inteinship with Analysis uioup in Boston, he has accepteu a
position to begin with the fiim aftei his Bates giauuation.
Qasim Nahmoou '1S giauuateu with uouble majois in Economics anu Nathematics.
With two summeis of woik at Baiclays Bank in Kaiachi anu New Yoik anu a
summei of ieseaich foi a Bates Economics piofessoi in housing maikets, he
accepteu a position with Baiclays in New Yoik following giauuation.

Page 64 of 7u
uhulam Awais Rana '1S is a majoi in Economics with a minoi in Nath, with inteiests
in inteinational politics anu economics. Be is stuuying at 0xfoiu 0niveisity foi his
}unioi yeai.

Enunotes:

l
National Association foi College Aumission Counseling, Septembei, 2uu8.
http:www.nacacnet.oigieseaichieseaich-
uataBocumentsTestingComission_FinalRepoit.puf, 7,1u,21.

ll
National Association foi College Aumission Counseling, Septembei, 2uu8.
uataBocumentsTestingComission_FinalRepoit.puf, 7,1u,21.

lll
William Biss anu Kate Boiia, "Befining Piomise: Twenty-five Yeais of 0ptional Testing at
Bates College: 1984-2uu9". " Papei given at Confeience: Biamonus in the Rough: Why
Eveiy Institution of Bighei Leaining Will Want to Attiact anu Aumit Byslexic Stuuents,
Stanfoiu Business School, }une 4, 2u11. http:www.bates.euuaumissionoptional-
testing

lv
Seveial impoitant iecent books anu stuuies have examineu how testing can shape college
aumissions ciiteiia: William u. Bowen anu Beiek Bok, K$2 J$&@2 )6 #$2 L"<27D M)38 K27(
G)3*2N02392* )6 G)3*"127"38 L&92 "3 G)''282 &31 O3"<27*"#5 ?1("**")3* (Piinceton: Piinceton
0niveisity Piess, 1998), William C. Bowen, Natthew N. Chingos & Nichael S. NcPheison,
G7)**"38 #$2 P"3"*$ M"32D G)(@'2#"38 G)''282 &# ?(27"9&Q* >0:'"9 O3"<27*"#"2* (Piinceton:
Piinceton 0niveisity Piess, 2uu9), "Repoit on the Commission on the Stanuaiuizeu Tests in
0nueigiauuate Aumissions" (NACAC, 2uu8), anu most iecently }oseph A Soaies, Eu. J?K
!&7*D K$2 G&*2 6)7 K2*# R@#")3&' ?1("**")3* (New Yoik: Teacheis College Piess, Columbia
0niveisity, 2u12), an aumiiable collection of essays on optional testing fiom vaiious
ieseaicheis, thinkeis anu aumissions piofessionals. Also ielateu is a iecent stuuy on the
low peicentages of high talent-low income stuuents who apply to oui selective institutions:
Caioline Boxby anu Chiistophei Aveiy, "The Nissing '0ne-0ffs: The Biuuen Supply of Bigh-
Achieving, Low Income Stuuents" ( National Buieau of Economic Reseaich, 2u12). Like
many college aumissions ueans, we offei a ueep bow of appieciation to Bowaiu uaiunei
anu Robeit Steinbeig foi theii seminal libeiating woik on multiple intelligences anu goou
woik, anu to }onathan Kozol foi his books on the vastly uiffeient euucational expeiiences of
the iich anu pooi in Ameiica. Foi Bill Biss as an aumissions uean, uaiunei anu Kozol have
been both louestones anu Noith stais.
v
No single ieseaich papei shoulu pietenu to offei inteinational solutions to the complex
issues connecteu to stanuaiuizeu testing. But inteiestingly, theie have been
communications about this stuuy with euucational leaueis, ueans, college counselois, anu
jouinalists fiom seveial countiies which have contentious national entiance examinations:

Page 6S of 7u

South Koiea, }apan, China, vietnam, Inuia anu Isiael. They, like we, aie tiying to
unueistanu how to maximize the euucational potential of theii people.

vi
0ne of the iegulai questions fiom institutions in the stuuy was how effectively anu
ethically to collect scoies fiom non-submitteis. Suiely this woulu be a useful topic foi an
NACAC }ouinal aiticle oi as pait of a National Confeience panel.

vll
At some levels, oui gioupings aie juugment calls: one of oui minoiity-seiving institutions
is piivate, but laigei than some of the public univeisities, anu with a focuseu aits
cuiiiculum, so it coulu have been counteu in thiee of oui foui categoiies. We have pioviueu
some institutions in the stuuy with multiple aggiegate uata analyses when they fell into
multiple categoiies.

vlll
As mentioneu in the text, ieaueis shoulu not attach too much piecision to these
scatteiplot giaphs. Theie aie appioximately 88,Suu stuuent iecoius in the fiist giaph anu
1uS,Suu iecoius in the seconu. All available uata was useu, so the N counts foi the two
giaphs aie uiffeient, laigely uue to laigei numbeis of stuuents foi whom colleges uiu not
ieceive oi iecoiu a final BSuPA, oi foi whom theie was no college Cumulative uPA. To offei
a ieauable foimat, on each giaph theie aie Su uots foi submitteis, Su foi above-aveiage-
testing non-submitteis anu Su foi below-aveiage-testing non-submitteis. The fiist giaph
was geneiateu by aggiegating all stuuent uata with BSuPA anu College Cum uPA. Each
stuuent pool was uiviueu into Su gioups, with a computeu aveiage BSuPA anu Cum uPA
foi each gioup.

The seconu scatteiplot giaph, with SAT scoies anu college cumulative uPA, was geneiateu
in the same way, by aggiegating all stuuent uata with SAT anu College Cum uPA. Theiefoie,
theie aie Su uata points foi Submitteis anu Su each foi above-aveiage-testing non-
submitteis anu below-aveiage-testing non-submitteis.

lx
Biss anu Boiia, "Befining Piomise: Twenty-five Yeais of 0ptional Testing at Bates College:
1984-2uu9."

x
0ne piivate institution pioviueu six class cohoits of uata insteau of foui. Foi puiposes of
consistency, in calculating the finuings foi the aggiegate analysis of the 2u piivate
institutions, we useu foui class cohoits fiom this institution, so baseu oui analysis on
S6,8S9 iecoius insteau of S7,611.

xl
}onathan Kozol, J&<&82 S32N0&'"#"2*D G$"'1723 "3 ?(27"9&Q* J9$))'* (New Yoik: Baipei
Peiennial, 1992), 2S6-2S7.

xll
Foi peihaps unueistanuable ieasons, examination of stanuaiuizeu testing policy has not
been a top piioiity at many minoiity-seiving institutions. uiven that testing geneially uoes
not woik in favoi of unueiiepiesenteu minoiity stuuents in competition foi aumission, we
in ieciuiting institutions foi the stuuy weie suipiiseu to finu that most well-known

Page 66 of 7u

minoiity-seiving institutions iequiie testing anu uon't seem to have consiueieu a test-
optional policy. So it will be impoitant to examine the pieuictive value of both the SATACT
anu Accuplacei tests. Who shoulu peifoim this ieseaich is not a tiivial issue. The CEEB has
pioviueu evaluation iepoits on the pieuictive value of theii testing, anu they have
enoimous histoiical files foi compaiison. But with the College Boaiu committeu to
ieuesigning the SATs, it is not cleai how the vaiious pieuictive stuuies will be affecteu. A
numbei of inuepenuent stuuies point to BSuPA as a veiy stiong pieuictoi of college
success, with SATACT testing as a minoi factoi. See, foi example, Bowen, Chingos, anu
NcPheison, G7)**"38 #$2 P"3"*$ M"32, anu Soaies, Eu., J?K !&7*B Theie may be wisuom in
using the statistical methouologies in these volumes to uesign in-house stuuies, oi contiact
with inuepenuent highei euucation ieseaich fiims foi the stuuies.
xlll
All foui of the stuuent ieseaicheis who have woikeu on this stuuy aie inteinational
stuuents at Bates. They aie fiom Buima, South Koiea anu Pakistan, geneially uouble
majois in Economics anu Nathematics with poweiful statistical anu analytic skills, anu
each able to woik in foui languages. See the biogiaphical notes at the conclusion of this
papei.

xiv
Natthew Quiik, "The Best Class Noney Can Buy: The Rise of the 'Eniollment Nanagei'
anu the Cutthioat Quest foi Competitive Auvantage. The Seciet Weapon: Financial-Aiu
Leveiaging." The Atlantic. 444B#$2&#'&3#"9B9)(T(&8&U"32TBBB"#$%;&'($$;)*+#,;&(+;
"-,T-/V-/WT Nov 1, 2005

xv
Bowen,. Chingos, anu. NcPheison, G7)**"38 #$2 P"3"*$ M"32D G)(@'2#"38 G)''282 &#
?(27"9&Q* >0:'"9 O3"<27*"#"2* (Piinceton: Piinceton 0niveisity Piess, 2uu9). Foi BSuPA as a
stiongei pieuictoi than SATACT scoies, see pp. 4S, 7u, 11S, 117,122-124. Foi BSuPA as a
moie accuiate anu auvantageous cieuential foi low SES stuuents, pp. 116, 127. Foi foui-
yeai Cum uPA anu giauuation iate as a moie accuiate measuie of college achievement than
college fiist-yeai uPA, p. 11S. Foi a lack of eviuence foi "ovei-matching" in college
selection, pp. 1uu, 198, 2u9, 214. G7)**"38 #$2 P"3"*$ M"32 (p. 114) founu that when contiols
weie auueu foi the quality of high schools, the pieuictive powei of SATACT testing
uisappeaieu, anu often hau a negative coiielation with college peifoimance. 0ui stuuy uiu
not attempt any analysis of the quality of high schools, but in oui iepoits back to the
institutions in oui stuuy, we often iecommenueu that the college oi univeisity conuuct
theii own stuuy of the pieuictive powei of testing, to test both foi potential "false
negatives" among low income anu fiist-geneiation-to-college stuuents, anu potential "false
positives" among high income stuuents, who hau iaiseu theii test scoies thiough intensive
coaching, but peihaps not theii college giaues anu giauuation iates.

xvl
Bowen, Chingos anu NcPheison. G7)**"38 #$2 P"3"*$ M"32, 114-116.

xvll
Biss anu Boiia. "Befining Piomise: Twenty-five Yeais of 0ptional Testing at Bates
College: 1984-2uu9."

Page 67 of 7u

xvlll
http:meuia.collegeboaiu.comuigitalSeivicespufieseaich2u1STotaluioup-
2u1S.puf, 2.

xlx
!onaLhon Kozol, ?(&U"38 =7&92D K$2 M"<2* )6 G$"'1723 &31 #$2 G)3*9"2392 )6 & X&#")3 (New
Yoik: Ciown Publisheis, 199S), pp 7-8, 133-136, 170-173. Bill Biss began his teaching caieei
in the Noiiisania section of the South Bionx, assigneu to teach 7
th
anu 9
th
giaue
Nathematics in legal violation of his NYC license to teach English. Be has nevei foigotten
}onathan Kozol's uiscussion in Amazing uiace of the high concentiations of leau paint in
public housing anu public schools, anu of NYC public incineiatois in South Bionx
neighboihoous without the political oi financial powei to iesist the placement of these
incineiatois, both of which contiibute to veiy high iates of asthma anu leaining uisabilities
in chiluien in these neighboihoous.

K(A5(16-,<4.

Auelman, C. "Answeis in the Tool Box: Acauemic Intensity, Attenuance patteins, anu
Bacheloi's Begiee AttainmentB" Washington, BC: 0.S. Bepaitment of Euucation, }une, 1999.
Retiieveu }uly 12, 2uu9, fiom http:www.eu.govpubsToolboxtoolbox.html.

Bowen, William u., Natthew N. Chingos, Nichael S. NcPheison. G7)**"38 #$2 P"3"*$ M"32D
G)(@'2#"38 G)''282 &# ?(27"9&Q* >0:'"9 O3"<27*"#"2*B >7"392#)3. XYD >iinceton 0niveisity
Piess, 2uu9.

Bowen, William C., anu Beiek Bok. K$2 J$&@2 )6 #$2 L"<27D M)38 K27( G)3*2N02392* )6
G)3*"127"38 L&92 "3 G)''282 &31 O3"<27*"#5 ?1("**")3*. Piinceton, N}: Piinceton 0niveisity
Piess, 1998.

Cambiano, R.L., u.S. Benny anu }.B. Be voie, (2uuu). "College Stuuent Retention at a
Niuwestein 0niveisity: A Six-Yeai Stuuy." K$2 Y)073&' )6 G)''282 ?1("**")3*. No. 166
(Wintei 2uuu): 22-9u.

College Entiance Examination Boaiu, "2u1S College Bounu Seniois: Total uioup Piofile
Repoit." http:meuia.collegeboaiu.comuigitalSeivicespufieseaich2u1STotaluioup-
2u1S.puf
FaiiTest: The National Centei foi Faii anu 0pen Testing. "SATACT 0ptional 4-Yeai
0niveisities, Test Scoie 0ptional List", http:www.faiitest.oiguniveisityoptional

uaiunei, BowaiuB P7&(2* )6 Z"31D K$2 K$2)75 )6 Z0'#"@'2 S3#2''"82392*. New Yoik: Basic,
198S.

uaiunei, BowaiuB Z0'#"@'2 S3#2''"82392*D X24 I)7"U)3*. New Yoik: Peiseus, 2uu6.

Page 68 of 7u

Biss, William C., Elizabeth A. BeLaBunt, Richaiu W. Noll, Zina Zacque. "Aumissions as
Ninistiy." Closing Confeience Session, NACAC National Confeience, Cincinnati, 0B, 198S.
Publisheu in Y)073&' )6 G)''282 ?1("**")3*, Wintei, 1986, 16-24.

Biss, William C., Kate Boiia. "Befining Piomise: Twenty-five Yeais of 0ptional Testing at
Bates College: 1984-2uu9." Papei piesenteu at confeience, Biamonus in the Rough: Why
Eveiy Institution of Bighei Leaining Will Want to Attiact anu Aumit Byslexic Stuuents,
Stanfoiu, CA: Stanfoiu 0niveisity Business School, }une 4, 2u11.
www.bates.euuaumissionfiles2u14u12Sth-yeai-SAT-iepoit-Stanfoiu-6.S.11-wch.ppt.

Biss, William C., Kate Boiia. "Rethinking Piomise: Twenty-five Yeais of 0ptional Testing at
Bates College, 1984-2uu9." Papei piesenteu at confeience, Biamonus in the Rough: Why
Eveiy Institution of Bighei Leaining Neeus to Know about Byslexic Stuuents, New Baven,
CT: Yale Centei foi Byslexia anu Cieativity, Yale Neuical School, Yale 0niveisity, Nay 18,
2u1u.

Biss, William C. anu Piem R. Neupane . "2u Yeais of 0ptional SATs at Bates, (1984-2uu4)",
papei piesenteu at NACAC National Confeience, Nilwaukee, WI, Sept., 2uu4.

Biss, William C. anu Piem R. Neupane. "What Bave We Leaineu fiom 2S Yeais of 0ptional
SAT's at Bates College." Papei piesenteu at confeience, Yale Confeience on Byslexia anu
Cieativity, New Baven, CT: Yale Centei foi Byslexia anu Cieativity, Yale Neuical School,
Yale 0niveisity, }une 24, 2uu9

Boxby, Caioline anu Chiistophei Aveiy . "The Nissing '0ne-0ffs: The Biuuen Supply of
Bigh-Achieving, Low Income Stuuents", National Buieau of Economic Reseaich, 2u12.
444B3:27B)78T@&@27*T4+[\[H , oi
www.biookings.euu~meuiapiojectsbpeaspiing...2u1Sa_hoxby.puf

Buntei, }.u., W. Samtei. "A College Aumission Test Piotocol to Nitigate the Effects of False
Negative SAT Scoies." K$2 Y)073&' )6 G)''282 ?1("**")3*. No. 168 (Summei 2uuu): 22-9u.

Kiiby, E., S. White, anu N. Aiuguete, N . "Pieuictois of White anu Ninoiity Stuuent Success
at a Piivate Women's College." G)''282 J#0123# Y)073&'. vol. 41 Issue 2 (}une 2uu7): 46u-
46S.

Kozol, }onathon. ?(&U"38 =7&92D K$2 M"<2* )6 G$"'1723 &31 #$2 G)3*9"2392 )6 & X&#")3@ New
Yoik: Ciown Publisheis, 199S

Kozol, }onathon. J&<&82 S32N0&'"#"2*. New Yoik: Baipei Peiennial, 1991

Lemann, Nicholas. K$2 C"8 K2*#D K$2 J2972# I"*#)75 )6 #$2 ?(27"9&3 Z27"#)97&95. New Yoik:
Faiiai, Stiaus anu uiioux, 1999.

Page 69 of 7u

Nessick, S. "valiuity of Psychological Assessment." ?(27"9&3 >*59$)')8"*#. \/ (1996), 741-
749.

National Association of College Aumission Counseling. "Repoit of the Commission on the
0se of Stanuaiuizeu Tests in 0nueigiauuate Aumission." NACAC, 2uu8.
uataBocumentsTestingComission_FinalRepoit.puf

Naumann, W.C., B. Banualos anu T.B uutkin,. "Iuentifying vaiiables That Pieuict College
Success foi Fiist-ueneiation College Stuuents." K$2 Y)073&' )6 G)''282 ?1("**")3, n181 (Fall
2uuS), pp. 4-9.

Patteison, Biian F. anu Kiista B. Nattein. "valiuity of the SAT foi Pieuicting Fiist Yeai
uiaues: 2u1u SAT valiuity Sample." The College Entiance Examination Boaiu, publisheu as
PBF, Apiil 1u, 2u1S. Statisticaliepoit-2u1S-2-valiuity-sat-1
st
-yi-gpa-2u1u-sample-puf.
http:ieseaich.collegeboaiu.oig.

Quiik, Natthew. "The Best Class Noney Can Buy: The Rise of the 'Eniollment Nanagei' anu
the Cutthioat Quest foi Competitive Auvantage. The Seciet Weapon: Financial-Aiu
Leveiaging." The Atlantic. 444B#$2&#'&3#"9B9)(T(&8&U"32TBBB"#$%;&'($$;)*+#,;&(+;
"-,T-/V-/WT Nov 1, 2005

Rowen, Cate anu Nichael Robinson. "Nt. Bolyoke College's Expeiience with 0ptional
Submission of SAT Scoies: An Analysis of the Fiist Foui Yeais." 0npublisheu, Nt. Bolyoke
College Institutional Reseaich 0ffice. 2uu6.

Sackett, Paul R., Nathan R. Kuncel, Auam S. Beatty, }ana L. Riugon, Winny Shen anu Thomas
Kigei. "The Role of Socioeconomic Status in SAT-uiaue Relationships anu in College
Aumissions Becisions." >*59$)')8"9&' J9"2392, publisheu online 2 August 2u12. Accesseu
August 28, 2u12. http:ux.uoi.oig1u.1177u9S67976124S87S2.

Sacks, Petei. J#&31&71"U21 Z"31*D K$2 I"8$ >7"92 )6 ?(27"9&Q* K2*#"38 G0'#072 &31 !$&# !2
G&3 F) K) G$&382 S#B New Yoik: Ba Capo Piess, 1999

Soaies, }oseph A., Eu. J?K !&7*D K$2 G&*2 6)7 K2*#;R@#")3&' G)''282 ?1("**")3*. New Yoik:
Teacheis College Piess. 2u12.

Steinbeig, }acques. K$2 =]22@27*D S3*"12 #$2 ?1("**")3* >7)92** )6 & >72("27 G)''282.
New Yoik: viking, 2uu2.

Teienzini, P. T., L. Spiingei, P.N. Yaegei, E. T. Pascaiella, anu A. Noia. "Fiist-ueneiation
College Stuuents: Chaiacteiistics, Expeiiences, anu Cognitive Bevelopment." L2*2&79$ "3
I"8$27 ^109&#")3 S7 (1996): 1-22.

Page 7u of 7u

Ting, S. N. "Pieuicting Fiist-Yeai uiaues anu Acauemic Piogiess of College Stuuents of
Fiist-ueneiation anu Low-Income Families." Y)073&' )6 G)''282 ?1("**")3. +\[ (1998): 14-
2S.

William C. Biss
Piincipal Investigatoi
"Befining Piomise: 0ptional Stanuaiuizeu Testing Policies in Ameiican College anu
0niveisity Aumissions 0ffices"
26 Baufielu Roau
Ninot. NE u42S8

CE: 2u7-S76-4497
PB: 2u7-784-7726
whissbates.euu

The Value of Entrance Exams in Predicting College Success

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Value of Entrance Exams in Predicting College Success

Uploaded by

Copyright:

Available Formats

BEFININu PR0NISE: 0PTI0NAL STANBARBIZEB TESTINu P0LICIES IN ANERICAN C0LLEuE

ANB 0NIvERSITY ABNISSI0NS

You might also like