f6
04").
9t“

u.
v. '

.1 f
‘17?

 

niESiS
\

J Illilll'lllllill]Illiiililflilllilllllllli

3 1293 01559 2540

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

The Effect of Attitude Similarity On

Interpersonal Attractign: A Meta-Analysis

presented by

James Kuuipo Choy Ah Yun, Jr.

has been accepted towards fulfillment
of the requirements for

Ph. D . degree in Communication

 

Q (A/ ftu

Major professor

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE ll RETURN BOXtoromovothb checkoutfrorn your record.
TO AVOiD FINES roturn on or More data duo.

DATE DUE DATE DUE DATE DUE
C" i .1‘ [5894‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU loAn Aflirrnuttvo WWII Opportunity Intuition
mm:

7*? 7 *7 rfimm if , .__—__—_

ABSTRACT

THE EFFECT OF ATTITUDE SIMILARITY ON
INTERPERSONAL ATTRACTION: A META-ANALYSIS

By

James Kuuipo Choy Ah Yun, Jr.

This dissertation performed a meta-analysis on the effect that attitude similarity
has on interpersonal attraction. While attitude similarity is the extent to which people
share common attitudes, interpersonal attraction is the degree to which people like others
and has been typically measured by Byme's two-item interpersonal attraction measure.
Byme's interpersonal attraction measure includes an item asking subjects to rate how
much they like another target person and an item asking subjects to estimate how much
they would enjoy working with the target other.

This dissertation tested three hypotheses. The first hypothesis predicted a positive
relationship between the percentage of agreement of attitude similarity items and ratings
of interpersonal attraction across studies. The second hypothesis predicted a positive
relationship between the number of attitude items employed across studies and ratings of
interpersonal attraction. The final hypothesis predicted that interpersonal interaction
would moderate the effect that attitude similarity has on interpersonal attraction.

Eighty studies were used in this meta-analysis with a total of 92 effect sizes. To
standardize effect sizes across studies, each study finding was converted into a correlation

prior to the analyses and corrected for error of measurement.

The findings indicate that percentage of agreement was an important contributor
in explaining variance across studies, while the number of attitude items was not.
Additionally, two findings were revealed with respect to the effect that initial
interpersonal interaction has on the relationship between attitude similarity and
interpersonal attraction. In the comparison between the no-interaction and interaction
conditions, interpersonal interaction was found to moderate the effect of attitude
similarity on interpersonal attraction. In an analysis of the interaction conditions only, it
was revealed that interpersonal interaction initially (S-minute interactions) eliminated the
effect that attitude similarity has on interpersonal attraction, but over time (interactions

greater than 5-minutes, but not exceeding 30-minutes), an effect was found.

ACKNOWLEDGEMENTS

Many people assisted me in the completion of this dissertation. I would like to take this

opportunity to thank them:

Gary, Annie, and the nice folks at Mt. Pleasant - Through your wisdom I learned the
true meaning of probability.

Marge - You always kept me out of trouble, therefore I begrudgingly reliquish my
duties as office sitter. Thanks for all of your help.

Linda - The best raquetball partner a friend could have. Thanks for putting up with
me, keeping me on my toes, and letting me win a few.

Ron, Jack, and Frank - Taking your classes and having you on my committee made
me a better scholar and person. Thanks for your guidance and support. Because of you, I
leave MSU with a lifetime of knowledge.

Mom and Dad - You were there when I needed you. Thank you for your support and
in providing me with the work ethic that I needed to complete my education.

Sandi - I couldn’t have asked for a better advisor. Thanks for the research team
experience, teaching experience, and for the help on my dissertation. Most importantly,
thank you for being a friend.

Monique - You’re my best friend. Thanks for the support while I wrote this
dissertation and for the help in the final stages. You made my dissertation experience an

enjoyable one.

iv

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. vi
INTRODUCTION ............................................................................................................... 1
LITERATURE REVIEW .................................................................................................... 3
THEORY AND HYPOTHESES ......................................................................................... 8
Strength of the Attitude Similarity Manipulation .................................................... 9
Number of Attitude Items ...................................................................................... 10
Initial Interpersonal Interaction .............................................................................. 15
METHODS ........................................................................................................................ 22
Study Selection ...................................................................................................... 22
Criteria for Study Inclusion ................................................................................... 23
Meta-analytic approach .......................................................................................... 24
RESULTS .......................................................................................................................... 26
Attitude Similarity Effect Size ............................................................................... 26
Attitude Similarity Percentage of Agreement and Effect Size .............................. 27
Number of Attitude Items and Effect Size ............................................................. 29
Initial Interaction as a Moderating Variable .......................................................... 30
Multiple Regression of Interpersonal Attraction Predictors .................................. 32
DISCUSSION .................................................................................................................... 33
Overall Similarity Effect Size ................................................................................ 33
Percentage of Agreement ....................................................................................... 34

Attitude Items and Effect Size ............................................................................... 35

Interaction as a Moderator ..................................................................................... 36
Limitations to Meta-Analysis ................................................................................ 37
Future Research ..................................................................................................... 39
CONCLUSIONS ................................................................................................................ 45
APPENDIX A: Studies used for Meta-Analysis .............................................................. 47
APPENDIX B: Overall Effect Size and Subgroup Analysis ............................................ 52
APPENDIX C: Plot of Number of Attitude Items with Effect Sizes ............................... 53
APPENDIX D: McCroskey & McCain Attraction Measure ............................................ 54
LIST OF REFERENCES ................................................................................................... 55

vi

LIST OF TABLES

Table 1. Meta-analysis findings for all effect sizes used .................................................. 26
Table 2. Interpersonal attraction means across studies ..................................................... 27
Table 3. Meta findings of percentage of agreement ......................................................... 29
Table 4. Meta findings of interaction absent or present .................................................... 31
Table 5. Meta findings of varying interaction levels ........................................................ 31

vii

INTRODUCTION

While some posit that the most established and well-known finding in the
interpersonal literature is that attitude similarity creates interpersonal attraction
(Berscheid & Walster, 1983; Cappella & Palmer, 1990; De Wolfe & Jackson, 1984;
Fishbein & Aijzen, 1972; Parks & Adelman, 1983), others contend that this belief is
unfounded (Sunnafrank, 1992) and have gone so far as to pronounce the issue dead
(Bochner, 1991). Clearly, extreme differences exist in the perceived effect that attitude
similarity has on interpersonal attraction.

Although several narrative accounts have explored the effect that attitude
similarity has on interpersonal attraction (Byme 1969; Byrne & Griffitt, 1973; Byrne,
Clore, & Smeaton, 1986; Simons, Berkowitz, & Moyer, 1970), these summaries are
limited to historical reviews of the progression of research in this area. As such, these
accounts offer little assistance in resolving whether attitude similarity enhances
interpersonal attraction.

In an effort to explore the differences in opinion with regard to the effect that
attitude similarity has on interpersonal attraction, Communication Monographs invited
scholars (Byme, 1992; Sunnafrank, 1992) holding different views to discuss their ideas.
While Byrne maintained in his Communication Monographs article that attitude
similarity increases interpersonal attraction, Sunnafrank countered in his article by
arguing that the effect of attitude similarity on interpersonal attraction disappears in

normal developing relationships. Unfortunately, these articles did little to bridge the gap

2

in ideological differences. In fact, research continues without apparent resolve (Tan &
Singh, 1995).

This dissertation will illuminate some of the issues in the attitude similarity debate
by employing a meta-analytic approach to the attitude similarity literature. Performing a
meta-analysis on this literature will yield valuable findings. In particular, it will provide
a cumulative examination of the effect that attitude-similarity has on interpersonal
attraction and provide an opportunity to test variables which may moderate the
relationship between attitude similarity and interpersonal attraction.

More specifically, this dissertation will present the uncorrected and corrected
effect size (correlation coefficient) between attitude similarity and interpersonal
attraction, assess the effect that differences in the manipulation of attitude items across
studies have on the relationship between attitude similarity and interpersonal attraction,
test the extent to which initial interaction moderates the effect that attitude similarity has
on interpersonal attraction, and conclude with future research that should be conducted in
the attitude similarity and interpersonal attraction domain.

In addition to providing a more detailed account of the research concerning the
effect of attitude similarity on interpersonal attraction, using a meta-analytic approach is
warranted for several reasons. First, at the individual study level, sampling error exerts a
significant toll on findings (Hunter & Schmidt, 1990). Meta-analysis aids in overcoming
sampling error by combining findings from many studies and weighting them by their
sample size. As such, the effect size obtained from a meta-analysis is a more accurate

estimate of the actual effect size than findings from any individual study. Second, meta-

3

analysis allows for a summed effect size that is corrected for error of measurement.

Given that error of measurement systematically lowers the correlation, meta-analysis
allows for the curnulation of studies as if they had perfect measurement. Finally, meta-
analysis enables the identification of moderating variables in the literature. If there is
variance in the effect sizes across studies, meta-analysis provides the tools to determine if
these differences are artifactual (Hunter & Schmidt, 1990). If the differences are not
artifactual, potential moderating variables can be examined by dividing studies into
appropriate conditions.

While meta-analysis is useful in cumulating findings across studies, it has several
limitations. First, it is constrained to examining only variables that have been previously
addressed by researchers. Consequently, while important relationships between variables
may be hypothesized, meta-analysis cannot test these relationships unless it has
previously been examined by others. Second, if potential moderating variables are
discovered in the literature, a distribution of studies into appropriate subgroups may
provide unequal comparisons. For example, while one subgroup may have many studies
to estimate an effect size, its counterpart subgroup may be limited to only a handful of
comparison studies. As such, while one subgroup may produce a very stable estimate of
an effect size, another subgroup made up of only a few studies may produce a less stable
estimate of an effect size.

LITERATURE REVIEW
As early as the fourth century BC, Aristotle suggested that friends regard the

same things as good and evil (translated, 1932). Working from the similarity principle

4

introduced by Aristotle, Pearson and Lee (1903), conducted a correlational study using
the variable of attitude similarity. Their findings revealed high similarities between
spouses with respect to their attitudes on a host of items. Following Pearson and Lee's
study, several researchers successfully replicated their initial findings (Hunt, 1935;
Kirkpaka & Stone, 1935; Morgan & Remmers, 1935; Newcomb & Svelha, 1937;
Schiller, 1932; Schooley, 1936).

Winslow (193 7) extended the correlational analysis beyond spouses and attitude
similarity. In his study on friends, he found high positive correlations between the
attitudes of friends. Later research replicated Winslow's findings (Richardson, 1940).

By 1960, research concerning attitude similarity and close relationships merely
examined the extent to which people in close relationships, e.g., spouses and friends, held
similar attitudes. In an attempt to broaden the understanding of the potential effects of
attitude similarity, Newcomb (1961) studied the effect of attitude similarity on attraction
longitudinally. In his study, Newcomb assembled groups of housemates and asked them
to complete attitude questionnaires. By collecting interpersonal attraction data on all
housemates toward one another at several time points, his findings revealed that attitude
similarity predicted attraction in later relational stages, but not early ones.

In an attempt to explain his findings, Newcomb suggested that people lack
adequate attitudinal information upon which to make inferences about others in the early
stages of relationships. As such, people become information gatherers, seeking to obtain
enough attitudinal information to make interpersonal attraction judgments. Therefore, he

suggested that people in the initial stages of a close relationship will not be more

5

interpersonally attracted to attitudinally similar others in comparison to dissimilar others,
because they have not revealed that they hold similar attitudes.

According to Newcomb, if people spend considerable time with one another, e.g.,
roommates, they obtain sufficient attitudinal information to enable them to make
adequate assessments about interpersonal attraction. Consequently, in long term
relationships, individuals will regard attitudinally similar others as interpersonally
attractive and attitudinally dissimilar others as interpersonally unattractive.

All of the research regarding attitude similarity and interpersonal attraction prior
to Byme's research (1961) was correlational. One problem with correlational data is that
the interpretation of causality in not always clear. Attempting to prove that attitude
similarity enhances interpersonal attraction, social scientists embarked upon a research
program in which they experimentally induced attitude similarity and subsequently
assessed the extent to which people were interpersonally attracted to attitudinally similar
others.

Research concerning attitude-sirnilarity and interpersonal attraction has followed a
similar pattern (Byme, 1992; Cherry, Byme, & Mitchell, 1977; Bond, Byme, &
Diamond, 1968; Byme, Griffitt, & Golightly, 1966; Byme & Griffitt, 1966; Byrne, Ervin,
& Lamberth; 1970, Curran & Lippold, 1975; Griffitt, 1969). Typically, subjects
completed a host of questions designed to represent their attitudes. Items used to assess
attitudes measured issues such as school desegregation, politics, gardening, birth control,
dating, and the advisability of freshman having cars on campus (Byme & Rhamey, 1965).

About one week after subjects completed attitude questionnaires, they were asked to

6

evaluate their interpersonal attraction toward a bogus stranger given a list of the same
attitude items that the stranger supposedly completed previously. To induce varying
levels of attitude similarity, answers from the bogus stranger's attitude items were varied
such that they were highly similar, e.g., six similar and two dissimilar or highly
dissimilar, e.g., two similar and six dissimilar to the subject's responses to the same
attitude items. Finally, subjects reported judgments of interpersonal attraction toward the
bogus stranger.

Byrne and Griffitt (1973), define interpersonal attraction as an individual's
affective evaluation of another, and interest in interpersonal attraction has received
attention by communication scholars for two primary reasons. First, interpersonal
attraction is positively related to the extent to which people communicate with others.
Consequently, interpersonal attraction is one antecedent to predict communication
partners that people have. Second, interpersonal attraction is positively related to the
amount of influence that others have on us in interpersonal exchanges (Berscheid &
Walster, 1969). As such, interpersonally attractive others can potentially play an
influential role in our behaviors. Given the significance of interpersonal attraction in our
lives, research concerning this variable is important.

A Predominantly, researchers have measured interpersonal attraction through the
sum of two seven-point Likert-type items taken from Byme's Interpersonal Judgment
Scale (Byme, 1971 ). The interpersonal attraction items ask subjects to report the extent
to which they (1) like a target person and (2) would enjoy working with a target person.

After conducting and reviewing a host of studies using the bogus stranger technique in

7

which attitude similarity was induced, Byme and Nelson (1965) proposed a linear
function of Y = 5.44X + 6.62 to describe the effect of attitude similarity (X) on attraction
(Y)-

A series of studies which followed the introduction of Byme and Nelson's
proposed linear function appeared to offer support (Cherry, Byme, & Mitchell, 1976;
Bond, Byrne, & Diamond, 1968; Byrne, Griffitt, & Golightly, 1966; Byme & Griffitt,
1966; Byme, Ervin, & Lamberth; 1970, Curran & Lippold, 1975; Griffitt, 1969).
Consequently, Byme (1992) reported that a lawful relationship exists between attitude
similarity and interpersonal attraction. However, some studies have reported findings
which are inconsistent with the linear function proposed above (Curran & Lippold, 1975;
Sunnafrank, 1983; Sunnafrank, 1984, Sunnafrank, 1985; Sunnafrank, 1986; Sunnafrank
& Miller, 1981).

There are two major arguments against the lawfirl relationship that has been
proclaimed between attitude similarity and interpersonal attraction. The first posits that
communication is a moderator that eliminates the effect that attitude similarity has on
interpersonal attraction (Sunnafrank, 1986; Sunnafrank, 1985; Sunnafrank, 1984;
Sunnafrank & Miller, 1983), by making attitudinally dissimilar others appear more
interpersonally attractive during interaction processes.

A second explanation is the repulsion hypothesis (Rosenbaum, 1986). According
to the repulsion hypothesis, attitude similarity does not heighten interpersonal attraction,

because similarity is expected. However, the discovery of attitude dissimilarity is

8

unexpected and aversive, resulting in interpersonal repulsiveness which decreases
interpersonal attractiveness.

In short, the repulsion hypothesis suggests that the important relationship is not
between attitude similarity and interpersonal attraction, but between attitude dissimilarity
and interpersonal repulsion. While some research has been conducted to test the
repulsion hypothesis (Rosenbaum, 1986), at present an insufficient amount of evidence is
available to confirm it as a plausible explanation. In particular, the testing of the
repulsion hypothesis requires a no-attitude information condition to provide an accurate
test, which has not been sufficiently addressed (Byme, 1992).

A final way to examine the relationship between attitude similarity and
interpersonal attraction is to propose the opposite path, that interpersonal attraction
increases attitude similarity between interactants (Berscheid & Walster, 1983; Stephen,
1985). However, an insufficient number of studies (k=12) have been performed to
examine this relationship adequately. Consequently, the only attention given to the idea
that interpersonal attraction produces greater attitude similarity is in the final chapter
regarding firture research.

THEORY AND HYPOTHESES

As suggested previously, meta-analysis is useful in understanding large bodies of
research examining the same variables. This study provides a rationale to predict
differences in effect sizes across studies with regard to the percentage of agreement when
inducing attitude similarity, the number of items used in the attitude similarity induction,

and the effect that interaction has as a moderating variable between attitude similarity and

9

interpersonal attraction. Each of the above differences are discussed in turn and
predictions of their effects are presented.

According to Byme's (1969) reinforcement-affect model of attraction, people have
more positive affective responses toward people holding similar attitudes, because they
like people who view the world in the same way. In short, Byme's reinforcement-affect
model suggests a positive linear relationship between perceived attitude similarity and
interpersonal attraction.

Within the attitude similarity research there is extreme variation in the percentage
of agreement in the manipulation of the attitude similarity variable. For example, Singh's
(197 5) study was designed such that subjects in the attitude similar condition agreed on
all of the items and subjects in the attitude dissimilar condition did not agree on any of
the items. In contrast, Tesser's (1971) study was designed such that subjects in the
attitude similar condition agreed on two-thirds of the items and subjects in the attitude
dissimilar condition agreed on one-third of the items.

If Byme's model is accurate, then variation in perceived similarity should be
directly related to the degree in which people assess the interpersonal attractiveness of
another. More specifically, given that there are differences in the degree to which people
are presented as attitudinally similar, the greater the percentage of agreement, the greater
effect it should have on interpersonal attraction.

A second reason to expect that the greater percentage of agreement in the attitude

similarity manipulation will be positively related to its corresponding effect size is with

10

regard to restriction in range. Because restriction in the range of an independent variable
systematically attenuates the effect on the dependent variable, it follows that as the range
or percentage of agreement between the attitude similarity and dissimilarity variable
increases, it will produce a greater effect on the dependent variable of interpersonal
attraction. Given the apparent relationship between the percentage of agreement used in
the attitude similarity manipulation and its effect on a person's judgment of interpersonal
attraction, the following hypothesis is offered.

H1: As the difference in percentage between attitude similar and dissimilar

conditions increases across studies, so will the correlation between attitude
similarity and interpersonal attraction.

NumhcmLAtlimdellcmfi

One reason to expect that the number of attitude items used in a study will
influence the extent to which others are perceived as interpersonally attraction is the idea
that some issues are more important to people than others. For example, an extremely
religious person who is disinterested in sports will weigh attitude similarity on the belief
that God exists as more important than the belief that Big Ten basketball teams are
generally better than Pac Ten basketball teams.

One area where topic or issue importance has been studied is in the persuasion
literature. For example, prior research has shown that as an issue becomes more salient
to an individual, it increases their motivation to listen to messages concerning that
particular topic (Petty & Cacioppo, 1979; Petty, Cacioppo, & Goldman, 1981; Petty,
Cacioppo, & Schumann, 1983). Consequently, item importance appears particularly

relevant to this dissertation in that some issues, those that are deemed important by

11

individuals, will receive greater attention and serve as a stronger cue in assessing the
interpersonal attraction of others.

A persuasion theory addressing the relative importance of particular issues to
individuals is the Theory of Reasoned Action (Fishbein & Ajzen, 1975). According to
the theory of reasoned action, important attitudes or beliefs are given greater weight than
less important ones in a person's decision making process.

The idea that people give more weight to important issues has been addressed by
research. For example, Warshaw (1980) examined the decision making process with
regard to consumer purchases. In this study, subjects were asked to make a list of
favorable and unfavorable reasons to make a certain purchase decision. In some cases, it
was revealed that although subjects listed more unfavorable reasons than favorable ones,
they still had intentions to make the purchase. The author of this study maintained that in
these cases, a few highly weighted favorable purchase reasons were strong enough to
outweigh a greater number of lower weighted unfavorable purchasing reasons.

In another study researching the idea of the weighting of important issues,
Bowman and Fishbein (1978) examined individuals' attitudes toward an Oregon nuclear
safeguard initiative. This study revealed that an indicator of voting behavior was the
weight that people placed on reasons to vote for or against the initiative. That is, the
stronger the weight of a belief about the initiative, the greater effect it had on a person's
overall voting decision.

The idea that people weigh certain attitudinal items as more important than others

when making interpersonal judgments of others seems reasonable. For example, consider

12

a likely dilemma that two friends may face with the respect to their attitudes toward
military spending, drug legalization, and abortion. While the pair may attitudinally differ
on issues that they consider unimportant, such as military spending and the need for drug
legalization, they may agree on an extremely important issue to them such as abortion. In
this case, attitude similarity on the abortion issue would likely have greater weight than
on the issues in which there is dissimilarity. Consequently, although there is overall
attitude disagreement (assuming that there are only three attitude issues in this
relationship), it is likely that these friends would still find each other interpersonally
attractive.

So why should it be expected that the number of items used in a study will explain
varying effect sizes of interpersonal attraction found across studies? One possible
explanation is that as the number of attitude items increases in a study, so does the
opportunity for people to agree or disagree on an issue that is important to them.
Additionally, if attitudes are revealed on issues that a person regards as important, then
he/she is likely to use this information to form a stronger judgment of interpersonal
attraction than another who fails to uncover information about a topic that is important to
them.

Numerous studies have shown a positive relationship between the amount of
information that people have about others and their willingness to make inferences about
them (Condon & Crano, 1988; Schwartz & Smith, 1976; Yabrudi, 1978). Several
theories, such as Uncertainty Reduction Theory (Berger & Calabrese, 1975), have

attempted to explain the above relationship.

13

According to Uncertainty Reduction Theory, a primary concern of people when
dealing with strangers is to increase their predictive power of a stranger's behavior. In
presenting their model of uncertainty reduction, Berger & Calabrese maintain that there is
a positive relationship between communication and the reduction in uncertainty, because
communication increases the amount of information we have about others. Applied here,
this would suggest that as the number of attitude items used in a study increases, the
certainty of making inferences about others would increase, resulting in stronger
interpersonal attraction judgments.

For example, consider that a person meets two strangers at the same time and
engages in a discussion of attitudes. If at the end of the discussion ten attitudes are
known about one of the strangers and only one attitude is known about the second one, it
seems reasonable to expect that there would be differential certainties about perceptions
that would be held about these strangers. More specifically, it is likely that the above
person would feel more confident and more likely to make a strong interpersonal
judgment based on the attitudes toward the stranger revealing his or her attitudes on ten
issues over the other revealing only one attitude. Given that information is an important
element in reducing uncertainty, it seems reasonable to expect that people will be more
certain and will consequently make stronger judgments about the interpersonal attraction
of others as they receive more information about them.

Another reason that more information about another's attitudes is expected to be
positively related to a person's judgment about them is with respect to consistency

theories. In the most general sense, consistency theories suggest that people seek to

14

maintain consistency across attitudes they hold. As such, people experience internal
tension when they view others similar to themselves negatively or dissimilar others
positively. In order to avoid internal tension, it is argued that people will adapt either
their attitudes or attraction judgments about others who appear similar or dissimilar in
order to maintain a consistent world View (Festinger, 1957; Heider, 1946; Osgood &
Tannenbaum, 1955). Because attitudes are somewhat enduring (Trenholm, 1989), a
person realizing inconsistency between their attitudes and interpersonal judgments is
more likely to realign their interpersonal judgments when there are discrepancies between
their attitudes about issues and interpersonal judgments of others. In the present scenario,
this means that if one has more information about another's attitudes, then they will more
likely change their interpersonal attractions as opposed to their attitudes when presented
with inconsistencies. For example, if a horse race enthusiast dislikes another person
interpersonally and subsequently discovers that the disliked other is also a horse race
enthusiast, then it is more likely that this person would realign their interpersonal
attraction judgment instead of their attitude toward horse racing. Given the notion of
consistency, it seems reasonable that people will have stronger judgments of interpersonal
attraction towards another as they receive greater amounts of information concerning
their attitudinal similarity/dissimilarity.

A final reason that more attitude items might result in a greater corresponding
interpersonal attraction judgment is with regard to the reliability phenomena. That is, as
the number of items increase in a measure, so does its subsequent reliability of that

measure. Consequently, when two variables are correlated with one another and the

15

reliability of one of those variables is increased, by definition, so will the correlation
between those two variables.

More specifically, as the number of attitude items increases, the reliability of the
attitude similarity variable increases. Therefore, if attitude similarity and interpersonal
attraction are correlated with one another, then the correlation between these variables
will be greater as the reliability of the attitude measure increases.

Because considerable variance exists in the number of attitude items used across
studies examined in this meta analysis (seven to 56) and that information quantity can be
expected to effect interpersonal judgments of others, the following hypothesis is
presented:

H2: As the number of attitude items used in a study increases so will the
correlation between attitude similarity and interpersonal attraction.

1..” ll .

Several studies (Sunnafrank, 1983, 1984; Sunnafrank & Miller, 1981) have varied
their research from the typical bogus-stranger technique employed in the attitude
similarity-interpersonal attraction research by adding the variable of interaction. For
example, Sunnafrank (1983, 1984) first had subjects complete attitude inventories.
Subsequently, attitude similar or dissimilar partners were formed and provided with the
attitude inventory completed by their partner. After reading their partner's attitude
inventory items, couples were brought together and they engaged in a 5-minute get-

acquainted interaction. Upon completion of the 5-minute interaction, interactants were

l6

separated and asked to complete measures of interpersonal attraction towards their
partner. Each of the dyads used here were sarne-sex pairs.

In another series of studies researching the effect that interaction has on the
relationship between attitude similarity and interpersonal attraction, such as Curran and
Lippold (1975) had subjects complete attitude similarity measures. Based on the
completed measures, subjects were apportioned into either attitude similar or dissimilar
conditions. Subsequently, partners were provided with sufficient fimds to go on a 30-
minute "coke date" where they were given the chance to interact with one another outside
of the laboratory setting. After their "coke date", subjects were asked to complete a
measure of interpersonal attraction toward their partner. Each of the "coke date" dyads
were opposite sex partners.

Two key elements in the above studies make them different from Byme's typical
bogus-stranger research technique. The first, is the presence of interaction. The second,
is that there were true attitude agreements or disagreements with real others.
Consequently, the question here is whether these differences will moderate the
relationship between attitude similarity and interpersonal attraction?

The first reason to expect that initial interaction will influence the effect that
attitude similarity has on interpersonal attraction is because in normal initial interactions,
attitudes are usually not uncovered. Given the lack of attitude information, it cannot be
used as a judgment of interpersonal attraction. In their work on the types of information
revealed in initial interactions, Berger & Calabrese (1975) introduced their proposed

stages of interaction. According to them, the first stage that people encounter is the entry

17

phase. In this phase, people are governed by social norms and rules, which limit
conversation to low risk topics such as demographic information. Although attitude
information may be explored in this phase, it is not information typically central to a
person and tends to concern low involvement issues.

A more specific analysis of the types of information that people expect to be
revealed in initial interactions has been conducted by Berger, Gardner, Clutterbuck, and
Shulrnan (1976). In their study, subjects read a list of 150 preselected topics and placed
them into 15 minute time blocks as to when they thought these topics would be discussed.

Their study revealed that only informational background topics were expected to be
discussed in the first 15 minutes of interaction, relatively few and innocuous attitude
topics were expected to be discussed between fifteen to thirty minutes, e.g., the Bears are
a lousy football team, and more attitude relevant items were expected to be discussed
after 30 minutes, e.g. Any American who is a communist should be deported from the
country.

Expecting that attitude similarity will not influence interpersonal attraction in the
early stages of new relationships is not a new discovery. In fact, Newcomb (1961) found
that attitude similarity had little effect on the attraction of housemates in the early stages
of their relationship, but was an important one in later stages.

A second reason to expect that interaction will suppress the effect of attitude
sirrrilarity on interpersonal attraction regards the richness of different types of
communication. In their writings on media choice, several researchers have maintained

that face-to-face interaction is the richest medium and that written communication is one

18
of the poorest (Daft, Lengel, & Trevino, 1987; Sitkin, Sutcliffe, & Barrios-Choplin,
1992). Therefore, many nonverbal cues come in to play that might affect interpersonal
attraction.

Why would the fact that face-to-face interaction is a richer form of
communication than written communication lend support to the idea that interaction will
moderate the relationship between attitude similarity and interpersonal attraction?

The first reason to expect a difference is that attitude similarity researchers have
varied the level of communication. For example, in the bogus-stranger method, the only
communication about another was the written communication (attitude inventory of
another). In other attitude similarity research, there has been the traditional written
communication and a rich face-to-face interaction.

In particular, when presented with face-to-face interactions, people perceiving that
the interaction will be potentially aversive will increase positive non-verbal cues, such as
smiling to facilitate a more pleasant interaction. In their work on interpersonal
expectancies, Ickes, Patterson, Rajecki, and Tanford (1982) provided subjects with
information about another subject and were informed that they would be asked to engage
in a future interaction with this person. In the study, subjects were lead to believe that
they would either have a future interaction with a friendly or unfriendly other. Given that
people infer that attitudinally similar others are likely to be friendly and dissimilar ones
will be unfriendly, this study appears to relate reasonably well to the variable of attitude

similarity (Byme, 1961). The results of this study revealed that when subjects interacted,

19

those in the unfriendly condition (negative expectations) smiled more at their partner than
those in the friendly condition (positive expectations).

In the second study conducted by Ickes, Patterson, Rajecki, & Tanford (1982),
subjects were provided with a host of items revealing that they were similar or dissimilar
in personality to another subject and then the pair was brought together and asked to
engage in discussion. Similar to the first study, when subjects subsequently interacted,
subjects holding negative interaction expectations (those perceived as being dissimilar in
personality) smiled at their partner with greater frequency than subjects holding positive
interaction expectancies (those being perceived as being similar in personality).

Cappella and Palmer (1990) present the most specific research on the relationship
between attitude similarity and smiling. Their findings were consistent with the idea that
attitudinally dissimilar partners smile more at one another. In particular, they found that
attitude similarity was negatively related to smiling toward one's interacting partner.

Although research is mixed on the effect that smiling has in making judgments
about others, most authors maintain that it is negatively related to the perception of others
only when it is perceived that the smiling is being used as a mask to deceive (Mehrabian,
1971). Because there is no apparent reason why people would initially expect another to
deceive them in a non-persuasive situation, such as an initial introduction conversation,
then this effect seems unlikely in the studies examined in this dissertation.

Research examining the potential positive effects of smiling in situations where
interactants approach the interaction with negative expectancies show that subjects may

use nonverbal involvement, such as smiles, to make the interaction more pleasant

20

(Rosenfeld, 1966). Consequently, if subjects leave the initial face-to-face interaction with
a pleasant feeling about the other, they might weigh this as a stronger cue to make a
judgment of interpersonal attraction than the fact that they have attitudinal differences.
Additionally, the effect of smiling in face-to-face interactions may be strong enough to
increase judgments of interpersonal attraction with dissimilar others to levels found with
similar others.

Another reason to expect that srrriling might lessen the effect of negative feelings
toward dissirrrilar others is with regards to mimicking. Dabbs (1969) found a positive
correlation between similar non-verbal behaviors such as leaning forward, touching
shoes, and fidgeting and liking. Therefore, if subjects perceive that smiling by an
attitudinally dissimilar other is mimicking their own smiling, then they would be more
interpersonally attracted to them in comparison to an attitudinally dissimilar non-
interacting other.

A third reason to expect that initial interpersonal attraction will lower the effect
that attitude similarity has on interpersonal research concerns conversational goals and
societal norms that people operate under. More specifically, in initial face-to-face
interactions, people are generally pleasant. Consequently, the positive effects of initial
interactions may work to suppress potential negative feelings that people may hold
toward attitudinally discrepant others.

In their research on initial interactions, Burleson and Denton (1992) suggested

that people engaged in initial interactions are just trying to enjoy the interaction. Since

21

most people dislike confrontation, people are likely to highlight their similarities and
downplay dissimilarities, which could potentially inflate perceived similarity by partners.

Brown and Levinson's theory of politeness (1978) addresses the norms of initial
interaction. According to Brown and Levinson, in initial interactions people seek to
establish common ground with others by selecting only safe conversational topics. If
people only discuss "safe topics" in initial interactions, they could potentially leave that
initial interaction with the belief that future interactions would follow a similar pattern.
Consequently, the polite discussion would seem likely to mitigate the effect of any
discrepant attitudinal information that people receive about others prior to an initial
interaction period.

Another way that politeness might mitigate the effect that attitude similarity has
on interpersonal attraction is through the selection of topics on which the interactants
share similar views. That is, by avoiding potential conflicts that may arise from issues
that they disagree on, interactants may inflate perceived similarity by only discussing
topics that they agree upon. If perceived similarity is inflated in the attitude dissimilar
conditions, it could eliminate any differences in perceived dissimilarity that may have
existed prior to the interaction.

The effect of initial interpersonal interaction on the relationship between attitude
similarity and interpersonal attraction can be examined in two ways. First, limited
interactions, c.g., interactions lasting fewer than 30-minutes, create situations where few
attitude issues are uncovered and even if they are uncovered, social norms cause people to

downplay these dissimilarities. Consequently, in initial interpersonal interactions, the

22

effect of attitude similarity on interpersonal attraction should be lower than when no
interaction is present, because forces are acting to mitigate the effect of attitudinal
disagreements. Therefore, the following hypothesis is presented:

H3a: Interaction between people will moderate the effect of attitude similarity
on interpersonal attraction, such that the correlation between them will
decrease.

While the effect of attitude similarity on interpersonal attraction will be attenuated
in initial interpersonal interactions due to social forces such as initial interaction norms
and the lack of attitudinal information, these factors will be overcome as interaction
increases. In particular, if Berger & Calabrese (1976) are correct in their idea that time
elapsed in a relationship will result in more attitudinal knowledge of others and if attitude
similarity is indeed an indicator of interpersonal attraction, then the effect of attitude
similarity, at least in the initial stages of interaction should be positively related to
judgments of interpersonal attraction. Therefore, the following hypothesis is presented:

H3b: The effect of attitude similarity on interpersonal attraction will increase as
time elapses in initial interpersonal interactions.

METHODS
MW
Relevant studies were obtained by initially starting with the Byme and
Sunnafrank's articles published in Communication Monographs in 1992. All relevant
articles from the bibliography of these articles were obtained and the bibliographies of

these studies were searched. The bibliography searching process continued until no

23

further germane articles were obtained. Using the method of the bibliography search
yielded a diverse number of sources, including: Communication Monographs, Human
Communication Research, Human Relations, Journal of Applied Social Psychology,
Journal of Experimental Research in Personality, Journal of Personality, Journal of
Personality and Social Psychology, Journal of Research in Personality, Journal of
Social Psychology, Psychological Reports, Social Cognition, and The Western Journal of
Speech Communication.

Additionally, the Social Science Citation Index and computer-based searches in
PSYC INFO (Psychology Literature) and ACAD (Expanded Academic Index) were
searched using the keywords of attitude similarity and attraction. The initial search
yielded 134 studies that could potentially be used in this meta-analysis. The list of
obtained studies through the process used here might not be a complete coverage of all
studies, but given the large number of studies collected, it is unlikely that a few omitted
studies would change the general conclusions of this study.

[I . . E S l I l .

First, the study had to focus on the relationship between attitude similarity and
interpersonal attraction in order to be included in this study. Examples of studies that
were excluded given the criterion established here included those measuring the effect of
personality similarity (Atkinson & Schein, 1986) or use of non-interpersonal attraction
measures as the dependent variable (Coombs & Chang, 1981). 42 of all of the studies

collected were excluded by this criterion.

24

Second, each study was required to include original data. Articles reviewing or
re-reporting data were not included in the selection process (Byme, 1992). Seven of all
of the studies collected were excluded by this criterion.

Finally, studies failing to provide sufficient information to allow computation of
the correlation between attitude similarity and interpersonal attraction was excluded. For
example, Curran's (1973) study was excluded from analysis. Although he researched
attitude similarity and interpersonal attraction, he did not provide the sufficient statistical
information necessary to reproduce the correlation coefficient. Five of all of the collected
studies were excluded by this criterion. Given the above criteria to include studies in this
meta analysis, 80 studies remained leaving 92 effect sizes for the analyses (See appendix
A).

Methods

The general meta-analytic approach used in this dissertation is that presented by
Hunter, Schmidt, and Jackson (1982). Essentially, the procedures advanced by Hunter et
a1. involves the estimation of effect sizes between the variables being examined. After
each of the effect sizes for individual studies are obtained, they are weighted by sample
size and cumulated.

To provide a standard measure across studies, each of the findings were
transformed into correlation coefficients. The primary source for transforming F values
to correlations was Hunter's FTOR program. When transforming other statistics (e. g., t-
tests) the formulas presented by Hunter et al. (1982) were used. The full meta-analyses

was performed using Hunter's ngare program. Most studies used Byme's two-item

25

measure to assess interpersonal attraction. However, the reliability of these two items
vary. Reported reliabilities are as low as .75 (Sunnafrank, 1986) and as high as .90
(Sunnafrank, 1985). The most commonly reported alpha is .85 (Byme, 1965). Given that
the mean between the highest and lowest report of alpha for these items (82.5) is
extremely close to the reliability presented by Byme, then .85 was used as the reliability
estimate for Byme's two item interpersonal attraction measure. Several studies used only
Byme's liking question to measure interpersonal attraction. For these studies,
Spearmann-Brown's prophecy formula for estimating the reliability of shortened scales
from a scale with a known reliability was employed. Using Spearmann-Brown's
prophecy formula, a reliability of .74 was calculated as the reliability for only the liking
item.

The reported reliabilities for studies not using Byme's two item interpersonal
attraction measure were employed to correct their effect sizes for attenuation due to error
of measurement. In cases where no reliability measure was presented and there was
insufficient information to calculate an estimate of the reliability measure, then perfect
measurement was assumed. Only two of the 92 effect sizes used in this study required

the assumption of perfect measurement.

26
RESULTS

To estimate the effect of attitude similarity on interpersonal attraction across all of
the effect sizes used in this meta analysis (k=92), the correlations computed for each of
the individual studies were averaged.

The overall uncorrected and weighted effect size for the relationship between
attitude similarity and interpersonal attraction is r=.46 (k=91, sd=.l9, n=10,588). When
corrected for attenuation due to error of measurement, the weighted effect size is .51
(k=91, sd=.22, n=10,588). Because corrected correlations have larger standard errors
than corrected ones, a .95 confidence interval was calculated around the uncorrected
effect size and each end was subsequently corrected for error of measurement. For these
data, the estimated confidence interval for the corrected correlation is
P(.46_<_MeanRho_<_.56)=.95 (See Table 1; for a compilation of all meta-analysis findings,
see Appendix B).

Table 1. Meta-analysis findings for all effect sizes used
K N Rho SdRho

92 10,588 .51 .20

As a final method of obtaining a general understanding of the effect that attitude
similarity has on interpersonal attraction, the interpersonal attraction mean scores were

calculated across non-interacting and interacting conditions (see Table 2).

27

Table 2. Interpersonal attraction means across studies

Mean SD K

No interaction AS 10.61 .89 51
AD 7.17 1.09 51

5 minute AS 11.07 .27 4

Interaction
AD 11.21 .09 4

30 nrinute AS 11.67 -- 1
Interaction

AD 1 0.60 -- 1
Interaction AS 1 1.39 .28 5
conditions
combined AD 1 1.00 .24 5

AS=attitude similar condition
=attitude dissimilar condition

 

Hypothesis one predicted a positive relationship between percentage of agreement
and the study effect sizes. To test the effect that manipulation strength has on the
corrected effect sizes of the studies used here, two analyses were conducted. The first
analysis was a correlation of the uncorrected effect sizes with the percentage difference
between the largest attitude similarity condition less the smallest attitude similarity

condition. For example, for a study using 12 attitude items and defining attitude

28

similarity as having nine of 12 items similar and attitude dissimilarity as having three of
12 items, the percentage of agreement was calculated. In this particular case, the
percentage of attitude similarity was calculated as .75 (9 divided by 12) and attitude
dissimilarity was calculated as .25 (3 divided by 12). As such, the percentage of
difference in this example is .50 (.75 less .25).

A second analysis testing manipulation strength was conducted by calculating the
variance between attitude conditions across studies and correlating that variance with
effect sizes. For example, for a study using 12 attitude items and defining attitude
similarity as having nine of 12 items similar and attitude dissimilarity as having three of
12 items, the variance of these conditions was calculated. In this particular case, the
percentage of attitude similarity was calculated as .75 (9 divided by 12) and attitude
dissimilarity was calculated as .25 (3 divided by 12). Each of the attitude similarity
values (.75, .25) were subtracted from the mean of the variances (.50), squared, summed,
and divided by the number of conditions. As such, the variance of these conditions is .06.

Because the results from these two tests resulted in approximately the same
finding, only the percentage test is presented. The percentage test was chosen because it
is a method that is more easily understood.

For these data, the correlation between percentage of agreement and attitude
similarity across studies and the uncorrected effect sizes is .48 and .53 when corrected for

error of measurement (k=65, n=7,282, see Table 3).

29

Table 3. Meta findings of percentage of agreement

K N Rho SdRho
Percentage of Agreement

Top Third 20 2,796 .61 .13
Middle Third 24 2,070 .53 .15
Bottom Third 21 2,416 .45 .11

There were slightly fewer studies used for the analyses here than all of the studies used
in this meta analysis (k=92), because (a) data from interaction studies (k=1 1) were not
used in this analysis and (b) some of the studies did not report data needed to calculate
strength (k=16). Interaction studies were analyzed separately, because if interaction
attenuates the relationship between attitude similarity and interpersonal attraction, then
their lower effect sizes would blur the test of the relationship between manipulation
strength and its effect size, because it is believed that the effect size in these studies will
be zero. Given the above findings, the percentage of agreement of attitude items had the
expected effect on the correlation between attitude similarity and interpersonal attraction.
II] [9.11 11308 S'

Hypothesis two stated that there would be a positive relationship between the
number of items employed in the attitude manipulation and the effect size across studies.

For these data, the correlation between number of attitude items used in a study and

effect size was .05 (k=76, See table 2).

30

There were slightly fewer studies used for the analyses here than all of the studies
used in this meta analysis (k=92), because (a) data from interaction studies (k=11) were
not used in this analysis and (b) some of the studies did not report the number of attitude
items used (k=5). A plot of attitude items with the corrected effect sizes was conducted
to determine if there is a nonlinear relationship between the number of attitude items and
effect sizes of studies (see Appendix C). The plot of attitude items with the corrected
effect sizes showed no signs of a non-linear relationship. Given the above findings, the
number of attitude items did not have the expected effect on the correlation between
attitude similarity and interpersonal attraction.

The full meta-analysis of all effect sizes used in this study had a large standard
deviation and less than 10 percent of the variance across these studies could be attributed
to sampling error. Consequently, there is evidence that there is at least one moderating
variable in these data.

Hypothesis 33 stated that interaction moderates the effect of attitude similarity on
interpersonal attraction, such that the effect size would be lower in interaction studies
than studies having no interaction. Because all of the studies (k=92) used in this meta
analysis could be coded with respect to interaction, none of them were excluded to test
hypothesis three. To test hypothesis 3,, a subgroup analysis was performed.

The subgroup analysis testing hypothesis 3a was a comparison of the corrected

and weighted effect sizes for studies in which interaction was present or absent. For these

31

data, the subgroup analysis for the no interaction studies yielded r=.58 (k=81, n=8,5 72).

In comparison, the interaction studies yielded r=.18 (k=1 1, n=2,016, see Table 4).

Table 4. Meta findings of interaction absent or present
K N Rho SdRho
No-interaction 8 1 8,572 .58 . 16

Interaction Present 11 2,016 .18 .08

Hypothesis 3,, stated that the effect of attitude similarity on interpersonal
attraction would increases in the course of an initial interpersonal interaction. To test
hypothesis 3b, a subgroup analysis breaking the interaction conditions into either initial
interaction only (5-minutes) or beyond initial interaction (approximately 30 minutes) was
calculated. For these data, the average effect size for initial interaction studies is r=.04

(k=5, n=362) and r=.21 (k=6, n=1,654, see Table 5) for beyond initial interaction studies.

Table 5. Meta findings of varying interaction levels
K N Rho SdRho

5-minute Interaction 5 362 .04 .07

30-minute Interaction 6 1,654 .21 .09

No-interaction 81 8,572 .58 .16

32

A final test to examine the effect that interaction has on the effect size of attitude
similarity/interpersonal attraction studies was conducted by coding no interaction, initial
interaction only, and beyond initial interaction studies with the values of 1 (k=81,
n=8,572), 2 (k=6, n=1,654), and 3 (k=5, n=362) , respectively. The correlation of the
new study groupings (1, 2, or 3) with the effect size is r=-.69 and r=-.73 when corrected
for error of measurement. Given the above findings, the data show that initial
interactions that are 30 or fewer minutes lessen the effect that attitude similarity has on
interpersonal attraction and that as time elapses, at least from five to 30 minutes, the
effect of attitude similarity on interpersonal attraction increases.

I l l . l B . [I l 9 . E 1.

In this meta-analysis, the effect of manipulation strength, number of attitude items
used, and interaction on interpersonal attraction were studied. To understand the
cumulative effect of percentage of agreement, number of attitude items, and presence or
absence of interaction on the effect size of studies, a regression analysis was performed
on the unweighted and corrected effect sizes of studies. For these data, Multiple R=.69
with beta=.25 (p=<.001) for percentage of agreement of the attitude similarity
manipulation, beta=.07 (p=.43) for the number of attitude items, and beta=-.70 (p=<.001)
for interaction.

A separate regression analysis was performed with the coding of interaction
studies in the three levels previously established. For this analysis, Multiple R=.70 with

beta=.25 (p=<.001) for percentage of agreement of the attitude similarity manipulation,

33

beta=.04 (p=.68) for the number of attitude items, and beta=-.73 (p=<.001) for
interaction.

The regression analyses performed here yields interesting findings. In particular,
while the correlation between manipulation strength and the effect size of interpersonal
attraction is r=.53 in the analysis without interaction studies, the effect drops to beta=.25
in the regression analysis. One reason that the effect of manipulation strength drops in
this analysis is that the interaction studies that were excluded in testing hypothesis one
were included here. As an additional test, the efiect that manipulation strength has on
effect sizes across interaction studies was conducted. While r=.53 between manipulation
strength and the corrected effect size for the no interaction studies, r=-.36 (k=7, n=539)
for the same relationship with interaction studies. The confidence interval for the
correlation between manipulation strength and effect sizes for the interaction studies is,
P=(-1.0<rme<.34)=.95.

Although there were 11 interaction studies that could be used for this analysis,
some of the studies (k=4) were not included, because there was insufficient information
to determine the manipulation strength of these studies. The extremely small number of
cases available to perform this test should lend caution to interpretation of this finding.

DISCUSSION

W. The average corrected effect studies (r=.51) for all
studies employed here showed a strong positive relationship between attitude similarity
and interpersonal attraction. Given that the correlation between the uncorrected and

corrected correlation was large (r=.99), there was no evidence that error of measurement

34

was a factor in differences across studies. One obvious reason that a large positive
relationship between the uncorrected and corrected correlation was found can be
attributed to the virtual consistent use of Byrne's two-item interpersonal attraction
measure as the dependent variable of interpersonal attraction.

The high positive correlation between manipulation strength and effect size in the
no interaction studies provides support for Byme's reinforcement affect model previously
discussed. However, given the limited number of interaction studies (k=7) that could be
used to test this idea, insufficient evidence exists to make claims about Byme's
reinforcement affect model when people interact, because there was little variance in the
manipulation strength across interaction studies (five of the seven studies had the same
variance).

WWW. Given the high correlation between the strength of the
attitude manipulation and effect sizes across studies r=.53 (when corrected for error of
measurement), there is strong evidence to suggest that perceived percentage agreement of
items has a strong influence on interpersonal attraction.

Two reasons were presented to suggest that a positive relationship would be found
between percentage of agreement and interpersonal attraction. The first reason stated that
as attitude similarity increased, so would positive affective responses toward people
holding similar attitudes, which would lead to greater interpersonal attraction. While no
direct test could be made to determine if attitudinally similar others had greater affective

responses to others in comparison to attitudinally dissimilar others, the effect of greater

35

interpersonal attraction was found. Consequently, while this model was not entirely
affirmed, present evidence is insufficient to discount it as an explanation.

The second reason that was provided to support the relationship between
percentage of agreement and greater interpersonal attraction stated that as range
restriction decreased across attitude similar conditions, there would be a greater effect on
interpersonal attraction. Given the high correlation uncovered between the percentage of
agreement and interpersonal attraction, support is provided for this explanation.

W23. The correlation between the number of attitude
items used in each study and the effect size F.05 was minimal. The evidence suggests
that the two variables are not related in a positive linear manner as hypothesized.

Two reasons were provided to support the idea that the number of attitude items
used in a study would influence the interpersonal attraction effect sizes across studies.
The first indicated that people weight the importance of issues differentially.
Consequently, the greater the number of attitude items that are used in a study, the more
likely that an important one would be found causing a more extreme judgment on
interpersonal attraction. The findings did not support the proposed relationship.

The second reason was a simple methodological explanation stating that as the
number of attitude items increased, so would the reliability of this variable. Assuming
that a relationship exists between attitude similarity and interpersonal attraction, the
greater the reliability of either of the variables would by definition result in a larger

correlation between the two variables. No evidence was found to support this rationale.

36

Given the low correlation between the number of attitude items and interpersonal
attraction effect sizes across studies one of two conclusions can be drawn. First, the
above rationales used to predict a positive linear effect between the use of a greater
number of attitude items and stronger judgments of interpersonal attraction could be
flawed. That is, the idea of issue importance and the reliability phenomenon are incorrect
in this context.

A second conclusion that can be drawn is that factors within this meta-analysis
prevented an accurate test of the relationship between the number of attitude items and
effect sizes across studies. In particular, the limited variance in number of attitude items
across studies could have prevented an accurate test of this relationship. Given strong
support that there was a weak test of the relationship between the number of attitude
items and interpersonal attraction effect sizes in the studies used for this meta-analysis,
any conclusions drawn from these findings should be taken with caution.

WW. Two tests were employed to examine the effect that
interaction has on the relationship between attitude similarity and interpersonal attraction.
Not only did the sub-group analysis of interaction present or absent in studies reveal
extreme difierences in average effect sizes (r=.18 and r=.58), but the correlation of varied
interaction levels with the effect size also showed evidence that interaction effects the
relationship between attitude similarity and interpersonal attraction (r=.04, r=.21, r=.58).
Since a majority of the studies were done without any interaction, there was a substantial
difference in the number of studies used for the comparison. However, given the extreme

differences in average effect sizes across studies, it seems reasonable to conclude that

37

interaction, at least in the early stages of relationships, lessens the effect that attitude
similarity has on interpersonal attraction.

Two ideas were tested with this hypothesis. The first is that when subjects are
presented with face to face interaction, other variables, such as interaction norms and
nonverbal cues eliminate the effect that attitude similarity has on interpersonal attraction.

Given that the effect size between attitude similarity and interpersonal attraction is
substantially lower in the interacting than the non-interacting condition, evidence appears
to support this rationale.

A second idea tested in this hypothesis is that over time, in the interaction
condition, the effect size between attitude similarity and interpersonal attraction will
increase, because factors which initially suppress the relationship between attitude
similarity and interpersonal attraction such as interaction norms and the lack of attitudinal
information will be lessened in time. As hypothesized, the effect size between attitude
similarity and interpersonal attraction increased over time. Consequently, while
communication appears to initially erase the effect of attitude sirrrilarity on interpersonal
attraction, the positive effect of attitude similarity on interpersonal attraction emerges, at
least at time periods greater than 5-minutes.

I . . . I l _ E l .

One of the limitations of this dissertation was similarity in the methods employed
by researchers to study the effect that attitude similarity has on interpersonal attraction.
Although replication of studies is useful in obtaining a more stable estimate of an effect

size between variables, the lack of variability limits the overall understanding of the

38

conditions under which the relationship is influenced. Consequently, although we have a
fairly stable estimate of the effect that attitude similarity has on interpersonal attraction in
unfamiliar bogus-stranger situations, we are unsure how these variables interact with one
another in more familiar relationships.

While some researchers have sought to branch out and explore the relationship
between attitude similarity and interpersonal attraction with respect to initial interactions
(Capella & Palmer, 1990; Sunnafrank, 1986; Sunnafrank & Miller, 1981), these studies
fail to show the long-term effects of interactions. More specifically, the rationale used to
explain why initial interactions suppress the attitude similarity to interpersonal
attractiveness relationship are based on the structure of initial conversations. The more
interesting question would be how does interaction effect long-term relationships?
Current research does not provide a sufficient base to answer this question.

A final concern of this meta-analysis is the measure used to assess interpersonal
attraction. As suggested previously, most of the studies used Byme's two-item
interpersonal attraction scale to measure the dependent variable of interpersonal
attraction. One problem with Byme's interpersonal attraction measure is that it lacks face
validity. Although the liking item seems to measure interpersonal attraction, the second
item on working does not seem an appropriate measure of interpersonal attraction. In the
work domain, there are many variables which influence a person's desire to work with
another. For example, while people value characteristics of expertise (Conrad, 1990),
organization (Daniels & Spiker, 1991), and objectivity (Hanna & Wilson, 1988) in

working relationships, these same characteristics are not necessarily important

39

components of general interpersonal attraction. Consequently, Byme's working item
seems a better measure of perceived work competency as opposed to interpersonal
attraction.

A second problem with Byme's two item measure is that it cannot be tested for
validity. With only one correlation available (the correlation between the two items) a
test of internal consistency is impossible.

Research that has been undertaken to develop a scale to measure interpersonal
attraction has been conducted by McCroskey and McCain (1974). Their scale includes
the dimensions of social, physical, and task attraction (see Appendix C). If interpersonal
attraction is indeed multi-dimensional as proposed by McCroskey and McCain, then there
is evidence that a better measure of interpersonal attraction measure should be used in
lieu of the Byme measure.

An additional benefit of the McCroskey and McCain measure is that it has enough
items to test the measure for internal consistency and parallelism since they provide at
least five measures for each dimension. Unfortunately, McCroskey and McCain do not
address the issue of internal consistency or parallelism in their interpersonal scale paper.
Wasatch

While the effect that attitude similarity has on interpersonal attraction has been
examined in over 100 studies, we have advanced little in our understanding of these
variables except that in bogus stranger situations attitude similarity has a strong effect on

interpersonal attractiveness and that initial interactions moderate the effect of attitude

40

similarity. Clearly, a research program is needed to enhance our understanding in this
domain.

One area of research that has generated considerable controversy in the attitude
similarity concerns the effect of initial interactions. Within this dissertation, it is
suggested that people are operating within particular rules and norms in initial
interactions. Future research should test whether initial interaction norms operate to
suppress the effect of attitude similarity on interpersonal attraction in newly formed
relationships.

To test the above idea, subjects could be presented with attitudinally similar or
dissimilar partners and asked to engage in initial conversations in which they either
engage in initial interactions that are typical of first time interactants or engage in a
discussion of specific attitude items.

In addition to testing types of conversations, research should also be conducted on
interactions over longer time periods than have previously been used. In particular, the
research on initial interactions have had subjects engage in conversations between five
minutes (Sunnafrank, 1985) and 30 minutes (Curran and Lippold, 1975). Although this
research has been useful in understanding the effect of initial interactions in the short
term, it does little to help us understand how attitude similarity influences interpersonal
attraction in relationships that are more enduring.

An interesting finding in this meta-analysis is that as the time of interaction
increases, so did the effect that attitude similarity has on interpersonal attraction.

Although these findings are limited to comparing the 5-minute interaction (r=.04) with

41

the 30-minute interaction (r=.21) conditions, they do add interesting insight into the
attitude similarity effect. In particular, it may suggest that the effect of attitude similarity
on interpersonal attraction in developing relationships.

To obtain a more complete understanding of the effect that attitude similarity
exerts on interpersonal attraction over time, research needs to examine interaction
relationships greater than the 30 minute maximum that has been used. This research
would determine if the attitude similarity effect on interpersonal attraction remained
stable after a certain period of time or if it fluctuated.

A second area of interest in the attitude similarity debate that has received
considerable recognition involves an alternative explanation of the traditional attitude
similarity effect on interpersonal attractiveness. As stated earlier, Rosenbaum 's (1986)
repulsion hypothesis maintains that attitude similarity does not increase interpersonal
attraction, but that attitude dissimilarity decreases interpersonal attraction. If people are
repulsed by people who are attitudinally dissimilar, then it would provide a completely
different perspective on the effect that attitude similarity has on interpersonal attraction.
To test the repulsion hypothesis, Rosenbaum replicated the traditional attitude similarity
study (which includes only attitudinally similar and dissimilar conditions) and added a no
information control condition.

In his research, Rosenbaum apportioned subjects into attitude similar, dissimilar,
or no attitude information conditions. His research revealed that there were no significant
mean differences between the attitude similar and no attitude conditions with respect to

the interpersonal attractiveness of a bogus stranger. Additionally, the results showed that

42

the attitude dissimilar others were rated a significantly less interpersonally attractive than
both the attitude similar and no attitude information conditions.

Although Rosenbaum's findings were consistent with his predictions, some
researchers (Byme, 1992) argue that it is impossible to have a no information control
condition. That is, it has been argued that in the absence of attitudinal information about
others, people assume that the anonymous person shares similar attitudes to them.

How might one go about forming a true no information control group? The first
step would be to create individuals that subjects believe that they cannot make inferences
about their attitudes. For example, a researcher might provide subjects with a description
of another in which limited information is provided, e.g., only name and non-descriptive
background information. In this case, the lack of information about another could
potentially heighten the realization that there is insufficient information to make
attitudinal predictions. Once a no inference condition is established, further testing of the
repulsion hypothesis can be undertaken.

Even in the absence of a no inference condition, the repulsion hypothesis could be
tested. If the repulsion hypothesis is true, then repulsion should be positively correlated
with the amount of time of initial interactions, because people reveal more attitudinal
information that would be perceived as repulsive to others.

Evidence to test the repulsion hypothesis given the findings of this meta-analysis
is available through an analysis of the mean scores for interpersonal attraction over time
(see Table 2). If attitude dissimilarity leads to interpersonal repulsion, then the mean

attraction scores in the attitude dissimilar condition should decrease between the 5-minute

43

and 30-minute conditions, because greater number of dissimilar attitudes should be
uncovered. While there is some evidence supporting the idea that attitudinally dissimilar
others are perceived as less attractive over time (Mad at 5-minutes = 11.21, k=4 and Mad at
30-minutes = 10.60, k=1), the limited number of studies that are available to compare
these conditions do not provide enough cases to sufficiently test the repulsion hypothesis
using the interpersonal attraction means.

A third area of research that should be conducted is testing the theories to explain
why some variables, such as interaction, influence the relationship between attitude
similarity and interpersonal attraction. For example, Sunnafrank consistently argues
across several articles that initial interaction moderates the relationship between attitude
similarity and interpersonal attraction because without interaction people feel that their
goals of achieving predictable, controllable, and stable environments are difficult to
achieve. Consequently, in non-interaction research, subjects rate attitude dissimilars as
interpersonally unattractive since they perceive them as a threat to their interaction goals
while attitude similars are rated as interpersonally attractive, because they are perceived
as interaction goal facilitators. However, once people interact, Sunnafrank argues that
people realize that they can achieve the above interpersonal goals and, consequently, the
effect of attitude similarity is eliminated.

To test Sunnafrank's theoretical explanation provided above, subjects could be
given a copy of attitude items from another individual that are either consonant or

disconsonant from their own attitudes. At this point, subjects could be asked to complete

44

a host of measures designed to measure the extent to which they feel a future interaction
with this stranger will be predictable, controllable, and stable.

Upon completion of the initial interpersonal measures, subjects could be asked to
interact with the person they received information about their attitude items. If the theory
presented by Sunnafrank is correct, results should show that attitude similarity prior to
interaction should be positively related to the belief that a future interaction with the
stranger will be unpredictable, uncontrollable, and unstable. After a discussion, the effect
of attitude similarity on the above three variables should be eliminated if Sunnafrank's
reasoning for differences in interaction versus no interaction conditions is correct.

A final area of research that has received some notice, but little attention by
researchers, involves testing the effect that interpersonal attraction has on attitude
similarity. Although some research has examined this relationship, it has been conducted
with small sample sizes and with static data.

One method that could be employed to examine the effect that interpersonal
attraction has on attitude similarity is to examine whether attitudes of relational partners
converge over time. In particular, assuming that people who date are interpersonally
attracted to one another, measures on attitude items could be collected at varying times in
new forming relationships and trends could be examined to see if attitude convergence
occurs over time.

Although convergence of attitudes over time might be confounded with other

variables, such as power, self-esteem, and importance of being in a particular

45

relationship, these variables could be controlled for and examined as variables that might
influence attitudinal convergence.
CONCLUSIONS

As early as the fourth century B.C., Aristotle suggested that friends regard the
same things as good and evil (1932). Taken literally, Aristotle would agree with Byme's
notion that attitude similarity influences interpersonal attraction or conversely, that close
relationships induce similar attitudes.

While thousands of years have passed since Aristotle's revelation, researchers are
still questioning the conditions under which the above statement is true. While this
dissertation provides one step toward sorting out what we know about the relationship
between attitude similarity and interpersonal attraction there remain many unanswered
questions.

Communication scholars interested in interpersonal communication should find
this dissertation especially interesting. In particular, it reveals that communication,
especially in initial interactions plays an important role in assessments of interpersonal
attraction.

A second reason that interpersonal communication scholars should find interest in
this dissertation is with the respect to the theoretical explanations that have been provided
to describe the effect that interpersonal interaction has on the relationship between
attitude similarity and interpersonal attraction. In particular, this dissertation provides
new explanations, such as politeness theory, and nonverbal cues displayed in

interpersonal interactions to explain why people make certain interpersonal attraction

 

46

judgments. While this dissertation does not necessarily provide a definitive explanation
of the interpersonal interaction effect on the relationship between attitude similarity and
interpersonal attraction, it at least provides more reasons to explain current findings.

In addition to providing new theoretical explanations to understand human
behavior, this dissertation has also taken the abstract significance tests applied by
individual studies and converted them into effect sizes. Consequently, future scholars can
discuss the effect that attitude similarity has on interpersonal attraction with a better
understanding of the level of effect as opposed to the vague notion that an effect exists.

While a better understanding of how attitudes are revealed and discovered within
relationships need to be examined, this dissertation provides a working framework to

conduct future research that is theoretically supported.

APPENDICES

 

APPENDIX A: Studies used for Meta-Analysis

47

 

Author D Al N I r r'
Arenson, Davis, & Jones (1983) - - 36 N .46 .50
Arrowood & Short (1973) - 12 240 N .41 .45
Bleda (1973) .60 10 90 N .40 .43
Bond, Byme, & Diamond (1968) .60 10 80 N .49 .53
Brink (1977) .42 07 72 N .38 .41
Byme (1961) - 26 34 N .88 .96
Byme (1961) - 26 84 N .66 .72
Byme (1962) - 12 112 N .64 .70
Byme, Baskett, & Hodges 1 .84 24 40 N .85 .92
(1971)
Byme, Baskett, & Hodges 2 .66 12 56 N .76 .83
(1971)
Byrne & Clore (1967) exp 1 .66 12 80 N .58 .63
Byrne & Clore (1967) exp 2 - 12 70 N .57 .62
Byme, Clore, & Grifitt (1967) .66 56 31 N .68 .74
Byrne & Ervin (1969) .60 40 176 N .34 .37
Byrne, Ervin, & Larnberth (1970) .34 10 49 N .28 .30
Byrne, Gouaux, Griffitt, .37 15 506 N .26 .28

Lamberth,Murakawa,
Prasad, Prasad,
& Ramirez (1971)

Byme, Griffitt, Hudgins,
& Reeves 1 (1969)

Byme, Griffitt, Hudgins,
& Reeves 2 (1969)

Byme & Griffitt (1966)
Byme & Griffitt (1966)

Byme, Griffitt, & Golightly l
(1966)

Byme, Griffitt, & Golightly 2
(1966)

Byme, London, & Griffitt
(1968)

Byme, London, & Reeves (1968)
Byme & Nelson (1965)

Byme & Rhamey (1965)
Cappella & Palmer (1990)

Cherry, Byme, & Mitchell
(1976)

Clore & Baldridge (1970)
Clore & Gormly (1974)
Condon & Crano (1988)
Curran & Lippold 1 (1975)
Curran & Lippold 2 (1975)
Davis (1984)

DeWolfe & Jackson (1984)

1.0

.49

.50

.66

.66

.50

.66

.67

1.0

.50

1.0

.50

1.0

1.0

48

08

12

12

07

12

12

14

12

12

36

12

12

l2

12

30

30

15

84

82

260

30

85

55

80

205

168

180

80

192

64

48

226

588

196

60

70

Z

2 <1 2 Z Z

Z Z < -< Z Z Z

.60

.34

.21
.57

.44

.63

.53

.50
.56
.65
.39

.61

.50
.42
.54
.15
.01
.88

.50

.65

.37

.23

.62

.48

.68

.58

.54

.61

.71

.43

.66

.54

.46

.59

.16

.01

.96

.68

 

Erwin (1981)
Erwin (1982)
Franklin (1971)

Gonzales, Davis, Loney,
Lukens, & Junghans (1983)

Good & Good (1972)

Good & Nelson (1971)

Gormly & Clore (1969)

Gouaux (1971)

Gouaux & Summers (1973)
Hoyle (1993)

Insko & Wetzel (1974)

Jackson & Mascara (1971)
Jarnieson, Lydon, & Zanna (1987)
Johnson (1974)

Johnson & Johnson (1972)
Kaplan (1972)

Kaplan & Olczak (1970)

Kleck & Rubenstein (1975)
Layton & Insko (1974)

Lydon, Jarnieson, & Zanna (1988)

McGinley (1 980)

.50

.85

1.0

.66

.60

.60

1.0

.50

.67

.60

.61

.50

.50

1.0

.69

1.0

.80

49

12

12

O7

20

12

12

10

2O

12

12

10

10

18

12

12

12

15

08

10

2O

50

48

120

270

165

142

96

110

228

82

128

108

48

80

32

80

135

41

320

81

183

Z Z Z Z

Z Z Z -< -< Z Z Z Z 2 Z Z Z Z Z Z Z

.34

.38

.58

.73

.56

.41

.59

.68

.41

.75

.61

.52

.79

.21

.64

.67

.28

.20

.48

.41

.45

.37

.41

.63

.79

.61

.45

.64

.74

.45

.82

.66

.57

.86

.28

.86

.73

.30

.22

.52

.55

.49

McGinley, Nicholas, & McGinley
(1978)

McGinley & Reiner (1979)
Olczak & Goldman (197 5)
Orpen (1984)

Palmer & Kalin (1985)

Pandey & Rastogi (1978)
Posavac & Pasko (1971)
Rosenbaum (1986)

Sachs (1975) exp 1

Sachs (1975) exp 2

Santee (1976)

Scott (1973)

Shaikh & Kanekar (1993)
Shuntich (1976)

Singh (1973)

Singh (1974)

Singh (1975)

Smeaton, Byme, & Murmen (1989)
Smith, Meadow, & Sisk (1970)

Stroebe, Insko, Thompson,
& Layton (1971)

Sunnafrank (1983)

.80

.80

.60

.72

.50

.66

1.0

1.0

.70

1.0

1.0

.60

.60

.60

1.0

.88

.66

1.0

1.0

50

20

10

15

18

O7

09

08

12

12

12

20

12

15

10

10

10

03

12

08

02

165

190

40

614

271

32

81

168

60

90

112

73

76

54

75

9O

30

59

44

180

53

Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z —< Z Z

.50

.57

.60

.27

.48

.70

.68

.54

.30

.43

.64

.84

.90

.64

.68

.64

.68

.38

.53

.76

.65

.54

.62

.65

.29

.52

.76

.74

.59

.33

.47

.70

.91

1.21

.70

.74

.70

.74

.41

.58

.83

.71

51

Sunnafrank (1983) 1.0 02 98 Y .00 .00
Sunnafrank (1984) 1.0 02 67 N .64 .71
Sunnafrank (1984) 1.0 02 98 Y .00 .00
Sunnafrank ( 1985) 1.0 02 64 N .45 .49
Sunnafrank (1985) 1.0 02 60 Y .00 .00
Sunnafrank 1 (1986) 1.0 02 52 N .74 .80
Sunnafrank 2 (1986) 1.0 02 50 Y .18 .20
Sunnafrank & Miller (1981) 1.0 02 58 N .73 .79
Sunnafrank & Miller (1981) 1.0 02 56 Y .11 .12
Tesch, Huston, & Indenbaum 1.0 32 100 N .73 .79
(1973)
Tesser (1971) .33 06 22 N .72 .78
Touhey (1974) .84 12 90 N .62 .67
Williams, Ryckman, Gold, .60 10 80 N .78 .85
& Lenney (1982)
Yabrudi & Diab (1978) exp 1 - 08 40 N .75 .82
Yabrudi & Diab (1978) exp 2 - 08 40 N .83 .90

I = Initial interaction

AI = Number of attitude items

D = Percentage difference between most similar and dissimilar attitude conditions
N = Sample size

r = Uncorrected correlation

r' = Corrected correlation

52

APPENDIX B: Overall Effect Size and Subgroup Analysis

K N Rho SdRho

Overall Meta-analysis 92 10,5 88 .51 .20
No-interaction 81 8,5 72 .5 8 . l 6
Interaction Present 11 2,016 .18 .08
5-minute Interaction 5 362 .04 .07
30-minute Interaction 6 1,654 .21 .09

Percentage of Agreement
Top Third 20 2,796 .61 .13
Middle Third 24 2,070 .53 .15

Bottom Third 21 2,416 .45 .11

53

APPENDIX C: Plot of Number of Attitude Items with Effect Sizes

 

 

 

 

9 2 l
1 1 1
222
l l 1 1 1
21 135 2 1 1
Effect .6 2124 1
Sizes 1 14 1
1112 l
1 35 1
1 3 1
3 12 1
1
1 1 1 1 1
1
O 3 1
8 24 40 56
0 16 32 48
Number of Attitude Items

Appendix C. Plot of the number of attitude items employed in individual studies with
their reported effect size between attitude similarity and interpersonal attraction.

54

APPENDIX D: McCroskey & McCain Attraction Measure

Social Attraction

SAPP’Ni"

I think he (she) could be a friend of mine.

It would be difficult to meet and talk with him (her). *

He (she) just wouldn't fit into my circle of friends. *

We could never establish a personal friendship with each other. *
I would like to have a friendly chat with him (her).

Physical Attraction

9°>'.°‘.V':“P’!":"

I think he (she) is quite handsome (pretty).

He (she) is very sexy looking.

I find him (her) very attractive physically.

I don't like the way he (she) looks. *

He (she) is somewhat ugly. *

He (she) is not very good looking. *

He (she) wears neat clothes.

The clothes he (she) wears are not becoming. *

Task Attraction

.V‘PWNI"

He (she) is a typical goof-off when assigned a job to do. *
I have confidence in his (her) ability to get the job done.
If I wanted to get things doc I could probably depend on him (her).
I couldn't get anything accomplished with him (her). *
He (she) would be a poor problem solver. *

* reflected items

LIST OF REFERENCES

LIST OF REFERENCES

* Indicates that study was used in the meta-analysis.

Aristotle. (1932). We (L. Cooper, Trans, pp. 37-108). New York:
Appleton-Century-Crofts. (Original work published 330 BC.)

*Aronson, J .S., Davis, M.C., & Jones, LC. (1983). Sequential effects of
attitudinal stimuli. WM, 112, 257.260.

*Arrowood, A.J., & Short, J .A. (1973). Agreement, attraction, and self-esteem.
Canadianlcumalnfflehavimaisgensc, 5, 242-252.

Atkinson, D.R. & Schein, S. (1986). Similarity in counseling. Iheflgmsfling
Eastman. 14, 319-357.

Berger C.R. & Calabrese, R.J. (1975). Some explorations in initial interactions
and beyond: Toward a developmental theory in interpersonal communication. Hmnan
Commumcatmrifiesearch, 1, 99-112.

Berger, C.R., Gardner, R.R., Clutterbuck, B.W., & Shulman, LS (1976).
Perceptions of information sequencing in relationship development. Hmnan
Commnmcatronfiescarch, 3, 1976.

Berscheid, E., & Walster, E. ( 1983). MW. Reading, MA:

Addison-Wesley.

*Bleda, PR. (1973). Attitude similarity-dissimilarity and attraction in the middle

eastern culture. IhelcumaLdLSccialehclcgy, 21, 153-154.

55

56

Bochner, AP. (1991). On the paradigm that would not die. In J .A. Anderson
(Ed.), Communicationlearbcok, 14 (p484-491). Newbury Park, CA: Sage.

*Bond, M., Byme, D., & Diamond, M.J. (1968). Effect of occupational prestige
and attitude similarity on attraction as a function of assumed similarity of attitude.

W23, 1167-1172.

Bowman, C.H., & Fishbein, M. (1978). Understanding public reaction to energy
proposals: an application of the Fishbein model. lmunalanpnliedSocialehologx,
8, 319-340.

*Brink, J .H. (1977). Effect of interpersonal communication on attraction. Journal

officrsonahtxandSocralfisxcholch 35., 783-790.

Brown, B.J., & Levinson, S. (1978). Universals in language usage: Politeness
phenomena. In E.N. Goody (Ed.), Quosuonsandpolitonoss (pp. 56-289). Cambridge:
Cambridge University Press.

Burleson, B.R., And Denton, W.H. (1992). A new look at similarity and
attraction in marriage: Similarities in social-cognitive and communication skills as
predictors of attraction and satisfaction. CommunioationMonogmphs, 52, 268-287.

Byme, D. (1992). The Transition from controlled laboratory experimentation to
less controlled settings: Surprise! Additional variables are operative. Communication
Monographsifl, 190-198.

Byme, D. (1969). Attitudes and attraction. In L. Berkowitz (Ed.), Aduanoesjn
ExperimentaLBmholcax

(pp. 178-224). Academic Press: New York and London.

57

*Byme, D. (1961). Interpersonal attraction and attitude similarity. Loumulof
Ahaonnalanifiociaifimholw, 62, 713-715.

*Byme, D. (1961). Interpersonal attraction as a function of affiliation need and
attitude similarity. W8, 3., 233-289-

Byrne, D. (1971). Ihootuaouonparadigm. New York: Academic Press.

*Byrne, D., Baskett, G.D., & Hodges, L.H. (1971). Behavioral indicators of

interpersonal attraction. JoumaloflApolieiSocjaLPsychology. 1, 137-149.

*Byrne, D., & Clore, G.L. (1967). Effectance arousal and attraction. Journalof
WW, 6, 1-18.

*Byrne, D., Clore., G.L., & Griffitt, W. (1967). Response discrepancy versus
attitude similarity-dissimilarity as determinants of attraction. PsychonomonioSQjonco, 1,
397-398.

Byme, D., Clore, G.L., & Smeaton, G. (1986). The attraction hypothesis: Do
similar attitudes affect anything? WW, 6, 1167-
1170.

*Byme, D., & Ervin, CR. (1969). Attraction toward a negro stranger as a
function of prejudice, attitude similarity, and the stranger's evaluation of the subject.
HummRolauons, 22, 397-404.

*Byrne, D., Ervin, C.R., & Lamberth, J. (1970). Continuity between the
experimental study of attraction and real-life computer dating. Joumaloffiersonalituand

Watchman. 16, 157-165.

58
*Byrne, D., Gouaux, C., Griffitt, W., Lamberth, J ., Murakawa, N., Prasad, M.,
Prasad, A., & Ramirez, M. (1971). The ubiquitous relationship: attitude similarity and
attraction. HumanRelau'ons, 24, 201-207.
Byme, D. & Griffitt, W. (1973). Interpersonal attraction. Annualfieyjmof

13mm, 25, 317-336.

*Byme, D. & Griffitt, W. (1966). A Developmental Investigation of the Law of
Attraction. ,LoumaLoflRersonalmLondSooioLBsyghology, 4, 699-702.

*Byrne, D., & Griffitt, W. (1966). Similarity versus liking: a clarification.
W, 6, 295-296.

*Byrne, D., Griffitt, W., & Golightly, C. (1966). Prestige as a factor in
determining the effect of attitude similarity-dissimilarity on attraction. lournalof
Borsonuljty, 34, 434-444.

*Byrne, D., Griffitt, W, Hudgins, W., & Reeves, K. (1969). Attitude similarity-
dissirnilarity and attraction: generality beyond the college sophomore. IhoJoumaLof
SocialPsxchclch, 19, 155-161.

*Byrne, D., London, 0., & Griffitt, W. (1968). The effect of topic importance
and attitude similarity-dissimilarity on attraction in an intrastranger design.
BsxchcncmcnicScienss, 11, 303-304.

*Byrne, D., London, 0., & Reeves, K. (1968). The effects of physical

attractiveness, sex, and attitude similarity on interpersonal attraction. JournaLof

Bermalitx, 36, 259-271.

59

*Byrne, D., & Nelson, D. (1965). Attraction as a linear function of proportion of

positive reinforcements. WW, 6, 659-663.

*Byrne, D. & Rhamey, R. (1965). Magnitude of positive and negative
reinforcements as a determinance of attraction. Momma
Psychology, 2, 884-889.

*Cappella, J .N., & Palmer, M.T. (1990). Attitude similarity, relational history,
and attraction: The mediating effects of kinesic and vocalic behaviors. Communioation

Monographs, 51,161-183.

*Cherry, F ., Byme, B., & Mitchell, HE. (1976). Clogs in the Bogus Pipeline:

Demand Characteristics and Social Desirability. JoumaLoflRosearohjnPersonaliu, 19,
69-75.

Cialdini, RB- (1990). WWW—0mm New
York: Williams Morrow.
*Clore, G.L., & Baldridge, B. (1970). The behavior of item weights in attitude-

attraction research. LoumalnfExnerimentaLSocialPsxchclogx 6, 177-186.

*Clore, G.L., & Gormly, J .B. (1974). Knowing, feeling, and liking: A

psychophysiological study of attraction. Wu, 8, 218-230.
*Condon, J .W., & Crano, W.D. (1988). Inferred evaluation and the relationship
between attitude similarity and interpersonal attraction. WWW
251121191932, 54, 789-797. \
Conrad, C. (1990). WWW. Holt, Rinehart, &

Winston, Inc.

6O

Coombs, L.C. & Chang, M. (1981). Do husbands and wives agree? Fertility

attitude and later behavior. Ropulationondfinyironmont, 4, 109-127.

Curran, J .P. (1973). Examination of various interpersonal attraction principles in

the dating dyad. IcumaLQfLExnetimentalResearchinllersonalirx 6. 347-3 56.

*Curran, J .P. & Lippold, S. (1975). The effects of physical attraction and attitude

similarity on attraction in dating dyads. W, 43, 528-539.

Dabbs, J .M. (1969). Similarity and gestures and interpersonal influence.

on “cr' 0 .1) 0.41.2, orV‘ruo r‘éu‘u 9| ' Ho- -. a o '-u',4

—9

337-338.

Daft, R.L., Lengel, R.H., & Trevino, L.K. (1987). Information richness: A new
approach to managerial behavior and organizational design. Wild
Bfihaxim, 6,191-233.

Daniels, T.M., & Spiker, B.K. (1991). W
oommunioation. Wm. C. Brown Publishers.

*Davis, J .M. (1984). Attraction to a group as a function of attitude similarity and
geographic distance. Sooialfiehaxionandjorsonaliu, 12, 1-6.

*De Wolfe, T.E., & Jackson, LA. (1984). Birds of a brighter feather: Level of
moral reasoning and similarity of attitude as determinants of interpersonal attraction.

PsychologicaLRemrts, 54, 789-797.

*Erwin, PG. (1981). The role of attitudinal similarity and perceived acceptance

evaluation in interpersonal attraction. Iholoumaloflflsyohology, 199, 133-136.

61

*Erwin, PG. (1982). The role of attitudinal similarity and direct acceptance
evaluations in attraction. Ihoimnnaloffisyohology, 111, 97-100.

Festinger, L. (1957). Ameomofsognmueorssonanoe. Stanford, Ca: Stanford
University Press.

Fishbein, M., & Ajzen, I. (1975). MW.
Reading, MA: Addison-Wesley.

Fishbein, M., & Ajzen, I. (1972). Attitudes and opinions. AnnuaLReyieuLof
Bsxehcloax, 23, 487-554.

*Franklin, B.J. (1971). Attitude similarity-dissimilarity, dogmatism, and
interpersonal attraction. Bmhology, 31, 4-11.

*Gonzales, M.H., Davis, J .M., Loney, G.L., LuKens, C.K., & Junghans, OM.
(1983). Interactional approach to interpersonal attraction. WM
SociaLBsxcholch.6, 1192-1197.

*Good, L.R., & Good, KC. (1972). Role of vindication motivation in the attitude
similarity-attraction relationship. EsyohologioaLRoports, 31, 769-770.

*Good, L.R., & Nelson, DA. (1971). Effects of person-group and intragroup
attitude similarity on perceived group attractiveness and cohesiveness. Esyohonomonio
Soionoo, 23, 215-217.

*Gorrnly, A.V., & Clore, G.L. (1969). Attraction, Dogmatism, and attitude
similarity-dissimilarity. JoumaLoflExporimerualRosoamhinfiersonality, 4, 9-13.

*Gouaux, C. (1971). Induced affective states and interpersonal attraction.

JounalnfEerscnahnLarrdSncialfisxchclogx 20, 37-43.

62

*Gouaux, C., & Summers, K. (1973). Interpersonal attraction as a function of

affective state and affective change. loumaloflResearohjnflersonamy, 1, 254-260.
Griffitt, W.B. (1969). Personality similarity and self-concept as determinants of
interpersonal attraction. MW, 18, 137-146.
Hanna, M.S., & Wilson, G.L. (1988). WWW
Egofossronalfieumgs. Random House: New York.

Heider, F. (1946). Attitudes and cognitive organization. Lmnnaloflflsyohology,
21, 107-112.

*Hoyle, RH. (1993). Interpersonal attraction in the absence of explicit attitudinal

information. SooioLQogniLion, 11, 309-320.

Hunter, J .E., & Schmidt, FL. (1990). Methcdmfmetazanalxsis. Newbury Park,
CA: Sage.

Hunter, J.E., Schmidt, F .L., & Jackson, GB. (1982). W
W. Beverly Hills, CA: Sage.

Ickes, W., Patterson, D., Rajecki, T., & Tanford, S. (1982). Behavioral and

cognitive consequences of reciprocal versus compensatory responses to preinteraction

expectancies. W11, 1, 160-190.

*Insko, C.A., & Wetzel, C. (1974). Preacquaintance attraction as an interactive

function of the proportion and number of similar attitudes. RmsontafiueRosearohjn

SociaLEsxcthch, 5, 27-33-

63
*Jackson, L.A., & Mascaro, GM. (1971). Interpersonal attraction as a function of
attitude similarity dissimilarity and attitude extremity. Psyohonomonioficienoe, 23, 187-
188.
*Jamieson, D.W., Lyndon, J .E., & Zanna, MP. (1987). Attitude and activity
preference similarity: Differential bases of interpersonal attraction for low and high self-

monitors. loumachBerscnalitxandficcialehologx 53. 1052-1060.

*Johnson, CD. (1971). Competence motivation and interpersonal evaluation.
Bulletmnffisxchoncmcmcmm, 4, 199-200.

*Johnson, J ., & Johnson, D.W. (1972). The effects of other's actions, attitude
similarity, and race on attraction toward others. HumanRolations, 2, 121-130.

*Kaplan, M.F. (1972). Interpersonal attraction as a function of relatedness of
similar and dissimilar attitudes. WW, 6, 17-
2].

*Kaplan, M.F., & Olcsak, P.V. (1970). Attitude similarity and direct

reinforcement as determinants of Attraction. loumalofExneximomaLRosmhjn
Personalim, .4. 186-189-

*Kleck, R.E., & Rubenstein, C. (1975). Physical attractiveness, perceived
attitude-similarity, and interpersonal attraction in an opposite-sex encounter. Joumalof
Whom, 31, 107-114.

*Layton, B.D., & Insko, CA. (1974). Anticipated interaction and the sirnilarity-

attraction effect. Socjomouy, 2, 149-162.

64

*Lydon, J .E., Jamieson, D.W., & Zanna, MP. (1988). Interpersonal similarity
and the social intellectual dimensions of first impressions. SooiaLCogniuon, 4, 269-286.
McCroskey, J .C., & McCain, T.A. (1974). The measurement of interpersonal

attraction. SneoonMonogrophs, 41, 261-266.

*McGinley, H. (1980). A test for artifactual effects in an attitude
similarity/ interpersonal attraction study. BmleunofjhofisychonomonioSooim, 16,
137-139.

*McGinley, H., Nicholas, K., & McGinley, P. (1978). Effects of body position
and attitude similarity on interpersonal attraction and opinion change. Euchologioal
Remrts.42,127-138.

*McGinley, H., & Reiner, M. (1979). Contingency awareness and interpersonal
attraction. BullofinofjhofisyohonomonioSoojofl, 13, 175-178.

Mehrabian, A. (1971). A nonverbal betrayal of feeling. loumaloflfixmrimontal
Researchmflersonahm 5. 64-73-

Newcomb, TM. (1961). 111W. New York: Holt, Rinehart,

& Winston.

*Olczak, P.V., & Goldman, J .A. (1975). Self-actualization as a moderator of the
relationship between attitude similarity and attraction. ILLLQumalofEsuchology, 82,
195-202.

*Orpen, C. (1984). Attitude similarity, attraction, and decision-making in the

employment interview. IholonmaLofiBsychology, 111, 111-120.

65

Osgood, C.E., & Tannenbaum, RH. (1955). The principle of congruity in the

prediction of attitude change. W, 62, 42-5 5.

*Palmer, D.L., & Kalin, R. (1985). Dogmatic responses to belief dissimilarity in
the "bogus stranger" paradigm. WM, 48, 171-
179.

*Pander, J ., & Rastogi, R. (1978). Intolerance of ambiguity and response to
attitude similarity-dissimilarity. Raucholgia, 21, 104-106.

Parks, M.R., & Adelman, MB. (1983). Communication networks and the
development of romantic relationships: An expansion of uncertainty reduction theory.
HumanCommnnieatienR:s:areh,10, 55-79.

Petty, R.E., & Cacioppo, J .T. (1979). Issue involvement can increase or decrease
persuasion by enhancing message-relevant cognitive responses. Joimaloflfiemnumy
MW, 31,1915-1926.

Petty, R.E., Cacioppo, J.T., & Goldman, R. (1981). Personal involvement as a
determinant of argument-based persuasion. W
Bsxahelegxa 41, 847-855.

Petty, R.E., Cacioppo, J .T., & Schumann, D. (1983). Central and peripheral

routes to advertising effectiveness: The moderating role of involvement. loumaLof

Censumenliesaareh, 10, 135-136.
*Posavac, E.J., & Pasko, SJ. (1971). Interpersonal attraction and confidence of

attraction ratings as a function of number of attitudes and attitude similarity.

BsxehenemenieSeiane: 23, 433-435-

66

*Rosenbaum, ME. (1986). The repulsion hypothesis: On the nondevelopment of

relationships. loumaloflflersonflimaniSocjalfiychology, 31, 1156-1166.

Rosenfeld, HM. (1966). Approval-seeking and approval-inducing functions of
verbal and nonverbal responses in the dyad. Lmnnalofiflersonalflandfiooial
Embelegx. 4., 597-605.

*Sachs, DH. (1975). Belief similarity and attitude similarity as determinants of
interpersonal attraction. ,IomnaloflRejoarohjnPersonality, 2, 57-65.

*Santee, RT. (1976). The effect on attraction of attitude similarity as information

about interpersonal reinforcement contingencies. Sooiomeuy, 39, 153-156.

Schwartz, J.M., & Smith, WP. (1976). WWW
12mm, 3.4, 1268-1275.

*Scott, WC. (1973). The linear relationship between interpersonal attraction and
similarity: An analysis of the "unique stranger" technique. WW
Bentham, 21,117-125.

*Shaikh, T., & Kanekar, S. (1993). Attitudinal similarity and affiliation need as
determinants of interpersonal attraction. MW, 134, 257-
259.

*Shuntich, R.J. (1976). Some effects of attitudinal similarity and exposure of
attraction and aggression. Jounmoffiesoarohjnfiersonamy, 10, 155-165.

Sitkin, S.B., Sutcliffe, K.M., & Barrios-Choplin, JR. (1992). A dual-capacity
model of communication media choice in organizations. Humanfiommunioation

Research, 1.8. 563-598.

67

Simons, H.W., Berkowitz, N.N., & Moyer, J. (1970). Similarity, credibility, and

attitude change. W, 13, 1-l6.

*Singh, R. (1973). Attraction as a function of similarity in attitudes and

personality characteristics. IhelmrmaloflSooialPsyohology, 21, 87-95.

*Singh, R. (1974). Reinforcement and attraction: Specifying the effects of

affective states. Resoarohjnfiarsonalm, 8, 291-305.

*Singh, R. (1975). Reinforcement, affect, and interpersonal attraction.
Bsyohologia, 18, 142-148.

Sitkin, S.B., Sutcliffe, K.M., & Barrios-Choplin, J .R. (1992). A dual-capacity
model of communication media choice in organizations. Humanflommunioation
Researeh, 8, 563-598.

*Smeaton, G., Byme, D., & Murmen, SK. (1989). The repulsion hypothesis
revisited: Similarity irrelevance or dissimilarity bias? loumalofiPoLsonal'mLaniSocial
Psychology, 36, 54-59.

*Smith, R.E., Meadow, B.L., & Sisk, T.K. (1970). Attitude similarity,
interpersonal attraction, and evaluative social perception. Bsychonomonioficionco, 18,
226-227.

Stephen, TD. (1985). Fixed-sequence and circular-causal models of relationship

development: Divergent views on the role of communication in intimacy. Journalof

MauragaansLthaEamilx 41, 955-963.

68

*Stroebe, W., Insko, C.A., & Layton, B. (1971). Effects of physical
attractiveness, attitude similarity, and sex on various aspects of interpersonal attraction.
WW3! 18, 79-91.

*Sunnafrank, M. (1983). Attitude similarity and interpersonal attraction in
communication processes: In pursuit of an ephemeral influence. Communioauon
Monographs, 50, 273-284.

*Sunnafrank, M. (1984). A communication-based perspective on attitude
similarity and interpersonal attraction in early acquaintance. Communication
Monographs, 31, 372-380.

*Sunnafrank, M. (1985). Attitude sinrilarity and interpersonal attraction during
early communicative relationships: A research note on the generalizability of findings to
opposite-sex relationships. Wasnmloumfloffimoohflommunioation, 42, 73-80

*Sunnafrank, M. (1986). Communicative influences on perceived similarity and
attraction: An expansion of the interpersonal goals perspective. W

Saaaehflemmum'eatien 512. 158-170.

Sunnafrank, M. (1992). On Debunking The Attitude Similarity Myth.
CemmunieatiQnMenegraphs, 59, 164-179.

*Sunnafrank, M. & Miller, GR. (1981). The role of initial conversations in
determining attraction to similar and dissimilar strangers. Humanflommunioation

Research, 8, 16-25.

69

Tan, D.Y., & Singh, R. (1995). Attitude and attraction: A developmental study of
the similarity-attraction and dissimilarity-repulsion hypotheses. Remnalflondfiooial
W11, 975-986.

*Tesch, F.E., Huston, T.L., & Indenbaum, EA. (1973). Attitude similarity,
attraction, and physical proximity in a dynamic space. JournalofiAoplflSociol
Bayohology, 3, 63-72.

*Tesser, A. (1971). Evaluative and structural similarity of attitudes as
determinants of interpersonal attraction. lorunalofllersonalitmriSooiolfisuchology,
18, 92-96.

*Touhey, J .C. (1974). Situated identities, attitude similarity, and interpersonal
attraction. Sooiomouy, 31, 363-374.

Trenholrn, S. (1989). PorsuaslonongLSochnfluonoe. Prentice Hall, Englewood
Cliffs, NJ.

Warshaw, PR. (1980). Disentangling behavioral intention and behavioral
expectation. JournalofiExnorimontaLSoojalflsyohology, 21, 153-172.

*Williams, S., Ryckman, R.M., Gold, J .A., & Lenney, E. (1982). The effects of
sensation seeking and misattribution of arousal on attraction toward similar and dissimilar
strangers. loemalefkesaarehmfiersenahnr. 18, 217-226.

*Yabrudi, P.F., Diaz, & Lufiy, L.N. ( 1978). The effects of attitude sirrrilarity-

dissinrilarity, religion, and topic importance on interpersonal attraction among lebanese

university students. Iholoumaloffisyohology, 106, 167-71.

H GQN STAT

HIC I E UNIV. LIBRRRIES
llHlllllllllllllllllllllllllllllllllllllUIHIUIWI
31293015592540