Assessing Academic Content

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

The Journal of Special Education

Assessing Students With 46(2) 117­–127


© Hammill Institute on Disabilities 2012
Reprints and permission: http://www.
Significant Cognitive Disabilities sagepub.com/journalsPermissions.nav
DOI: 10.1177/0022466910379156

on Academic Content http://journalofspecialeducation


.sagepub.com

Jessica Goldstein1 and Peter Behuniak1

Abstract
Federal and state policies require all students with significant cognitive disabilities to participate in state assessments and
be included in measures of adequate yearly progress. Although these alternate assessments of grade-level content based
on alternate achievement standards have been in place for several years, little is known about the knowledge and skills
of students with significant cognitive disabilities or about the influence of these assessments on student learning. This
study uses an empirically evaluated classification scheme of data from one cohort of students’ participation in a statewide
alternate assessment, Connecticut’s Skills Checklist, to validate claims from teacher focus group interviews. Data from both
components of the study suggest that nearly half of the teachers found the content on the Skills Checklist inaccessible for
their students. Implications of this finding are discussed.

Keywords
validity, assessment, accountability, significant cognitive disabilities

Federal policy requires that all students with significant their AA-AAAS. By 2008, 15 states were still working on
cognitive disabilities participate in state assessments and be modifying their assessments to meet federal standards (U.S.
included in measures of adequate yearly progress (U.S. Department of Education, 2009). In a national study of
Department of Education, 2003). To meet this requirement, alternate assessment practices, respondents from every state
states developed alternate assessments for this population reported that the student’s special education teacher admin-
based on alternate academic achievement standards (AA- istered or assembled the alternate assessments (Cameto et al.,
AAAS) that are aligned to grade-level academic content 2009). Cameto et al. (2009) also reported that the student’s
standards. Though students with significant cognitive dis- classroom teacher was allowed to score the alternate assess-
abilities represent only approximately 1% of the general ment in 27 states. These scoring procedures are heavily
population (U.S. Department of Education, 2005), the dependent on disposition of the teacher (Quenemoen, 2008).
diversity of these students’ communication abilities and Task design and scoring place special education teachers at
learning challenges is quite complex. Developing standard- the center of accountability decisions for their students and
ized assessments for this diverse group of students requires have a significant influence on the psychometric quality of
a balance between the standardization required for large- the assessment results (Elliott & Roach, 2007).
scale assessments and the flexibility needed for the class- As the measurement field grapples with improving the
room teacher to successfully work with this population. technical quality of alternate assessments, an effort is
Although a body of literature on technical issues related to also under way to better understand the nature and needs
the development of AA-AAAS is slowly emerging (Browder of the student population targeted by these assessments.
et al., 2004; Gong & Marion, 2006; Perie, 2007), few stud- Results from one study of Learner Characteristics Inven-
ies exist to describe what is known about students with sig- tory (LCI; Kearns, Kleinert, Kleinert, & Towles-Reeves,
nificant cognitive disabilities and their participation in
large-scale assessment programs.
1
The struggle to balance flexibility for the students and University of Connecticut, Storrs, CT, USA
psychometric rigor is a national issue. The federal gov-
Corresponding Author:
ernment set high standards for the technical quality of Jessica Goldstein, University of Connecticut, Neag School of Education,
these instruments, which states have struggled to meet. In 249 Glenbrook Road, Storrs, CT 06269, USA
2006, 38 states did not meet peer review standards for E-mail: jessica.goldstein@uconn.edu
118 The Journal of Special Education 46(2)

2006) data from more than 12,000 students from seven is not yet known is the extent to which all students have
states differentiated students with significant cognitive dis- access to the curriculum, appropriate assistive technology,
abilities according to language use (Kearns, Towles-Reeves, and other means to demonstrate learning” (p. 333). Little
Kleinert, Kleinert, & Thomas, 2009). Their study indicated is known about these students and their learning environ-
that a majority of the students in this population communi- ments and the impact of these issues on accountability
cated expressively and receptively at the symbolic level decisions.
(i.e., they communicate with a wide variety of intents and This study uses teacher focus group interviews and an
follow directions independently). A second, much smaller empirically evaluated classification scheme of data from a
group was at the emergent symbolic level of expressive and statewide alternate assessment to examine the accessibility of
receptive communication, meaning that these students use the grade-level academic content for all students with signifi-
pictures, objects, signs, and gestures in addition to oral cant cognitive disabilities. The study was conducted in Con-
speech and require additional cues to follow directions. The necticut with focus groups composed of special education
final group was at the presymbolic level of communication. teachers responsible for implementing the state’s alternate
These students’ communication efforts require interpreta- assessment. Unlike most states in which alternate assess-
tion from the listener or observer, and definition of their ments have been administered only in the past few years,
receptive language skills is less clear. An earlier study of Connecticut has been employing a system of alternate assess-
LCI data from three states indicated that most students in ment since 2000. At that time the statewide assessments, the
the study had functional reading and mathematics skills but Connecticut Mastery Test (CMT) in Grades 3 to 8 and the
that those skills were strongly correlated with expressive Connecticut Academic Performance Test (CAPT) in Grade
and receptive communication skills (Towles-Reeves, Kearns, 10, were expanded to include the CMT/CAPT Skills Check-
Kleinert, & Kleinert, 2009). Almond and Bechard (2005) list (hereafter “the Skills Checklist”). The current version of
identified similar patterns of heterogeneity in their study of the Skills Checklist, implemented in 2006, employs teacher
165 students participating in a pilot alternate assessment ratings of student knowledge both for use in improving
program. With regard to communication, almost 40% of the instruction and for statewide accountability purposes. We
students were able to communicate using 200 or more words, used existing data from the Connecticut AA-AAAS program
but approximately 10% did not use words to communicate. to examine teacher perceptions of student academic content
Research evidence also exists to demonstrate that students knowledge based on teacher ratings. The following sections
with significant cognitive disabilities are capable of learn- of the article provide an overview of the CMT/CAPT Skills
ing some components of reading (Browder, Wakeman, Checklist, a description of the research methodology, and the
Spooner, Ahlgrim-Delzell, & Algozzine, 2006) and mathe- results of the qualitative and quantitative analyses.
matics skills (Browder, Spooner, Ahlgrim-Delzell, Harris,
& Wakeman, 2008).
Although these studies provide an important step in Instrumentation
understanding the ways in which this population may access The Second Generation CMT/CAPT Skills Checklist was
academic content, there currently exists neither research developed in response to 2003 federal guidelines requiring
evidence nor a theoretical consensus regarding the appro- (a) grade-level-specific alternate assessments that reflected
priate expectations for academic content mastery by this grade-level content standards, (b) the establishment of
student population. Quenemoen (2008) offered, alternate achievement standards in a valid, documented
standard setting procedure, and (c) the adoption of these
We do not know as yet what will work the best in alternate achievement standards by the state board of edu-
teaching and in assessing students with significant cation (Behuniak & Amenta, 2009). When the Skills Check-
cognitive disabilities in the academic content. We are list was developed in 2005, it was envisioned that it would
seeing evidence of remarkable achievement, but this transform the educational process for students with signifi-
group is so varied in characteristics and the field of cant cognitive disabilities. Specifically, it is stated in the
severe disabilities is still divided on what appropriate technical manual that teachers will use the instrument to
outcomes we can and should expect. (p. 23) “inform instruction, monitor student growth and progress,
and document achievement” and that use of the Skills
Still, it is a federal requirement that all states must assess Checklist will result in “(a) a change in the nature of ins­
these students’ academic content knowledge annually. It is truction for students with significant cognitive disabilities,
not surprising that this situation gives rise to questions reg­ (b) an improvement in the quality of instruction for these
arding the extent of students’ exposure to academic content students, and (c) the greater inclusion of these students in
and the availability of appropriate communication tools. general education settings” (Connecticut State Department
Browder, Wakeman, and Flowers (2009) suggested, “What of Education [CSDE], 2006, p. 9).
Goldstein and Behuniak 119

The Skills Checklist is a nonsecure, stand-alone working focus group interview protocol was based on prior survey
document that teachers can use throughout the school year research on teacher use of the Skills Checklist to inform
to inform instruction and assess academic skills. All stu- instruction and monitor student progress. The final protocol
dents are assessed in reading and mathematics in Grades 3 is included in the appendix. Note that the protocol addresses
through 8 and 10; science is assessed in Grades 5, 8, and 10. teachers’ use of the Skills Checklist throughout the year for
The child’s primary special education teacher rates the stu- multiple purposes. Only issues related to the accessibility of
dent’s performance on a 3-point scale (0 = does not demon- the academic content for all students with significant cogni-
strate skill, 1 = developing/support, and 2 = mastered/ tive disabilities are addressed in the current article.
independent). When rating an item, any mode of communi- Participants were recruited by the consultant responsible
cation or responding that is typically utilized by the child is for managing the Skills Checklist program at CSDE. A sub-
acceptable. Each test item consists of three downward ext­ set of special education teachers who submitted a checklist
ensions that address the essence of the content standard and for a student in March 2009 was randomly sampled and
expected performance statements but in a simplified form then further limited to a pool of teachers who worked at
that makes them more accessible for students with signifi- schools near the focus group location. Those teachers
cant cognitive disabilities. The CMT/CAPT Skills Checklist received an e-mail from the CSDE stating that the state was
is available online at http://www.csde.state.ct.us/public/cedar/ partnering with the University of Connecticut to learn more
assessment/checklist/index.htm. about how the Skills Checklist was used at schools and its
The Skills Checklist also includes a section called Access impact on the education of students with significant cogni-
Skills, which was designed to assess preacademic skills that tive disabilities. Efforts were made to recruit an equal num-
students without disabilities typically develop prior to school ber of teachers at each level (elementary, middle, and high
entry. The Access Skills section covers receptive communi- school). However, a greater number of teachers at the ele-
cation, expressive communication, social-interactive com- mentary level opted to participate in the research project.
munication, basic literacy, and spatial relationships. Scores Three focus groups were conducted in June 2009 sepa-
from the Access Skills portion of the Skills Checklist are not rately with elementary (n = 10), middle (n = 5), and high
used for accountability purposes. school (n = 5) special education teachers who had adminis-
tered the Skills Checklist in the 2008–2009 academic year.
During the focus group sessions, efforts were made to
Method encourage open and honest responses by assuring complete
The purpose of this study was to examine teacher percep- confidentiality. Focus groups were video recorded with the
tions of academic content knowledge for students with sig- permission of the participants, and each session lasted
nificant cognitive disabilities in the context of the Skills approximately 90 minutes. Participants received a gift card
Checklist. The following research question was of interest: for a nominal amount to a local bookstore as compensation
Do teacher ratings of student academic content knowledge for their time. The recorded data from the focus groups
from the Skills Checklist data validate claims from teacher were transcribed verbatim, and transcripts were coded using
focus group interviews that the grade-level academic con- grounded theory (Corbin & Strauss, 2007). Transcripts
tent included in the Skills Checklist may not be accessible were analyzed on a line-by-line basis by the research team,
for all students with significant cognitive disabilities? and open coding was used to identify key themes in the
Several methodologies were used to investigate this que­ data. Several themes emerged from the focus group data; in
stion. Focus groups interviews were designed based on the this article, we focus exclusively on teachers’ feedback
results of a survey of special education teachers who had regarding the range of abilities of students with significant
administered the Skills Checklist in 2009. Quantitative analy- cognitive disabilities and the appropriateness of the Skills
ses of 2009 checklist data were conducted to validate themes Checklist for all of the students for whom the instrument
from the focus group data related to assessment administra- was designed. Specifically, teachers at all levels consis-
tion and the manner in which teachers evaluate student tently commented that the content on the Skills Checklist
knowledge. These methodologies are described below. may not be accessible for the range of students they serve.

Qualitative Method Quantitative Method


Patton (2001) suggested that qualitative inquiry is most To validate the focus group data and learn more about the
appropriate for investigations in which the processes, administration of the assessment, teacher ratings of stu-
impacts, or both are largely unspecified or difficult to mea- dents’ academic content knowledge from the 2009 CMT/
sure. This is particularly true of teachers’ implementation CAPT Skills Checklist were analyzed. Specifically, we attem­
of the CMT/CAPT Skills Checklist. Development of the pted to identify those students who had mastered certain
120 The Journal of Special Education 46(2)

areas of content as well as those who were unable to dem- demographic data, student disability category data, mean
onstrate academic content mastery, using data for all fifth scores for each domain of the Access Skill section of the
grade students (n = 519). The Skills Checklist was first Skills Checklist, and several items related to communica-
administered in 2006, when these students were enrolled in tion skills from the LCI (Kearns et al., 2006).
second grade; thus, the Skills Checklist was in place for a
majority of their time in elementary school. In addition,
internal analyses at the state level indicated that younger Results
students tended to demonstrate mastery on a greater portion The results of the qualitative research based on the focus
of the Skills Checklist content. group interviews and quantitative analyses based on the LCA
Although checklist ratings submitted to the state were are presented below.
teachers’ summative evaluations of students’ skills, analy-
ses of these data presented an opportunity to study teacher
perceptions of student academic content knowledge. More- Qualitative Analyses
over, these ratings were assigned by teachers, and this The focus group interview protocol was designed to inves-
potentially introduces considerable bias in the data. These tigate ongoing use of the checklist for both instruction and
limitations notwithstanding, the notion that certain students accountability. In both contexts, the participating teachers
were unable to engage with the academic content explicated took great care to point out that the Skills Checklist did not
on the Skills Checklist warranted further investigation. necessarily accommodate all students with significant cog-
Latent class analysis (LCA; Lazarsfeld & Henry, 1968; nitive disabilities. One teacher said,
Magidson & Vermunt, 2004) was used to identify patterns
in the teacher ratings. LCA is a statistical technique used to It’s interesting for a test that’s designed for a popula-
explore latent variables, which are characteristics of people tion that is supposedly proportionally so small com-
or places that cannot be directly observed or measured. Exam- pared to the CMT and CAPT population, within that
ples of such latent variables in the social sciences include population there’s such a broad range of ability. It’s
phenomena such as economic development, racial prejudice, hard to use the same tool to assess all the students who
and religious commitment. LCA is used to explore classifi- we utilize it for.
cation structures in these latent variables through the analy-
sis of variation on observed measures or dependent variables This theme is particularly salient because it was add­ressed
(McCutcheon, 1987). LCA is a useful analytic technique in at the elementary, middle, and high school levels and refer-
that it offers detailed information about class membership ences to the issue tended to elicit emotional responses from
including the latent class proportions (the proportion of the teachers.
cases in each latent class) and conditional probabilities (the The teachers took time to establish a picture of their
probabilities that a case belongs to a specific class given the classroom in communicating their use of the Skills Checklist.
pattern of responses to the observed measures). One teacher said,
In this study, the latent variable was student academic
content knowledge and the dependent variables were the My students are just learning how to pull their pants
teacher ratings for each content area in mathematics and down. Some of them are still learning how to feed
reading. Separate analyses were conducted for the mathe- themselves, doing simple exchanges with pictures and
matics and reading checklist data for the fifth grade stu- not even understanding that, not even knowing how to
dents. For both subjects, the mean scores for each content play appropriately with any toy.
area were used as the observed dependent variables; corre-
lations among the mean scores were indicated in the model. Another elementary school teacher described working with
The language arts domains included reading and respond- literature with a particular student in her class:
ing, exploring and responding to literature, communicating
with others, and English language conventions/writing; the The one that I love the best was the difference
mathematics domains included algebraic reasoning, geom- between real and make believe. I have a child who
etry and measurement, numerical and proportional reason- will pull his pants down. We have this discussion
ing, and probability and statistics. Note that all checklist with the parent that he doesn’t know . . . that his mom
items were rated on a 3-point scale in which 0 = does not wants him to be able to only pull his pants down
demonstrate skill, 1 = developing/support, and 2 = mas- when he’s inside rather than outside. He doesn’t have
tered/independent. After the number of latent classes was that realization anyways. So we went through that
established, available additional student data were used to and I said, “How am I ever going to teach real and
describe the groups of students. Descriptive statistics included make believe?”
Goldstein and Behuniak 121

Yet another teacher, who was particularly emotional in it,’ but we’re exposed to everything.” Most, however, were
describing her students, said, concerned: “How well are we prepared to teach students
with significant cognitive delays? I don’t think we’re very
My class is the medically fragile class and it’s called prepared at all.” The teachers explained that a community
that for a reason. We’re running over because of practitioners working with a similar population does not
Danny’s. . . he’s not breathing and you have to go exist for them. One teacher said, “We’re so isolated and
take care of that before we worry about if he’s look- there are so few teachers. I just cross my fingers sometimes,
ing at the picture. hope that I’m on the right track.” One teacher explained,
“Those of us who are old are probably not very well pre-
When asked to describe their use of the instrument, the pared. We had one course in mental retardation in college.”
teachers spoke of a divide between those students who work She later added,
on Access Skills and those who can engage more deeply with
academic content. Access Skills were designed to assess It just sort of floored me that the stuff that you kind
communication and quantitative and preacademic skills that of learn through working with these kids that people
students without disabilities typically develop prior to school don’t grasp. You know, you do first, second, third.
entry. Access Skills are included in the Skills Checklist at You realize, for example, that they don’t know first
each grade, but data from this section are not incorporated is one, and second is two. And the nuances of low-
into performance level scores for accountability. Many ness. You teach them all kinds of separate skills and
teachers described classrooms similar to this teacher: then you have to teach them to generalize.

I would say, of the seven students I teach, I use the Another said, “I’m a specialist in this area and I’m still
Skills Checklist to drive my instruction for three of making up my own materials.” In spite of these comments,
them. The other students it’s way too high for. I have another teacher added, “I don’t think you can teach some-
a student who’s currently functioning on the 18-month body in a college course what this looks like.”
level. I feel horrible filling out the checklist because Within this context, many teachers suggested that two
cognitively none of those things are reasonable for levels of testing might be more appropriate for these stu-
her. The only thing that is reasonable for her is the dents: “One for students who just have Access Skills and
Access Skills. then [one for] the ones where time in the general education
environment is appropriate and beneficial.” This teacher
One teacher said, “I happened to bring a copy of [the Skills added that some students have a diagnosis that leads to a
Checklist for] my highest functioning child. I chopped his situation in which the student is “either regressing medi-
name off of it, but he’s got a few things in the Access Skills cally or they’re not going to progress past a certain point.”
and we don’t have anything else.” Yet another middle A different assessment would be appropriate for these stu-
school teacher described a subset of her students as having dents because “it’s important for parents to be able to, and
limited decoding ability and another subset with decoding teachers to be able to focus on the progress, as opposed to
skills but limited comprehension skills. When asked how all those zeroes.”
she uses the Skills Checklist, this teacher responded, “I look The notion that the Skills Checklist may not be appropri-
at the Access Skills, but the other part of the test really does ate for all of the students for whom it was designed was an
not apply to my students. They’re so far removed from that unanticipated and alarming outcome of the focus group
level of instruction.” Another teacher elaborated, “I have to interviews. However, the small sample size of this portion
really extrapolate down, down, down to their level to really of the study left the pervasiveness of this problem unclear.
picture that in my mind, to see how it would affect my Data from the 2009 administration of the Skills Checklist
students.” were analyzed to investigate whether such patterns exist in
The teachers were also asked to discuss the extent to teacher ratings of student academic content knowledge at
which they felt that special education teachers were pre- the state level.
pared to teach academic content to students with significant
cognitive disabilities. Several teachers mentioned that they
were included in all of the academic professional devel- Quantitative Analyses
opment activities at their schools. Those teachers were For both subjects, two-, three-, and four-class models were
more positive, sharing comments such as, “We’re held fit to the data. Only results from analyses of the mathematics
accountable. We’re involved in all of the regular ed. data are presented here because of space limitations (see
instruction” and “You’re teaching me how to teach this fifth Note 1). Measures of model fit and the proportion of students
grade math and I’m never ever going to get anywhere near in each class for each analysis are included in Table 1.
122 The Journal of Special Education 46(2)

Table 1. Model Fit Statistics and Proportion of Students by Class for the Mathematics Data
Statistic Two-class model Three-class model Four-class model
Akaike information criterion 966.01 842.68 775.43
Sample-size-adjusted Bayesian 986.57 868.54 806.69
information criterion
Entropy .87 .90 .88
Proportion of students by latent
class
Class 1 .65 .49 .46
Class 2 .35 .28 .24
Class 3 .23 .09
Class 4 .21

For mathematics, the Akaike information criterion (AIC) skill explicated in a specific downward extension, even with
and sample-size-adjusted Bayesian information criterion teacher support. Stated differently, teachers’ ratings indi-
(BIC) measures of model fit decreased as the number of cated that half of the students did not engage with the math-
classes increased; this suggested better fit for a higher num- ematics content on the Skills Checklist. For reading, 45% of
ber of classes. However, the three-class model had a higher the students were in the group with low ratings (M =
entropy value, providing evidence of its superiority over the 0.18, SD = 0.17, n = 231). Ratings for the other two groups
other models. In addition, the Vuong–Lo–Rubin likelihood of students indicated that teachers viewed students’ skills as
ratio test for the four-class model was not significant (p = at least developing or supported, but to varying degrees.
.19), which indicates that the three-class model is sufficient Students with moderate ratings (mathematics: M = 0.94, SD
to represent the data. Moreover, the three-class model rep- = 0.26, n = 144; reading: M = 0.69, SD = 0.26, n = 205) and
resents a more even distribution of students across classes high ratings (mathematics: M = 1.48, SD = 0.26, n = 119;
as compared to the four-class model, which facilitates inter- reading: M = 1.25, SD = 0.26, n = 83) engaged with the
pretation of the data given the purpose of the analysis. Thus, academic content on the Skills Checklist, at least minimally.
the three-class model was selected for the mathematics Moreover, more students had higher scores in mathematics
data, and students were classified into three groups. than in reading. Replication studies were conducted with
After examination of the model fit statistics for the read- eighth grade data and produced similar findings: Approxi-
ing data, a three-class model was selected as well. The AIC, mately half of the students had mean scores that were close to
the sample-size-adjusted BIC, and entropy statistics sug- zero (62% for mathematics and 45% for reading). In their
gested a four-class model. However, the Vuong–Lo–Rubin entirety, the quantitative analyses confirmed the findings of
likelihood ratio test was not significant in either the three- the focus groups: Approximately half of the population did
class (p = .18) or the four-class (p = .21) model, suggesting not engage with the academic content on the Skills Checklist.
that a two-class model was sufficient to represent these Available data were used to describe the three groups of
data. In the two-class model, however, nearly all of the stu- students. Again, only results from analyses of the mathemat-
dents fell into the first class; further analyses of these clas- ics data are presented here. Review of these data indicated
sifications would limit the utility of comparisons across that the distribution of the demographic variables (gender,
classes. Moreover, the proportion of students in Class 2 of ethnicity, free or reduced-price lunch status, and English
the two-class model and Class 3 of the three-class model language learner status) is approximately equivalent across
was similar, as were the means on the content domains. the groups within each subject (see Table 2).
This suggested that the three-class model incorporated a This provides evidence to support the fairness and
split of the students in larger class of the two-class model. validity of the instrument in terms of demographic sub-
Use of the three-class model for the reading data allowed group performance. Examination of the distribution of stu-
for a parallel structure to the mathematics data. Thus, the dents within each disability category across each of the three
three-class model was selected for the reading data as well. groups, however, may highlight some disparities. In review-
The three classes represent variable levels of teacher rat- ing these data, we consider only the category in which n is
ings of content mastery and are subsequently referred to as greater than or equal to 20. The mathematics data (see
the students with low ratings, moderate ratings, and high Table 3) indicate that most of the students with autism and
ratings. For mathematics, the data show that 49% of stu- with multiple disabilities fell in the group with low ratings
dents had mean scores across the downward extensions that (55% and 69%, respectively) and most of the students with
are very close to 0 (M = 0.20, SD = 0.22, n = 256). A rating speech-language impairment (46%) were in the group with
of 0 indicates that the student could not demonstrate the high ratings.
Goldstein and Behuniak 123

Table 2. Demographic Data by Mathematics Classification


Low ratings Moderate rat- High ratings All ratings
Demographic category (%; n = 256) ings (%; n = 144) (%; n = 119) (%; n = 519)
Gender
Male 61 62 73 63
Female 39 38 27 35
Ethnicity
Black 20 17 20 19
Hispanic 22 19 18 20
White 56 56 54 55
Asian American 3 6 8 5
American Indian — 1 — 0
Lunch status
Eligible 43 41 47 43
Ineligible 57 59 53 55
English language learner (ELL)
ELL 6 12 13 9
Non-ELL 94 88 87 92

Table 3. Distribution of Mathematics Classification for Each Student Disability Category


Disability category Total n Low ratings (%) Moderate ratings (%) High ratings (%)
Autism 134 55 28 18
Emotional disturbance   4 75 — 25
Hearing impairment   3 33 — 67
Intellectual disability 110 36 38 26
Learning disability 12 17 42 42
Multiple disabilities 116 69 13 18
Orthopedic impairment   0 — — —
Other health impairment 49 35 37 29
Speech-language impairment 26 27 27 46
Traumatic brain injury   2 100 — —
Visual impairment   0 — — —
Missing 63

Students with intellectual disabilities and other health with low ratings were more likely to use cries, facial exp­
impairments were approximately evenly distributed across ressions, or changes in muscle tone to communicate whereas
the groups. Similar trends are evident in the reading data. A students in the group with high ratings were more likely to use
large portion of the students with autism (44%) and a major- symbolic language to communicate. Similarly, students in the
ity of students with multiple disabilities (66%) were in the group with high ratings were more likely to follow directions
group with low ratings. Alternatively, there were few stu- without additional cues. In addition, approximately one third
dents with speech-language impairments in the group with of the students in the group with low ratings use augmentative
low ratings (12%). communication during instruction, whereas most of the stu-
The additional tables provide a more complete picture of dents in the other two groups do not.
teachers’ views of these students’ skills. Mean scores for
each domain in the Access Skills section were lower and
more variable for the group with low ratings in both subjects Discussion
(see Table 4). The teachers in the focus group interviews referred to two
Individual items from the LCI provide greater detail on the types of students with significant cognitive disabilities:
typical modes of communication for each cluster (Table 5). those for whom grade-level academic content is accessible
With regard to communication skills, students in the group and those for whom it is not. These findings were confirmed
124 The Journal of Special Education 46(2)

Table 4. Mean Scores and Standard Deviations for Access Skill Domains by Mathematics Classification

Low ratings Moderate ratings High ratings All ratings


(n = 256) (n = 144) (n = 119) (n = 519)
Access skill domains M SD M SD M SD M SD
Expressive communication 1.13 0.52 1.74 0.25 1.88 0.15 1.47 0.52
Receptive communication 1.30 0.52 1.84 0.16 1.93 0.10 1.60 0.48
Basic literacy 1.19 0.63 1.88 0.18 1.95 0.08 1.55 0.58
Spatial relations 1.55 0.49 1.89 0.23 1.95 0.16 1.74 0.42
Social interactions 1.01 0.49 1.52 0.38 1.70 0.31 1.31 0.42

Table 5. Responses for Communication Items From the Learner Characteristics Inventory by Mathematics Classification

Low ratings Moderate ratings High ratings All ratings


Communication items (n = 256; %) (n = 144; %) (n = 119; %) (n = 519; %)
Expressive communication
Communicates primarily 19 1 1 10
through cries, facial expressions,
change in muscle tone
Uses intentional communication, 25 7 5 16
but not at a symbolic language
level
Uses symbolic language to 44 81 87 64
communicate
Receptive language
Uncertain response to sensory 5 — 1 3
stimuli
Alerts to sensory input from 14 — — 7
another person BUT requires
actual physical assistance
Requires additional cues to 50 44 25 42
follow 1–2 step directions
Independently follows 1–2 step 19 44 66 37
directions, does NOT need
additional cues
Augmentative communication
No 54 95 94 69
Yes 39 5 6 20
Missing 30 17 9 11

by analyses of checklist data in mathematics and reading. These data highlight the challenges of bringing an aca-
Our research indicated that nearly half of the teachers found demic curriculum to a population of students who do not
the content inaccessible for their students. Moreover, the share the same cognitive and communication abilities as the
data showed these students had poor communication skills, general student population. Students with significant cogni-
used augmentative communication as part of their instruc- tive disabilities have specific information processing chal-
tion, and were more likely to have disability classifications lenges related to short-term memory and the need for increased
of autism or multiple disabilities. Ratings for the remaining frequency of skill repetition with instructional feedback
students suggested teachers felt their students could engage (Kleinert, Browder, & Towles-Reeves, 2009). Teachers of
with the academic content with some support and with students with significant disabilities encounter many more
varying degrees of success. Students who could engage obstacles in working to cover the curriculum with their stu-
with the academic content had higher ratings of their com- dents, even a curriculum judged against alternate achieve-
munication skills and were less likely to use augmentative ment standards. The finding that certain disability categories
or assistive communication devices. were more prominent in the group with low ratings is also
Goldstein and Behuniak 125

of interest. These data are only suggestive of a relationship Implications and Directions
between disability category and teacher ratings, and further for Further Research
investigation into this pattern and its implications for
accountability practice is needed. The present study also The Skills Checklist is more than an assessment of academic
brings focus to teachers’ inadequate preparation to instruct progress for students with significant cognitive disabilities.
their students on content of this nature. Several teachers in It is also expected to fill a role as a comprehensive guide for
the focus groups were quick to note deficits in their own teachers to instruct students on AA-AAAS. The present
knowledge and skills. Inclusive accountability practices study suggests that these standards may not be accessible for
have questionable meaning and impact in the absence of the entire population of students it was designed to serve.
effective instruction. This finding suggests the need for Specifically, the data indicate that additional classification
resources devoted to helping teachers develop instruction to of this small population of students may be necessary for
meet their students’ learning challenges through preservice both effective instruction and effective assessment. A dis-
teacher training programs as well as professional develop- tinction within this population would have important impli-
ment programs for practicing educators. Certainly, more cations for educators and researchers alike.
research is needed to explore the interplay of these learning This study raises a number of questions regarding rea-
challenges and teachers’ efforts to expose students to an sonable expectations for students at presymbolic levels of
academic curriculum in combination with functional skills. communication. Are alternate assessments based on grade-
The current study provides a preliminary examination of specific content an appropriate accountability measure for
variability in academic content knowledge for one cohort of these students in their current form? What is the relation-
students in one state. ship between information processing challenges and the
The results of this study suggest that stratification of the extent to which the academic curriculum can be covered?
population of students with significant disabilities may be Do these students have sufficient opportunity to learn the
necessary to improve both the instruction and the assessment target academic content? Would it be beneficial to focus
of these students using the Skills Checklist. This is not a new instruction, and subsequently assessment, primarily on
concept. Browder, Flowers, and Wakeman (2008) conducted “preacademic” skills such as communication and other
an empirical evaluation of a classification scheme for the adaptive activities for these students? Clearly, more research
communication abilities of students with significant cognitive is necessary to address these questions.
disabilities, based on teacher ratings of 189 students. Their One alternative to the study of student academic content
study provided support for the notion that students with sig- knowledge is to focus research on the classroom teacher.
nificant cognitive disabilities can be classified by symbolic What knowledge and skills must a special education teacher
level (awareness or presymbolic, early symbolic, and sym- have to effectively deliver academic content to students with
bolic), and they offer this classification scheme as a tool for limited communication abilities? Can we assume that all
academic planning. The authors further suggest that it may be special education teachers have such knowledge? Would
appropriate to establish alternate expectations for achieve- alternate assessments based on grade-specific content be an
ment on alternate assessments based on these levels of com- appropriate accountability measure for this population if
munication, as most alternate assessments are biased toward teachers felt comfortable providing grade-specific academic
students who communicate at the symbolic level. Using the content instruction for these students? Furthermore, do edu-
LCI, Kearns et al. (2009) found that students at the presym- cators have sufficient access to assistive technology or aug-
bolic level of receptive and expressive communication were mentative communication devices to allow for efficient
also identified as having other issues that increased the com- communication with these students? Do educators with
plexity of their learning needs. These issues included low access to these technologies understand how to use them
levels of social engagement, limited motor skills, sensory effectively? Although these questions cannot be answered at
impairment, and health-related issues that affected school this time, the results of this study provide evidence of the
attendance. Their data suggested these students, students importance of addressing these issues in the future.
with the “most significant cognitive disabilities,” represented
approximately 10% of the population. The present study,
which incorporates data from both the LCI and Connecticut’s Conclusion
Skills Checklist, suggests this number may be much higher. The development and validation of alternate assessments
Although the results of the current study can be considered based on alternate achievement standards for accountability
only preliminary, the finding that the grade-level academic constitute a nascent field. Connecticut’s Skills Checklist was
content of one state’s alternate assessment may not be acces- created to change the nature of education for students with
sible for a large proportion of the students for whom it was significant cognitive disabilities. The quantitative analyses
targeted must be viewed as a cause of concern. indicated that approximately half of the target population can
126 The Journal of Special Education 46(2)

engage with the academic content on the Skills Checklist. 3. Has the Checklist affected teacher or parent expecta-
However, for the other half of the target student population, tions? How?
the effects of participation in the alternate assessment are 4. In general, do you feel that special education teachers
unclear. In the focus groups, teachers explained that these are well prepared to teach academic content to students
students may focus more on communication and other preac- with significant cognitive disabilities? (15 minutes)
ademic skills while in school. Our study raises both practical
and conceptual issues that relate to preservice education, Sharing Student Checklist Data. Participants will be
ongoing professional development, availability of resources asked to explain how data from the Checklist is shared with
within schools, and the broad issue regarding the impact of a the planning and placement team during the school year.
large-scale assessment program on the educational environ-
ment of students with significant cognitive disabilities. 1. How well do general education teachers, administra-
tors, and specialists in your school understand the
purpose and uses of the Checklist? To your knowl-
Appendix edge, do they receive any formal training on these
Focus Group Protocol issues?
2. Do you share data with the child study team during
Introduction. UConn investigators will introduce them- the school year? How is this information used by the
selves to the participants. The facilitator will distribute the child study team?
consent forms, explain the purpose of the session, and 3. How well do parents understand the purpose and uses
explain the procedures to the participants. Participants will of the Checklist? (10 minutes)
be given time to review the consent form, ask questions as
necessary, and sign the consent form. (15 minutes) Training for Checklist Administration. Participants will
Participant Introductions. Participants will be asked to be asked to discuss their training to use the Checklist.
provide their first name, the grade range of the students they
teach, and the Checklist training sessions they have attended. 1. Do you feel the Standard Checklist training program
(5 minutes) or the Certified Rater Training program can be im-
Use of the Checklist in the Classroom. Participants proved in any way? (5 minutes)
will be asked to describe how the Checklist is used to
design instruction for students with significant cognitive General Comments. Participants will be asked to offer
disabilities in their schools. Specifically, they will be any additional comments about the Checklist training pro-
asked: gram or the Checklist administration.

1. Do any of you use the Checklist during the fall, before 1. Is there anything you would like to see changed in the
the testing window? How do you use it? way the Checklist is designed or used? (5 minutes)
2. How is this similar or different to the way in which
you use the Checklist during the CMT/CAPT testing Debrief. The facilitator will summarize the discussions and
window, from January through March? solicit any additional comments or questions from the par-
3. The Checklist was designed to make grade-level aca- ticipants. The facilitator will thank the participants for their
demic content accessible for students with significant time and effort.
cognitive disabilities. Do you believe this goal has
been achieved? Why or why not? Abbreviations: CMT, Connecticut Mastery Test; CAPT, Connecticut
4. Can you describe a situation in which the Checklist Academic Performance Test.
affected your decision to expose a particular student
to grade-level content? (15 minutes) Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interests with
Impact of the Checklist. Participants will be asked to respect to the research, authorship, and/or publication of this
describe how the Checklist has impacted the educational of article.
students with significant cognitive disabilities.
Funding
1. How has the Checklist impacted the education of stu- The author(s) disclosed receipt of the following financial support
dents with significant cognitive disabilities? for the research, authorship, and/or publication of this article: This
2. Has the Checklist changed the way students with sig- manuscript was supported, in part, by the U.S. Department of
nificant cognitive disabilities exposed to the general Education Office of Special Education Programs (Grant No.
curriculum? How? H373X070003). However, the opinions expressed do not necessarily
Goldstein and Behuniak 127

reflect the position or policy of the U.S. Office of Special Education Gong, B., & Marion, S. (2006). Dealing with flexibility in assess-
Programs and no official endorsement should be inferred. ments for students with significant cognitive disabilities (Syn-
thesis Report 60). Minneapolis: University of Minnesota,
Note National Center on Educational Outcomes.
1. Analyses of reading data are available from the authors. Kearns, J., Kleinert, H., Kleinert, J., & Towles-Reeves, E. (2006).
Learner Characteristics Inventory. Lexington: University of
References Kentucky, National Alternate Assessment Center.
Almond, P., & Bechard, S. (2005). In-depth look at students who take Kearns, J. F., Towles-Reeves, E., Kleinert, H. L., &
alternate assessments: What do we know now? Retrieved from Kleinert, J. O., Thomas, M. K. (2009). Characteristics of and
http://www.measuredprogress.org/assessments/inclusive/ implications for students participating in alternate assess-
articlespapers.aspx ments based on alternate academic achievement standards.
Behuniak, P., & Amenta, J. (2009). Assessing all students in The Journal of Special Education. Advance online publica-
Connecticut. In W. M. Schafer & R. Lissitz (Eds.), Alternate tion. doi:10.1177/0022466909344223
assessments based on alternate achievement standards: Policy, Kleinert, H., Browder, D., & Towles-Reeves, E. (2009). Models of
practice, and potential (pp. 157–170). Baltimore, MD: Brooks. cognition for students with significant cognitive disabilities:
Browder, D., Flowers, C., Ahlgrim-Delzell, L., Karvonen, M., Implications for assessment. Review of Educational Research,
Spooner, F., & Algozzine, R. (2004). The alignment of alter- 79, 301–326.
nate assessment content to academic and functional curricula. Lazarsfeld, P. F., & Henry, N. W. (1968). Latent structure analy-
The Journal of Special Education, 37, 211–223. sis. Boston, MA: Houghton Mifflin.
Browder, D. M., Flowers, C., & Wakeman, S. Y. (2008). Facili- Magidson, J., & Vermunt, J. K. (2004). Latent class models. In
tating participation in assessments and the general curriculum: D. Kaplan (Ed.), The Sage handbook of quantitative methodology
Level of symbolic communication classification for students for the social sciences (pp. 175–198). Thousand Oaks, CA: Sage.
with significant cognitive disabilities. Assessment in Education: McCutcheon, A. (1987). Latent class analysis. Newbury Park, CA:
Principles, Policy, and Practice, 15(2), 137–151. Sage.
Browder, D., Spooner, F., Ahlgrim-Delzell, L., Harris, A., & Patton, M. P. (2001). Qualitative research & evaluation methods
Wakeman, S. (2008). A meta-analysis on teaching mathemat- (3rd ed.). Thousand Oaks, CA: Sage.
ics to students with significant cognitive disabilities. Excep- Perie, M. (2007). Setting alternate achievement standards. Lex-
tional Children, 74, 407–432. ington: University of Kentucky, National Alternate Assess-
Browder, D., Wakeman, S., & Flowers, C. (2009). Which came ment Center, Human Development Institute.
first: The curriculum or the assessment? In W. M. Schafer Quenemoen, R. (2008). A brief history of alternate assessments
& R. Lissitz (Eds.), Alternate assessments based on alter- based on alternate achievement standards (Synthesis Report
nate achievement standards: Policy, practice, and potential 68). Minneapolis: University of Minnesota, National Center
(pp. 329–332). Baltimore, MD: Brooks. on Educational Outcomes.
Browder, D., Wakeman, S., Spooner, F., Ahlgrim-Delzell, L., Towles-Reeves, E., Kearns, J., Kleinert, H., & Kleinert, J. (2009).
& Algozzine, B. (2006). Research on reading instruction for An analysis of the learning characteristics of students taking
individuals with significant cognitive disabilities. Exceptional alternate assessments based on alternate achievement stan-
Children, 72, 392–408. dards. The Journal of Special Education, 42, 241–254.
Cameto, R., Knokey, A.-M., Nagle, K., Sanford, C., Blackorby, J., U.S. Department of Education. (2003). Title I—Improving the aca-
Sinclair, B., & Riley, D. (2009). National profile on alternate demic achievement of the disadvantaged; Final rule, 68 Fed. Reg.
assessments based on alternate achievement standards (NCSER 236 (December 9, 2003). Washington, DC: U.S. Department of
2009-3014). Menlo Park, CA: SRI International. Education, Office of Elementary and Secondary Education.
Connecticut State Department of Education. (2006). CMT/CAPT U.S. Department of Education. (2005). Alternate achievement stan-
Skills Checklist Second Generation: Technical manual. Hartford, dards for students with the most significant cognitive disabilities:
CT: Author. Non regulatory guidance. Washington, DG: Office of Elemen-
Corbin, J., & Strauss, A. (2007). Basics of qualitative research: tary and Secondary Education.
Techniques and procedures for developing grounded theory U.S. Department of Education. (2009). State and local implemen-
(3rd ed.). Thousand Oaks, CA: Sage. tation of the No Child Left Behind Act, Volume V: Implementa-
Elliott, S. N., & Roach, A. T. (2007). Alternate assessments of tion of the 1 percent rule and 2 percent interim policy options.
students with significant disabilities: Alternative approaches, Washington, DC: Office of Planning, Evaluation, and Policy
common technical challenges. Applied Measurement in Edu- Development. Retrieved from http://www.ed.gov/rschstat/
cation, 20, 301–333. eval/disadv/nclb-disab/nclb-disab.pdf

You might also like