Gender PDF

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

University of Utah

Gender and the Perception of Knowledge in Political Discussion


Author(s): Jeanette Morehouse Mendez and Tracy Osborn
Reviewed work(s):
Source: Political Research Quarterly, Vol. 63, No. 2 (JUNE 2010), pp. 269-279
Published by: Sage Publications, Inc. on behalf of the University of Utah
Stable URL: http://www.jstor.org/stable/20721489 .
Accessed: 06/03/2012 22:54

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Sage Publications, Inc. and University of Utah are collaborating with JSTOR to digitize, preserve and extend
access to Political Research Quarterly.

http://www.jstor.org
Political Research Quarterly
Volume 63 Number 2
June2010 269-279

Gender and the Perception ofKnowledge O 2010 University ofUtah


10,1177/1065912908328860

in Political Discussion
http://prq.sagepub.com
hosted at
online. sagepub.com
http://

JeanetteMorehouse M?ndez
Oklahoma State University

Tracy Osborn
University of Iowa

Differences in knowledge about politics between men and women have the potential to affectpolitical discussion. We
examine differences in the perception of political knowledge between men and women and the effects these differences
have on how oftenmen and women talk about politics.We find bothmen and women perceive women to be less knowl
edgeable about politics and men to be more knowledgeable, regardless of the actual level of knowledge each discussion
partner holds. This perceptual knowledge gap could have ramifications for discussion as political participation, since
people turnto those theyperceive to be experts to gather political information.

Keywords: women and politics; public opinion; political participation

on thegender gap inAmerican politics could potentially have positive effects on democratic
has
Research focused primarily on finding and explaining government. However, the extent towhich women and
differences among men and women in vote choice, men discuss politics together is limited (Huckfeldt and

party identification, issue positions, and political Sprague 1995). Social identities, such as gender, shape
participation (Burns 2002; Seltzer, Newman and the groups within which one talks about anything,

Leighton 1997; Plutzer and Zipp 1996; Shapiro and including politics, and these identities instruct the
Majahan 1986). Less explored are gender differences information one extracts from the discussion, while
in other areas ofmass political behavior, such as polit further defining the ways inwhich this social identity
ical discussion. Huckfeldt and various colleagues is pertinent to the discussion (Cramer Walsh 2004).
argue variation in political discussion patterns and What this means for the interaction of gender and

topics helps to explain one's knowledge about, partic democratic citizenship is that the theoretical benefits
ipation in, and interest in the greater political process of political discussion may not be realized within the

(Huckfeldt and Sprague 1995). More importantly, reality of political discussion between the sexes. More
political discussion and the level of disagreement understanding is needed about how men and women
in political discussion have important repercussions discuss (or do not discuss) politics with each other and
for the health of democratic government in theUnited the consequences of these patterns.
States (Huckfeldt, Johnson, and Sprague 2004; One important component in explaining the fre
Cramer Walsh 2004; Mutz 2002). quency of political discussion is one's body of knowl
Many theorists place primary importance on this edge about politics. As Delli Carpini and Keeter
need for substantive deliberation to take place before "Political is . . .a critical
(1996, 6) note, knowledge
meaningful results, such as political equality, can be and distinct facilitator of other aspects of good

produced from democratic government (Fishkin citizenship." Knowledge about politics can make the
1992). As Mutz (2002, 122-23) argues, even when this products of political discussion more fruitful for
deliberation occurs in themostsimple, less-than-ideal democratic citizens, because people attempt to choose
form of exposure to viewpoints different from one's
own, changes, such as a greater willingness to extend
civil liberties to disliked groups, can occur. Particularly Jeanette Morehouse M?ndez, Assistant Professor, Oklahoma State

for a political minority, such as women, inclusion in University; e-mail: jeanette.mendez@okstate. edu.

and participation in the process of discussion could Tracy Osborn, Assistant Professor, University of Iowa; e-mail:

further this type of exposure to new viewpoints and tracy-osborn@uiowa.edu.

269
270 Political Research Quarterly

experts with whom to confer and learn about politics less likelytobelieve theirspouse followedpoliticsand
(Huckfeldt 2001). However, recent evidence demon knew a lot about politics than wives who named a
strates a sizable "knowledge gap," or difference in spouse, even when controlling for objective factors,
objective political knowledge levels, betweenmen such as the spouse's education level and age
andwomen (Mondak andAnderson 2004). This gap (Huckfeldtand Sprague 1995,201).
mirrorsothergaps in political participationbetween More recent work supports the continued exis
men andwomen (Burns,Schlozman, andVerba 2001) tence of sexual segregation in discussion. In a study
and is not entirely accounted for by survey measure of politicaldiscussion in Ireland,
Miller,Wilford, and
ment challenges (Mondak andAnderson 2004).With Donoghue (1999) findgenderaffectspolitical discus
the importantrole thatknowledge plays in political sion in additional ways. They find women are less
discussion, it is likely these differences between men inclined to discuss politics than men. Additionally,
and women's political knowledge can have important theyfindwomen believe theyare less able to per
ramifications for discussion as a form of political suade and influence others' political opinions than
participation forwomen and men. men, even when they hold strong opinions about pol
To that end, in this article, we examine how women itics. Finally, they find women are more likely than
andmen perceive thepolitical knowledge levels of men to stay out of a political argument, because they
thosewithwhom theydiscuss politics and theeffects are more likely to feel the other person's political
theseperceptionshave on the frequencyof political views will win out over their own. Similar to
discussion between men and women. We find that Huckfeldt and Sprague (1995), theauthorsfindmen
both men and women perceive women with whom aremore likelytodiscuss politicswith people outside
theydiscuss politics to be less knowledgeable about of the family, while women are more likely to name
politics, andmen withwhom theydiscuss politics to family members and husbands as discussion partners.
be more knowledgeable, regardless of the level of It is importantto note that the historyof gender
objectivepolitical knowledgeheld by thediscussant. relations in Ireland (such as the very recent changes
However,we find thatthisbias inperceptiondoes not to divorce laws) differsfrom those in theUnited
cause women and mento talk about politics less fre States, which may make comparison between women
quently with each other. Because discussion continues and men across the two countries less fruitful.
while theperceptionof politicalknowledge isflawed, Nevertheless, some of the same conclusions as
these differences have the potential to discount Huckfeldt and Sprague's earlier survey are evident,
women's contributions to political conversation. and as recently as the 1990s, Verba, Bums, and
men weremore likely
Schlozman (1997,1056) report
to say they discussed politics, and that they enjoyed
Gender and Political Discussion
discussing politics, than women. This persistence of
gender disparities in political discussion, similar to
In previous studies of political discussion between
other disparities in political participation for women,
women and men, a consistent finding is thatmen and
is particularly intriguing, given changes in women's
women do not talk about politics much with each
other.Huckfeldtand Sprague (1995, 195) note thatin participation as political elites over time. Although a
few studies find electoral context, particularly the
their1984 surveyof SouthBend residents'discussion,
presence of women candidates, creates a greater level
a "sexual segregation" exists in whom women and
of interest and engagement among women (Hansen
men choose as their partners in political discussion.
1997; Bums, Schlozman, and Verba 2001), the over
They find that among spouses, 76 percent of women all conclusion from studies of women's political dis
name their husband as a discussant, but only 55 per
cussion is that it is segregated from and occurs less
cent of husbands name their wife as a discussant (p.
often than discussion formen.
197). Among nonrelatives, this sexual segregation is
more pronounced: eighty-four percent of men named
The Role of Political Knowledge
only men as discussion partners, and 64 percent of
women named only women as discussion partners Although these patterns in sexual segregation are

(p. 195).Huckfeldtand Spraguedemonstratethatthis stark, it remains unclear whether these patterns of

segregation, even between spouses (on whom they discussion between men and women reflect differ
concentrate), has important consequences for the flow ence or discrimination. With regard to difference, it
of political information. Primarily, husbands in their may be thatwomen are systematically different from

survey who named wives as discussion partners were men on the characteristics that are antecedents to
M?ndez, Osborn /Gender and Perception ofKnowledge 271

political discussion. Alternatively, sexual segregation thanotherpoliticalprocesses (DelliCarpini andKeeter


inpolitical discussion betweenwomen andmen may 1996, 146).This findingis supportedby thework of
reflect a systematic discounting of women's potential Burns,Schlozman,andVerba (2001),who findwomen
forpolitical discussion.This distinctionis important can supply less political informationthanmen when
for two reasons. First, to the extent that discussion asked about it in a survey and are less likely to be sen
fuels democratic participation among citizenry, sitivetopolitical cues,meaning theyare less likelyto
understanding why women and men do not talk about know where groups, like theAmerican Association of

politicswith each other could informotherdebates RetiredPersons, standinpolicy debates.They linkthis


about discussion, such as the extent to which private difference in political information to women's lack of

political discussionmirrors the ideal conditions for participationinpoliticscomparedtomen andfind that


deliberative democracy (Conover, Searing, and this gap in participation levels disappears when
Crewe 2002; Mutz 2002). Second, if sex segregation women's differences in knowledge, interest, and effi
indiscussion is theresultof discrimination,thisdis cacy in politics, as well as available participation
tinctioncould point us towarda possible remedyto resources, are accounted for (Verba, Bums, and
sex segregation in discussion and its connection to Schlozman 1997).
women's political participation. There are two primary explanations for these dif
We argue the distinction between difference and ferences in knowledge.Mondak (2001) argues that
discrimination in political discussion lies in two the use of "don't know" as a response in typical sur
related characteristics that are antecedents to political vey question batteries,includingthoseused byDelli
discussion: discussant knowledge and perception of Carpini and Keeter (1996), systematically measures
this knowledge. Discussant knowledge, part of polit not only political knowledge but also a respondent's
ical expertise,plays an important
role inpoliticaldis comfort with answering "don't know" to a question
cussion (Huckfeldt 2001). Simply, people talk (Mondak2001,225).Women aremore likelytoanswer
politicsmore oftenwith those theyperceive to be "don't know," accounting for some of the disparity

political experts.Huckfeldt (2001) contends that (theyestimateup to 50 percent)betweenmen's and


when people judgewho is an expert,theydo so based women's political knowledge found in previous work
on objective criteria, particularly one's knowledge (Mondak and Anderson 2004, 510). However, these
about politics, his or her education, and his or her differences are also the result of social role expecta

partisan extremity, and they do so without bias tions, similar to those linked to the gender gap in
towardthosewith whom theydo not sharepolitical issue opinions and political participation. Gender
opinions. Burns, Verba, and Schlozman (2001) also women itis
socializationhas theeffectof eithertelling
find one's education, from which knowledge would not important for them to know the information or
follow, to be themost important explanation for the making the information women have about politics not
gender gap in political participation.However, in fit into thedefinitionofwhat is traditionally
political
theirexaminationof 1984 surveydata,Huckfeldtand (Verba, Bums, and Schlozman 1997; Sapiro 1983).
Sprague (1995) find male spouses discount their Processes linked to gender role socialization (particu
wives' "competence" and interest in politics, even larly in past generations), such as choice of occupation
when factors influencing this, such as education, are or civic involvement and educational attainment, also
controlled. shape political participation and therefore the opportu
This discounting of women's political knowledge nities for women to accumulate knowledge about
may be an effect of perception, based on gender dis politics (Bums,Schlozman,andVerba 2001).
crimination, or an effect of women's actual knowledge Althoughourpurposehere isnot toexplainwhy dif
about politics. In their extensive study of Americans' ferencesexistandpersistinpoliticalknowledgeamong
knowledgeof politics,Delli Carpini andKeeter (1996, women and men, the fact that these differences do exist
145-46) findwomen are less informedabout politics in political knowledge between men and women may
than men, particularly in substantive issue areas. The be an explanation forwhy women and men do not talk
one exception to this finding is thatwomen are not
sig aboutpolitics.Ifpeople seekout expertswithwhom to
nificantlydifferentfrommen in theirknowledge of talk about politics, women's lesser political knowledge
women's rights, such as the women's suffrage move may make them less desirable discussion partners.
ment. They contend this difference in information is However, it may be the case that judgments about
"consistent with a motivational basis," in that issues women's levels of political knowledge are as important
regarding women's rights are more salient to women as the actual objective level of
knowledge women
272 Political Research Quarterly

possess. Gender stereotypes exist forwomen and men was going to vote for in the election. This main
candidates; for instance, people believe women candi respondent could name up to five discussion partners,
dates possess traits stereotypical of women or believe but did not have to do so; he or she could name
theywill be more credible at handling "women's issues" anywhere from zero to five discussants. The main
(Huddy and Terkildsen 1993; Sanbonmatsu 2002). respondent also gave contact information for these
Within political participation, Verba, Burns, and discussants. Other relevant information, such as the
men appear to
Schlozman (1997, 1070) conclude that main respondent's educational level and political
have more of a "taste for politics," because men's party affiliation, was also collected.
greaterknowledge about and interestin politicsdoes The second sample contains information from the
not disappear when other explanations for the disparity, discussants named by the main respondents in the
such as resources, are accounted for. It could be the first sample. The discussants answered a battery of
case that participation in political discussion follows questions similar to those asked of the main respon
more "male"
from thisbelief thatpolitics is inherently dents, such as their own vote choice. Upon comple
for whatever reason, leading discussion partners to tion of this second sample, the resulting information
make assumptions about their partners' political gives us two main views concerning discussion
capabilities based on beliefs about
that gender. For patterns. First, we have information about the main
instance, theymay believe thatwomen know less about respondents' discussion network. That is, how large
politics regardlessofwhat theyactuallyknow. is their network, and whom does it contain? Second,
Below, we test two relationships related to gender we have information about a set of dyads, or pairs of
and political discussion to uncover the role political main respondents and individual discussants. Here,

knowledgeplays infosteringsex segregationinpolit we focus on the dynamics of the discussion dyads.


ical discussion. We expect two relationships among To address the first of the two relationships
political knowledge, discussion, and gender to expected above, whether men and women perceive
emerge. First, we expect the perception of women's women's levels of political knowledge differently,
political knowledge in political discussion will be we utilize the difference between the discussant's
lower thantheperceptionof thepolitical knowledge objectively measured political knowledge ("real"
of men discussion partners. We expect this relation knowledge) and the main respondent's perception of
shipwill exist in addition to any objective difference the discussant's level of political knowledge ("per
in the level of knowledge between women and men. ceived" knowledge) as our dependent variable.2 We
Second,we expect thisdifferencein theperceptionof use this variable (real minus perceived) because it
women's political knowledgewill lead respondents accounts for thepossibilitywe wish to test,which is
to discuss politics with women less frequently than whether men and women perceive their discussion

they do with men, due to the importance of expertise partners to have different levels of political knowl
in political discussion. edge. However, it does so by accounting for the
actual level of knowledge of that particular discussant
upon which the perception of knowledge is based.
Data and Research Design We measure the perception of knowledge using the

question, "Generally, speaking, how much do you


To examine these two questions, we use data from name of discussant)knows aboutpolitics?"
think(first
the 1996 PoliticalNetworkElection Study.1This sur FollowingfromHuckfeldt(2001) and themeasurement
vey consists of two samples containing information recommendations ofDelli Carpini andKeeter (1993,
about individualcitizens as well as thepeople with 1996), to measure objective knowledge, we use a
whom they discuss politics. Both samples were gath objectiveknowledge batterypreceded
three-question
ered in three waves of data collection prior to and fol by an introductory statement: (1) Whose responsibil
lowing the 1996 election season. In the first sample, ityis it todetermineifa law is constitutionalor not?
that of main respondents, respondents were asked Is it the president, the Congress, or the Supreme
various questions about their political discussion Court? (2) Next, what are the first ten amendments in
habits. Within these discussion habits, the respondent theConstitutioncalled? (3) How much of a majority
named up to five people with whom he or she dis is required for theU.S. Senate and House to override
cussed politics, and then identified various character a presidential veto? These three questions, although
istics of each of these discussion partners, such as brief, include two of Delli Carpini and Keeter's
whom the respondent thinks that discussion partner (1993) five recommendationsforquestions thathold
M?ndez, Osborn /Gender and Perception ofKnowledge 273

Table 1
Composition ofDyads within Discussion Networks andMean Levels ofObjective
Political Knowledge within Dyads
TypeofDyad Percentage Knowledge
Respondent Discussant
Knowledge
Women main respondent-womandiscussant 34.03 1.98 1.99
Woman main respondent-mandiscussant 22.71 1.94 2.47
Man main respondent-womandiscussant 13.29 2.26 2.03
Man main respondent-mandiscussant 29.97 2.47 2.42

Table 2
Mean Levels ofObjective Knowledge, Perceived Knowledge, and Frequency ofDiscussion by Gender
Objective Knowledge
Perceived Frequency of
Gender
Respondent Discussant Knowledge Discussion

2.24 Men 2.44 1.43 2.01


2.06 Women 2.00 1.22 1.85
Note: Objective political knowledge is scored 0 to 3. Perceived knowledge ismain respondent'sperceptionof discussant's knowledge,
scored0 to 2. Frequency of discussion is scored 0 to 3.

enough difficultyto adequately distinguishrespon shows thepercentageof dyads in each of thesefour


dents' differencesin knowledge levels by reducing categories in our sample. Table 1 also gives themean

reliabilityproblems stemmingfrom easy guessing. levelofobjectivepoliticalknowledgeheldby therespon


Although theadditionalquestion (theBill of Rights dentand discussantpairs in thefourdyads,andTable 2
question)was not recommendedspecificallybyDelli shows the overall objective knowledge, perceived knowl
Carpini and Keeter (1993), it does possess the face edge, and frequency of discussion formen and women
validity they emphasize for choosing civics-based in thesample.Two thingsare of note inTables 1 and2.
questions. This leads us to believe we have constructed First, the sexual segregation in political discussion dyads
a valid index formeasuring objective knowledge. mentioned above is clear, in that single-sex (woman
For the second relationship we test, whether per woman and man-man) dyads make up 64 percent of the

ception of political knowledge affectsthe frequency sample,andmixed-sexdyads composeonly 36 percent


of political discussion,we use themain respondent's of the sample. Second, the knowledge gap evident in the
perception of the frequency of political discussion researchofDelli Carpini andKeeter (1996);Mondak
with that discussant as the dependent variable.3 For and Anderson (2004); and Verba, Burns, and Schlozman
the first relationship, we utilize ordinary least squares (1997) also exists in our survey data. In each of the
regression (OLS), since the dependent variable is mixed-sex dyads, women's mean political knowledge is
continuous. For the second relationship, because the significantly less than men's (p < .05).Women's per
values of each dependent variable are ordinal ceived knowledge is also less thanmen's in the overall
responses (a great deal, average amount, and not sample, as is theirfrequency of discussion.
much; and never, rarely, sometimes, and often, Within themodels discussed below, we also control
respectively)we utilize ordered logistic regression for other variables known from previous research to
(Long 1997). affect political discussion. These include other factors
Both of the relationships we expect above require thatcontributeto theperceptionof political expertise
thatwe controlforeach of thefourtypesof discussion in a discussant, such as discussant and respondent
dyad combinations possible between men and women: a interest in politics, education level, partisan
extremity,
man names a man as a discussant (man-man), a man and age.We controlforobjectiveknowledgeinTable 3,
names a woman as a discussant (man-woman), a woman in that objective knowledge is a strongly significant
names aman as a discussant (woman-man), and a woman
predictor of perceptions about discussant expertise
names a woman as a discussant (woman-woman). Table 1 (Huckfeldt2001).4We includea controlforwhether
274 Political Research Quarterly

the discussant pair is married, since this is common Table 3


within man-woman and woman-man dyads. Finally, The Main Respondent's Difference inKnowledge
we control for the level of agreement, both perceived Perception ofDiscussant by theDiscussant's
and objectivelydefined,due to the strongeffectsof Interest, Education, and Partisan Extremity and

disagreement in discussion (Huckfeldtand Sprague by Other Factors


1995;Huckfeldt,Johnson,and Sprague 2004; Verba, Factor Coefficient
Bums, and Schlozman 1997; Bums, Schlozman, and
Objective discussant knowledge 0.93 (0.03)
Verba 2001).
Discussant interest3 -0.28 (0.04)
Discussant education0 -0.10 (0.03)
Perceptions ofPolitical Knowledge Discussant partisanextremity0 -0.04 (0.03)
Discussant aged -0.01(0.001)
To address the first of the two research questions, Woman main respondent-womandiscussant6 0.39 (0.07)
we determinewhether the differencebetween the Man main respondent-womandiscussant 0.39 (0.08)
level of objective political knowledge (real knowl Man main respondent-mandiscussant8 0.24 (0.07)
Objectively definedmain respondent
edge) and theperceived level of political knowledge -0.01
knowledge (0.03)
(perceivedknowledge) varies among each of thefour Main respondent interest3 -0.06 (0.04)
typesof genderdyads. Table 3 shows the resultsof Main respondenteducationb -0.01 (0.03)
knowl
the OLS regression of real minus perceived Main respondentpartisanextremity0 0.03 (0.03)
on the characteristics of the Perceived agreementbymain respondent11 -0.16 (0.06)
edge dyads.
Objectively defined agreement1 -0.08 (0.06)
As we expect, we find that the difference between
Political networkname generator* 0.02 (0.05)
real and perceivedpolitical knowledgevaries signifi 0.002
Main respondentaged (0.001)
cantlywith thegendercombinationof thediscussion Spouse 0.10 (0.09)
dyad.The scale of thedependentvariable (realminus Constant -1.29(0.18)
perceivedknowledge) is such thatthenegativeend of 1,356
the scale indicates the extreme position of a F/df/p 62/17/0.00
(-3.5)
R2 0.43
discussantwho has no objective knowledge but is
Root mean error 0.92
of
perceived toknow a greatdeal. The positiveend the
square

scale (3.3) indicates the extreme of a discussant who is Note: Regression model. Standard errors for coefficients and con

but objectively answered stant are shown in parentheses. Difference in knowledge


perceived to know nothing
percep
tion is main of discussant's knowledge
every question correctly. In the results, the three respondent's perception
subtracted from discussant's knowledge, where main
objective
positivecoefficientsforeach of thedyads (man-man, respondent's perception regarding how much
the discussant
woman-man is the
man-woman, woman-woman; knows about is 1 = not much, 2 = average amount, and
politics
omitted base category) show thatwith each change in 3 = great deal, and discussant's objectivelydefinedknowledge is
is toward the posi to knowledge
dyad combination, the pair moving
number of correct answers battery (0-3).
a. Campaign interest: 0 = not much interested, 1= somewhat
tive end of the scale. That is, for a woman main respon
interested, 2 = very interested.
dent who chooses a man as a discussant (the omitted
b. Education: years of school based on self report(range is 6-10).
base category), the constant of -1.24 indicates there is 1 = independent or nonpartisan, 2 = inde
c. Partisan extremity:
more knowledgeperceived thanthereisobjectivelyfor pendent
=
leaning to a party, 3 weak partisan,
4 =
strong partisan.

that discussant. Comparatively, when a man main d. Age: number of years of age.
discussant: 1 if dyad is
chooses a man as a discussant, the positive e. Woman main respondent-woman
respondent 0 all others.
and woman
moves them "up" the woman main discussant,
significantcoefficient(0.21)
respondent
f. Man main discussant: 1 if dyad is man
scale?the perceptionof knowledge is stillmore than
respondent-woman
main respondent and woman discussant, 0 all others.
theobjectiveknowledge,but thegap between thetwo g. Man main respondent-man discussant: 1 if dyad is man main

is smaller. For a woman main respondent who chooses respondent and


man discussant, 0 all others.
=
a woman as a discussant, the larger positive coefficient h. Perceived vote agreementbymain respondent: 1 respondent
discussant has same vote 0 = absence of
(0.36) moves them even farther up the scale, toward perceives preference,

agreement.
the point where objective knowledge becomes larger perceived
i. Objectively defined vote agreement: agreement regarding vote
than perceived knowledge. Finally, for the fourth cate choice; 1 = agreement based on self-report, 0
= no
agreement.
who chooses a woman as 1 = respondent
gory, a man main respondent j. Political network name generator: named dis

coefficient(0.39)
a discussant,thepositive significant cussant with whom they discuss politics, 0 =
respondent named

indicates they are the furthest up the scale. The women discussant with whom they discuss important matters.
M?ndez, Osborn /Gender and Perceptionof Knowledge 275

discussants chosen here have more objective knowledge Table 4


than they are perceived to have. The Main Respondent's Reported Frequency of
What these coefficients substantively mean is Political Discussion with theDiscussant by the
more easily seen in predicted values on the real Perceived Expertise of theDiscussant and by
scale for each of the four dyads, Various Other Factors
minus-perceived
holding all other variables constant at their means. Factor Coefficient
For the two dyads with men as discussants (man-man
on the knowledge scale Perception of discussant knowledge 1.13 (0.11 )
and woman-man), the values
Discussant interest0.12 (0.09)
are negative, -.02 for man-man and -.23 for
Discussant partisanextremity -0.09 (0.06)
woman-man. This indicates the main respondent Discussant age 0.003 (0.004)
(man or woman) perceives the men discussants to Woman main respondent-woman 0.02 (0.17)
know more than theyobjectively do. On the other discussant
Man main discussant -0.01
hand, for the two dyads with women as discussants respondent-woman (0.20)
Man main discussant 0.18 (0.17)
(man-woman and woman-woman), the values on the respondent-man
Main respondentknowledge 0.09 (0.06)
knowledge scale are positive, .16 for man-woman
Main respondentinterest 0.57 (0.09)
and .13 for woman-woman. This indicates themain Main respondentpartisanextremity -0.02 (0.07)
respondent (man or woman) perceives the women Perceived agreementbymain respondent 0.34 (0.14)
discussants know less than theyobjectively do. In Objectively defined agreement 0.12 (0.14)
Political networkname generator 0.18 (0.12)
short, both men and women main respondents under
Discussant's frequencyof discussion 0.36 (0.10)
estimate what women discussants objectively know,
Main respondentage -0.01 (0.004)
and inflate what men discussants objectively know.
Spouse 1.26(0.20)
We feel these results are remarkable for two rea Threshold (1) -0.43 (0.43)
sons. First, as we expected, men main respondents do Threshold (2) 2.19(0.41)
Threshold (3) 5.07 (0.44)
perceive women discussants to be less knowledge
able than they really are. However, it is interesting 1,234

that women main a woman dis xVdflp 307/16/0.00


respondents perceive Pseudo/?2 0.11
cussant's political knowledge to be lower than it
Note: Ordered logitmodel. Standard errorsfor coefficientsand
really is as well. Additionally, men discussants are
both men and women to be more knowl cutting-pointthresholdsare shown in parentheses.Frequency of
perceived by 0 = never, 1= 2 =
political discussion: rarely, sometimes, 3 =
edgeable about politics than theirobjective knowl often. Discussant's frequency of discussion: mean frequency of
score indicates?in fact, women inflate men's
edge political discussion reportedby thediscussant in his or her own
knowledge more thanmen do. Second,
these relation network of association.

ships are interesting because they exist while


accounting for the discussant's actual level of politi
cal knowledge within the variable. Women discus
In Table 4, three important patterns emerge. First,
sants may have lower objective levels of knowledge,
as shown in Table 1, but this is not the only distinc
thereisno significant
effectforanyof thethreetypes
of gender dyads for which we control. Substantively,
tion; they are also perceived to have lower levels of
thismeans theparticipantsin each dyad talk to each
knowledge relative towhat they actually know.5
other about politics at statistically indiscernible rates
from the other dyads. This discussion occurs despite
Perception ofKnowledge and the the fact that both women and men see women as less

Frequency ofPolitical Discussion "expert."6 Second, men and women main respondents
do not have a significantly
higher level of discussion
Given thesedifferences
betweengenderdyads in the with men discussants. One might expect men and
perception politicalknowledge,we expect thatboth
of women to speak more with men about politics, given
women and men main respondents will discuss politics their perception thatmen are more expert, but this is
less frequently with women and more frequently not the case. Third, despite the insignificantdyad
withmen, sincepeople choose todiscuss politicswith variables, theperceptionof knowledgegenerally is a
those theyperceive to be politicallyknowledgeable highly significant predictor of discussion frequency.
(Huckfeldt2001). Table 4 shows theordered logistic Therefore, the perception of knowledge generally is
regression of themain respondent's reported frequency an important determinant of how often one talks to a
of discussion on the dyad characteristics. discussant, but the differences among dyads are not,
276 Political Research Quarterly

Figure 1
Predicted Probabilities of Discussion Based on Perceived Expertise of
Discussant for Spousal and Nonspousal Dyads

GrapMa. Predicted Probability of Graph 1b. Predicted Probability of


Discussion withWomen Nonexperts Discussion with Women Experts

0,
0.6 4
0.5
0.4 ?
2
0.3 3Non-Spouse Q.
I Spouse
0.2 ^
0.1
0
never rarely sometimes
Frequency of Discussion
Frequency of Discussion

Graph 1 c. Predicted Probabilityof Graph 1d. Predicted Probabilityof Discussion


Discussion withMen Nonexperts withMen Experts

3Non-Spouse
I Spouse

never rarely sometimes


Frequency of Discussion Frequency of Discussion

Note: Figure 1 illustrates the effect of the spousal relationship on the frequency of discussion.

despite differences in how the members of those we havethe ability to measure here is how often
dyads perceive their discussant's knowledge. people talk to the discussants they already have, and
It is an interesting puzzle to find that women's not why they choose the discussants they choose
political knowledge is perceived to be less by both from the universe
possible of discussants.
women and men, and men's
political knowledge is Furthermore, for men
and women dyads, many of
perceived to be greater by women and men, but this whom are likely to be spouses (over two hundred
does not significantly change the frequency of dis dyad pairs in this survey are spouses), different fre
cussion, despite the importance of knowledge in the quency of discussion
may take place regardless of
frequency of discussion. The variable for whether the knowledge level because they are so proximate to
main respondent is the discussant's spouse may each other (Huckfeldt and Sprague 1995).7
explain this seeming contradiction. This variable is Figure 1 shows the predicted probabilities of the four
positive and
strongly significant, meaning spousal categories of discussion frequency (never, rarely,
dyads talk to each other more frequently about poli sometimes, and always) for women and men experts
tics. This suggests that although expertise plays an (those who are perceived to know "a great deal") ver
function may be limited sus men and women nonexperts (those who are per
important role, its evaluative
to those possible discussants that are proximate to the ceived to have "average" or "not much" knowledge).
main respondent?in their neighborhood, at their job, These probabilities are then broken down for spousal
or in theirhome (Huckfeldtand Sprague 1995;Beck dyads (thedark line)versusnonspousaldyads (thelight
1991). That is, people may look to experts, but only line). In comparing graphs la and lc (the left-hand
those experts around them; alternatively, they may side) to graphs lb and Id (theright-handside), it is
talk to nonexperts because they are proximate. What clear that, generally, discussion happens "often" with
M?ndez, Osborn /Gender and Perception ofKnowledge 277

experts rather than nonexperts, regardless of sex. of politicalknowledge amongwomen who enterinto
However, spouses clearly talkoftenmore with each politicaldiscussion.This suggestsbothdiscrimination
other thannonspousal dyads do, particularlyif that and difference may be explanations for women's lack

person is perceived to be an expert of either sex.Women in thatthe"knowledgegap" identified


of involvement,
nonexpertsare slightlyless likelytobe talkedtooften in the literature has the potential to become a sort of

by their spouses thanmen nonexperts, but this is


not a heuristic for assessing women's knowledge in political

significant difference. Overall, it appears frequency of discussion. This seems to lend support toVerba, Bums,
more by thespousalrelationship
discussionis driven of and Schlozman's (1997) conclusion thatthereis a dif
the dyad rather than the gender composition. ference, perhaps based on assumptions about women's
Although part of thefrequencyof discussionmay involvement in politics, in the "taste for politics"
be drivenby theproximityof discussants,particularly between men and women.

spouses,anotherpossibilityis toconsiderwhy people We are leftwith the somewhatcontradictory but


discuss politicswith othersin thefirstplace. The dif intriguing
finding,however, that the perceptionsof
ferences we find may be a result of people's seeking women's knowledge do not have a significant effect on
different types of discussants for different reasons. how often men and women discuss politics with each

Perhaps a respondent chooses one type of discussant other.This isdespite thefactthatperceptionof knowl
when he or she wants to learn information; thismay edge generallyhas a strongsignificant effect,in that
be a job for a perceived expert. On the other hand, a people want to discuss politicswith experts.Though
respondent may deliberately choose a discussant he or Huckfeldt(2001) assertspeople can judge theexpertise
she perceives to be less of an expert in order to edu level of thosewith whom theydiscuss politics accu
cate or proselytize to that discussant. One may also rately,theperceptioneffects we findhere suggestoth
choose certain discussants he or she perceives as erwise; there is within the judgmentof
distortion
knowledgeable about a specific topic, but not neces expertise according to one's gender. We suspect that
sarily generally knowledgeable about politics. even though discussion continues to take place at com

Though we lack theability to testthesepropositions parable levels between men and women main respon
here, with more explicit data about the reasons for dis dents and men or women discussants, the assessment
cussant selection and specific types of political knowl thatwomen know less about politics, even if that is not

edge, these other possibilities may become clearer. truly the case, may affect the dynamics of discussion
within the dyad. More specific measures of the dynam
ics within discussion and the outcomes of this discus
Conclusion
definetheeffectsof theperceptionof
sionmay further
We believe the divergent results presented above knowledge. For instance, is information culled from
have important but still undetermined repercussions for women and men discussants considered to be equally

politicaldiscussionas a formof politicalparticipation reliable, or equally important, to themain respondent?


States. The How does information taken from discussion with men
in the United systematic discounting of
women's political knowledge by both men and women affect vote choice or opinion formation, compared to

may be one key to unlocking the continually puzzling information taken from women discussants? Does the
lack of women's participation inmany aspects of poli fact that women are more likely to answer "don't
tics at the mass level. We know there are many know" to a political knowledge question, creating a
antecedents of women's participation in politics, mismeasurement of the gender gap in objective knowl

including a lack of resources (particularly, education), edge (Mondak andAnderson 2004), also contributeto
socialization processes, and various indicators of polit how women are assessed by their discussion partner
ical engagement, such as interest in politics and the lack when they talk about politics, or even to the lower per
of objectivepolitical knowledge (Verba,Burns, and ceptionof theirpoliticalknowledgeby thediscussant?
Schlozman 1994; Burns, Schlozman, and Verba 2001). Perhaps the tendency to answer "don't know" to politi
However, our results here point to the need to recon cal knowledge questions and be perceived as a less
sider the role that perception of women's stance on knowledgeable discussant is part of the same dynamic
these indicators of participation, rather than these of verbal and nonverbal communication within political

objective indicators of political engagement alone, discussion between men and women. Though we
in becoming a barrier towomen's participation in cannot measure these items here, we feel theymay be
plays
politics.Here, itmay stillbe inparta lackof objective betterindicatorsin thefuture
of theramificationsof the
politicalknowledgebut also theperceptionof thelack perceptualknowledgegap we in
identify thisstudy.
278 Political Research Quarterly

Another factor we do not consider here thatmay Finally, these results return us to the impact of
play an important
role is thecontextinwhich thepolit gender on political discussion as a form of democratic
ical discussion between men and women takes place. deliberation. If, as Fishkin (1992), Mutz (2002),
Delli Carpini and Keeter (1996); Burns, Schlozman, Huckfeldt (2001), and others attest, deliberation
and Verba (2001); and Hansen (1997) indicate that among citizens is critical for the function of democra
women's knowledge about and
interest in politics tic government, these results suggest that the outcomes
increase when the subject matter becomes relevant to of deliberation,such as the acquisitionof additional
women's place in the political system. It is possible, political information or an increase in tolerance, may
then, that the perception of women's knowledge in dis be temperedby thesocial dynamicsof theinteraction.
cussion may change within an environment with a Some criticsof deliberativedemocracyquestion the
heightened focus on gender, such as an election with a extent towhich the ideal conditionsof deliberation
major woman candidate,like theprimariesof 2008. It actually exist in real political conversation (Conover,
is also possible thatdifferent
indicatorsof knowledge, Searing, and Crewe 2002). These conditions, namely,
specific to these contexts, may be more accurate mea public deliberation without illegitimate coercion
sures of women's political knowledge, as Delli Carpini and with standards of political equality, may be under
and Keeter (1996) suggest. That is,more women may mined by the typeof social dynamicwe see here
be classifiedas expertswhen knowledge indicatorsare (Conover, Searing, and Crewe 2002). Conover,
based on information they are more motivated to know Searing,and Crewe (2002, 56), in fact,find standards
because of their gendered connection to it. ofpoliticalequalityinconversationareviolated inpart
Also of note is the consistencyand difference because women were discouraged from entering polit
between our findings and those of Huckfeldt and ical conversation; perceptions that women know less
Sprague (1995). Huckfeldtand Sprague's (1995, 202) about politics could be antecedents of this discourage
assessment of the 1984 South Bend data reveals a ment. In future research, itwill be interesting to recon
similar pattern, although focused on spouses, of the dis sider the substantive products of discussion between
counting of women's political interest and "compe men and women with this information about the
tence" by men. However, Huckfeldt and Sprague do not perception of women's political knowledge in mind.
account for theobjective levelof politicalknowledge For example,thediscountingofpoliticalknowledgeof
among women and men (only for education, a separate women could potentially affect the accuracy of infor
variable), leavingopen thepossibilitythatlowpercep mation gained from political discussion between men
tionof women is due to theirown lack of objective and women or the effects of women's proselytizing
knowledge about politics. As we have shown here, this aboutwomen candidates(Hansen 1997). Similarly,the
is not entirely the case; although women have a lower effects of disagreement in discussion on tolerance of
objective level of political knowledge, they are per other'sviews (Mutz 2002) may be muted or changed
ceived by bothmen to know less about politics than by differences in how women are perceived as discus
they really do. Also distinct, and perhaps more surpris sion partners. Overall, the evidence presented here
ing, is our finding that outside of the spousal relation questions the accuracy of perception of knowledge
ship, women perceive women to be less knowledgeable within political discussion and points to important
about politics. This leads to a puzzling frustration over considerations in the future study of women's political
what is necessary to change the gap in both percep participation.
tional and objective knowledge betweenmen and
women, and whether it is necessary to change this gap
to improve women's in politics. What we Notes
participation
can conclusivelysay about the frequencyof political
1.This surveyis discussed extensivelyinHuckfeldt, Sprague,
discussion between men and women is that gendered
and Levine (2000) andHuckfeldt (2001). Here we focus on infor
perceptions of knowledge do not affect this frequency. mation about the survey that is relevant to this specificproject.
There are other reasons women and men do or do not The surveywas conductedby theCenter forSurveyResearch at
talk about politics with each other. However, the impor IndianaUniversity in 1996-1997. The population surveyedis reg
tance of this study lies in the possibility thatmen istered voters in the Indianapolis and St. Louis metropolitan areas.
For the survey of main respondents, = and for the
2,174, survey
and women view women as less capable discussion
=
of respondents' discussants, 1,475. At least one discussant
partners when they discuss politics. This perceptual was interviewed for 872 of the main respondents who were will

knowledgegap inpoliticaldiscussionopens a host of ing togivemore explicit information


about theirdiscussants, such
for women's at as a last name and contact information. In many cases, more than
possible consequences participation
was to
both themass and elite levels. one discussant for each main respondent interviewed
M?ndez, Osborn /Gender and Perception ofKnowledge 279

create the sample of 1,475.Any interviewforwhich discussant Conover, Pamela Johnston, Donald Searing, and Ivor Crewe. 2002.

information was available from the main respondent was pursued The deliberative potential of political discussion. British
(Huckfeldt,Sprague, and Levine 2000, 649-50). JournalofPolitical Science 32 (1): 21-62.
knowl
2. To createthisvariable (objectiveknowledge-perceived CramerWalsh, Katherine. 2004. Talking about politics: Informal
edge),we standardizedthevariablesby turningtheminto scoresfor groups and social identity in American life. Chicago:
objectiveknowledgeand perceivedknowledge.Then we subtracted UniversityofChicago Press.
thestandardized perceivedknowledgefromthestandardizedobjec Delli Carpini,Michael, and ScottKeeter. 1996.What Americans
tive knowledge. The new variable is scaled -3.58 to 3.15. know about politics and why itmatters. New Haven: Yale
3. The specific surveyquestion is, "How often do you talk University Press.
about politicswith (firstname of thediscussant)?" -. 1993.Measuring political knowledge:Puttingfirstthings
4.We includeobjective knowledge as an independentvariable AmericanJournalofPolitical Science 37 (4): 1179-1206.
first.
despite the inclusionof objective knowledge in themeasurement Fishkin, James. 1992. Democracy and deliberation: New directions
of thedependentvariable. This is because themodel of percep for democratic reform. New Haven: Yale University Press.
tiveknowledgewould be incorrectlyspecifiedwithout account Hansen, Susan. 1997. Talking about politics: Gender and contex

ing for theobjective level of knowledge on which theperception tual effectson political proselytizing.Journal of Politics 59
is based. However, the dependent variable would be measured (1): 73-103.
incorrectly without accounting for how far perception is from Huckfeldt,Robert. 2001. The social communicationof political
realityinsteadof onlyperception. Without accountingforpercep AmericanJournalofPolitical Science 45 (2): 425-38.
expertise.
tion relative to objective knowledge, one could assume percep Huckfeldt,Robert,Paul Johnson,and JohnSprague. 2004. Political
tion is low because objective knowledge is low (i.e., that the disagreement: The survival of diverse opinions within communi

perception is simply accurate). Our dependent variable dispels cation networks. New York: Cambridge University Press.
this interpretation and focuses on discrimination, or the extent to Huckfeldt,Robert, and JohnSprague. 1995.Citizens,politics,and
which perception is offfromtherealityof theobjectivemeasure. social communication. New York: Cambridge University Press.
5. Throughout theanalysis,we havemodeled thebehavior of Huckfeldt,Robert, JohnSprague, and JeffreyLevine. 2000. The
men and women identically,noting any gender effects through dynamics of collective deliberation in the 1996 election:
thedyad variables. It could also verywell be the case thatmen Campaign effects on accessibility, certainty, and accuracy.
andwomen behave differentlyin termsof perceptionand should American Political Science Review 94 (3): 641-51.
be modeled as such.To testfor this,we split the sample bymain Huddy, Leonie, andNay da Terkildsen. 1993.Gender stereotypes
respondents into one analysis for men and one analysis for and the perception of male and female candidates. American
women and used dummy variables for same- and opposite-sex Journal ofPolitical Science 37 (1): 119-47.
discussants. Based on these results, we find that in terms of J. Scott. 1997. Regression models for categorical and lim
Long,
perceptionof expertise,men and women do behave similarlyin ited dependent variables. Thousand Oaks, CA: Sage.
making their assessments, and we are confident in the interpreta Miller, Robert, Rick Wilford, and Freda Donoghue. 1999.
tionof our resultsfromthecombined sample inTable 2. Personal dynamics as political participation. Political
6.We use theperceptionof knowledge tomeasure knowledge Research Quarterly 52 (2): 269-92.
in Table 3, because theoretically, one chooses discussants based Mondak, Jeffrey.2001. Developing valid knowledge scales.
on the perception of expertise, even though this perception is American Journal ofPolitical Science 45 (1): 224-38.
based on the levelof objectiveknowledge toa degree (asmodeled Mondak, Jeffrey, andMary Anderson. 2004. The knowledge gap:
and explained inTable 2). However, alternative models of the rela A reexamination of gender-based differences in political
tionship in Table 3 replacing the perceptionwith the objective knowledge. Journal ofPolitics 66 (2): 492-512.
knowledgemeasure fromTable 2 orwith theperception-objective Mutz, Diana. 2002. Cross-cutting social networks: Testing demo
dependent variable from Table 2 yield similar results for the rest cratic in practice. American Political Science Review
theory
of the dependent variables. 96(1): 111-26.
7.Interestingly, the results of Table 2 are consistent across Plutzer,Eric, and JohnZipp. 1996. Identitypolitics,partisanship,
dyads made up of spouses and nonspousal dyads, as the control and voting forwomen candidates.Public Opinion Quarterly
for spousal dyads is insignificant. 60 (1): 30-57.
Sanbonmatsu, Kira. 2002. Gender stereotypes and vote choice.
American Journal ofPolitical Science 46 (1): 20-34.
References Sapiro, Virginia. 1983. The political integrationof women.
Urbana: Universityof Illinois Press.
Beck, Paul Allen. 1991. Voter's intermediation environments in Seltzer,Richard, JodyNewman, andMelissa Voorhees Leighton.
the 1988 presidential contest.Public Opinion Quarterly 55 1997. Sex as a political variable. Boulder,CO: LynneReinner.
(3): 371-94. Shapiro, Robert, and HarpreetMajahan. 1986. Gender differ
Burns,Nancy. 2002. Gender: Public opinion and politicalaction. In ences in policy preferences: A summary of trends from
The state of thediscipline III, ed. Ira Katznelson and Helen the 1960s to the 1980s. Public Opinion Quarterly 50 (1):
Milner.Washington,DC: AmericanPoliticalScienceAssociation. 42-61.
Burns,Nancy, Kay Lehman Schlozman, and SidneyVerba. 2001. Verba, Sidney,Nancy Burns, andKay Lehman Schlozman. 1997.
The private rootsofpublic action: Gender, equality,and polit
Knowing and caring about politics: Gender and political
ical participation. MA: Harvard Press. Journal
Cambridge, University engagement. of Politics 59 (4): 1051-72.

You might also like