Group Title: effects on credibility, attraction and attitude of reward and distraction through violations of personal space expectations /
Title: The effects on credibility, attraction and attitude of reward and distraction through violations of personal space expectations /
Full Citation
Permanent Link:
 Material Information
Title: The effects on credibility, attraction and attitude of reward and distraction through violations of personal space expectations /
Physical Description: viii, 224 leaves : ; 28 cm.
Language: English
Creator: Stacks, Don Winslow, 1949-
Publication Date: 1978
Copyright Date: 1978
Subject: Personal space   ( lcsh )
Interpersonal relations   ( lcsh )
Speech thesis Ph. D   ( lcsh )
Dissertations, Academic -- Speech -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 213-223.
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Don Winslow Stacks.
 Record Information
Bibliographic ID: UF00098651
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000085692
oclc - 05341559
notis - AAK1045


This item has the following downloads:

PDF ( 7 MBs ) ( PDF )

Full Text







During a time of great unrest several significant others put

forth great effort to quell feelings of inadequacy and apprehension;

for this I am thankful and appreciative.

To Dr. Judee K. Burgoon, a special note of thanks is due. I

am thankful for her gentle questioning and probing of thoughts

which just would not come out. What I am today professionally is

to a large extent a reflection of her professionalism and interest.

I am proud to say that I was her student.

To Dr. Michael Burgoon, to whom I feel both awe and friendship,

I owe my start. It has been a long path but his guidance and

advice have been of great importance tc me. Who would have thought

that I would have studied his area?

To Dr. Robert Scholes I wish to thank for his ever ready ear

over the last nine months. I also wish to acknowledge his interest

in my area and the ones we share together.

To Dr. Marvin Shaw, who taught me a new meaning for the word,

"gentleman," I thank him for his time and effort and criticism.

I would like to thank Dr. Norman Markel for his critique of

earlier work and counsel during times which we both would like to

forget but cannot.

A special note of appreciation and thanks go to my other half,

Robin. Without her not so gentle prodding and her commitment to

me this may not have been completed. To her I say, the future

lies ahead.

Finally, a note of thanks to Kris Arbuthnot and Gary Steinberg,

without whose help the logistics of this study may have kept it

in the conceptual stage forever.


ACKNOWLEDGEMENTS. .. . . . . . . . ... ii

ABSTRACT. . . . . . . . . . . . . i

Review of Literature . . . . . 2
Factors Influencing Distancing Norms. . 4
Interactant Variables. . . . . 5
Interactional Variables. . . ... 12
Environmental Variables. . . .. 13
Effects of Violating Interpersonal
Distancing Expectations . . . . 15
Theoretical Models . . . . 15
Research Paradigms . . . . . 17
Distraction and Violations of Inter-
personal Distancing Norms . . . 28
Derivation of Hypotheses . . . . . 40

CHAPTER II: METHOD . . . . . . . . . 44


RESULTS. . . . . . . . . .
Manipulation Checks. . . . . .
Hypothesis 1 . . . . . ....
Hypotheses 2, 3 and 4. . . . .
Supplemental Analyses on Hypotheses
1 4. . . . . . . . .
Hypothesis 5 . . . . . . .
Hypothesis 6 . . . . . . .
Hypothesis 7 . . . .......
Hypotheses 8 and 9 . . . . .

Propensity to Counterargue . .
Distraction. . . . . . .
Violations of Personal Space
Expectations . . . . .
Implications and Suggestions for
Future Research. . . . .
Summary. . . . . . . .

. . 86
. . 87
. . 89

. . 90

. . 93
. . 97





APPENDIX B: MESSAGE. . . . . . . . .




REFERENCES . . . . . . . . . . . .

BIOGRAPHICAL SKETCH. . . . . . . . . . .







Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy



Don Winslow Stacks

August, 1978

Chairman: Judee K. Burgoon
Major Department: Speech

The underlying rationale for the effects of distraction in a

communicative setting is based upon the disruption of an indivi-

dual's ability to counterargue. While research investigating the

distraction effect has generally relied upon environmental

disruptors, this investigation attempted to test the distraction

available to a persuader within the communicative setting through

the violation of personal spacing expectations. Additionally,

research has attempted to examine an individual's orooensity to

counterargue a message when distracted and the results of the

distraction on the source's credibility. Such research predicts

that high propensity-to-counterargue individuals will be more

resistant to both the message and source than will low propensity-

to-counterargue individuals. This investigation sought to: 1)

establish the relationship between violations of personal spacing

expectations as a distractor; 2) to test the relationship between

propensity to counterargue and message acceptance, source

credibility and attraction as mediated by both the distractor

and the individual's propensity to counterargue; and 3) to expand

previous research on violations of personal space expectations

to persuasive attempts.

The results of multiple regression analyses offered limited

support for the predicted relationships among reward, distance,

propensity and the communication outcomes of credibility, attraction

and persuasion. Two dimensions of credibility and one dimension

of attraction conformed to the predicted curvilinear relationships

when subjects interacted with a rewarding initiator who did or did

not violate personal spacing expectations; when interacting with a

punishing initiator, obtained outcomes appeared to conform more to

the expected relationships for rewarding initiators. For propensity,

where significant effects were obtained, the results were in the

opposite direction; low propensity subjects were more resistant to

the source and message than were high propensity subjects. The

relationship between distance and distraction was not supported.

Although subjects reported being distracted by the personal space

violations, the distraction scores were not significantly different

across distances. The relationships between distance, reward and

propensity were not supported.

The results were discussed in terms of violations of personal

space expectations, distraction and distance, the propensity-to-

counterargue measure and the operationalization of reward. A

number of research improvements and extensions, as suggested by

the findings of this investigation, were discussed.



In a recent review and commentary on persuasion research,

Miller and Burgoon (in press) note that an expanded view of

persuasion must "allow for reciprocal influence, it must also

recognize that persuasive transactions are an essential ingredient

of all social situations" (p. 16). This expanded view also suggests

that other outcomes than attitude or behavior change might be of

importance to persuasive research. They suggest, for example, that

attraction and credibility are but two variables which are rele-

vant outcomes of the persuasion process (pp. 16-17). Along this

same line, perhaps an expanded view of persuasion should include

strategies whereby a source can enhance his/her attractiveness and

credibility in a persuasive interpersonal situation.

Within any interpersonal interaction, a myriad of factors

serve to influence the potential communicative outcomes of the

interaction. Among these factors, interpersonal distancing has

been demonstrated to affect perceptions of attraction and credi-

bility between interactants and modify expected interaction

outcomes. Of the outcomes investigated in interpersonal distan-

cing studies, very few have focused upon the perceived credibility

and attraction of a "violator" who does or does not invade another

individual's personal space. Much of the research we do have has

not systematically set out to establish normative distances

before violating the expected norm and has instead relied on

gross cultural distances as normative behavior. With the excep-

tion of Burgoon (1978) and Burgoon, Stacks and Woodall (1977),

researchers have been content to approximate interaction distances

and ignore the factors which serve to establish actual norms

within an on-going interaction. Additionally, what research is

available has dealt with only one or two variables of interest to

communication researchers. Specifically, more research is needed

to integrate such outcomes as attraction, credibility and persua-

sion in interpersonal interactions. This investigation seeks to

demonstrate the effect on communication outcomes of violating

interpersonal distancing norms between two interacting individuals.

The outcomes of interest are a persuader's influence on another

individual and the reciprocal perceptions the targeted individual

has of the persuader.

Review Of Literature
The growing amount of research on spatial behavior over the

last few years is indicative of the interest in how man structures

his day-to-day interactions. The interest in what Hall (1963)

observed to be man's unconscious structuring of microspace. or
"proxemics," has led researchers to gain a greater understanding

of how we can consciously manipulate the space immediately around

us. Although the use of space may still be outside most individuals'

conscious control, research has and will lead to greater deliber-

ation about its effects on communicative interactions.

Definitions. The distance between two interactants has

been termed "personal space," which may be viewed as the distance

which limits and separates individuals during interaction.

Personal space has been equated to a body-buffer zone (Horowitz, Duff

and Stratton, 1964,p. 651), "the space immediately surrounding an

individual which he feels to be personal, to belong to himself"

(Dosey and Meisels, 1969, p. 93) or a bubble or sphere that acts

to protect an individual from intrusion by others. Hall, using

a similar term, personal distance, refers to a distance of

between 18 inches and four feet which serves to comfortably

separate coacting individuals (Hall, 1969, p. 119). As such, it

is but one of four distance zones defined in terms of comfort or

intimacy, the other three being intimate distance (less than 18

inches), social distance (four to 12 feet) and public distance

(greater than 12 feet). Although recent research has called into

question the accuracy of Hall's spatial zones (Burgoon and Jones,

1976; Mazur, 1977; Shuter, 1977), the relative distances still

appear sound (Heston and Garner, 1972; Connolly, 1975).

In reviews of personal space literature (e.g., Evans and

Howard, 1973; Pederson and Shears, 1973) one finds personal space

treated as both a dependent variable and independent variable.

First, as a dependent variable, personal space has been viewed as

a function of three sets of variables (Burgoon and Jones, 1976;

Watson, 1972): interactant, interactional and environmental;

and, secondly, as a variable which affects interaction outcomes.

Interactant variables are those variables which the individual

brings to the interaction. Interaction variables are those which

concern the context within which the interaction takes place.

Environmental variables reflect the setting within which the

interaction takes place. Second, as an independent variable

the effects of violating distancing norms have been examined,

both for the effects on the individual violating the norm and the

individual whose normative distance has been violated. The

effects studied so far have focused largely on generally compen-

satory and physiological reactions by the recipient of the

violation and his/her changes in perceptions of attraction, credi-

bility and persuasiveness of the violator.

Factors Influencing Distancing Norms

The effects of interpersonal distancing expectations are depen-

dent on three sets of variables. Interactant variables are those

variables the individual brings to the interaction. Interactional

variables are reflected in the type of interaction (e.g., formal,

casual, etc.) the individual finds himself/herself. Environmental

variables reflect the surroundings in ihich the interaction takes

place. Their importance lies in that distancing norms adopted by

any two individuals are a function of all three sets of variables

and, as such, are reviewed in detail.

Interactant Variables

Experiments investigating the effects of interactant variables

on distancing norms generally cite cultural, demographic and

personality variables as factors the individual brings with him/

her to any interaction (e.g., Evans and Howard, 1973; Pederson

and Shears, 1973; Patterson, 1978). The main thrust of cultural

research on distancing norms has been advanced on the assumption

that two types of cultures exist; individuals from "contact"

cultures are assumed to expect and maintain closer interpersonal

distances in an interaction, while those from noncontactt" cultures

are assumed to expect and maintain farther interaction distances

(Hall, 1969; Watson and Graves, 1966; Watson, 1970). Demographic

variables typically include age, sex, status, coorientation and race.

The inclusion of race as a demographic variable, however, is

questionable as it may be more a cultural or subcultural factor

in influencing distance norms (Patterson, 1973) and will be

treated as such here. Personality variables are idiosyncratic

patterns of behavior which serve to modify distancing norms

between individuals; they include such traits as extroversion-

introversion, anxiety and aberrent personality.

Culture. The anecodotal and observational work of Hall

(1959, 1966) first brought the effects of cultural differences in

spacing norms to the attention of researchers. Although several

researchers have concluded that such differences are not opera-

tionable (Forston and Larson, 1968; Jones, 1971; Mazur, 1977) or

possibly a feature of faulty statistical analysis (Jones, 1971),

others have found significant differences between cultures in

distancing norms (Watson & Graves, 1966; Baxter, 1970; Shuter, 1976).

Baxter observed the distances maintained by individuals from three

cultures and found that Mexican/Americans maintained the closest

interpersonal distances, Blacks the greatest and Anglos an inter-

mediate distance. The researcher assumed that Mexican/Americans

represented a contact culture while the Blacks and Anglos a non-

contact culture. Other support for the differentiation comes

from research by Lorenz (1976), utilizing a doll placement task

in which South Americans placed the stimulus dolls a greater

distance apart than did Russians, who, in turn placed greater

distance between dolls that did Iraquis. Similarily, Watson and

Graves (1966) and Little (1968) found less distances to be the

norm between contact than noncontact cultures. Shuter's (1976)

observations of normative distances within contact cultures indi-

cates a degree of variability even within supposed contact cultures:

Panamanians maintained greater distances than did Colombians

than did Cost Ricans; Cook (1970) found the same to be true of

noncontact cultures (British subjects maintained greater distances

than North Americans).

Within cultures, subcultural differences may also affect

distancing norms. Research indicates that racial differences

within a culture influence normative distancing but the actual

norms are still tentative. Willis (1966), for instance, found

that Blacks maintained a greater distance than did Whites when

interacting with members of their own race. Baur (1973), however,

found the inverse to be true as Blacks adopted closer "comfortable"

distances than did Whites when interacting with members of their

own race and sex. Other research suggests that sex and age may

also affect racial distancing norms. Rosegrant and McCroskey

(1975) found that during an interview, Black males maintained a

greater distance from an interviewer than did White males, Black

females and White females, and Whites maintained a greater distance

than did Blacks from a Black interviewer. Additionally, Jones

and Aiello (1973) found that age affected racial distancing among

children: first-grade Blacks adopted smaller interaction

distances than Whites, but by third-grade the differences were

reduced and by the fifth-grade began to reverse.

Why racial differences exist is not clear, but it may be

that Blacks and Whites view spatial and temporal use differently.

For instance, Connolly (1975) found that Black subjects, who

viewed a series of pictures with two interactants at differing

distances, perceived less space in an interaction to be appropriate

in each situation and indicated later that they may have viewed

the sequence of pictures as a continuing sequence of interaction.

White subjects, on the other hand, viewed each picture as a differ-

ent interaction.

It is also possible that the process of adapting to a

new culture influences these normative distances; hence, one

might expect that subcultural groups would begin to adopt the

distances of the dominant culture. The same analysis may be

true of Forston and Larson's (1968) finding of no difference

between contact and noncontact cultures -- the norms of the

dominant noncontactt) culture may have been adopted during the

experiment. Forston and Larson did report that contact culture

individuals appeared to stand closer together than their noncontact

counterparts both before and after the experiment, maintaining

distances approximating their true cultures.

Demographics. The age, sex and status of interactants are

major factors in determining distancing norms and expectations.

Generally, it has been found that in same sex dyads, females

maintain closer distances than do males (Dosey & Meisels, 1969;

Pellegrini & Emprey, 1970; Aiello & Jones, 1971; Rosegrant &

McCroskey, 1975; White, 1975). Similar results obtain in groups

larger than dyads (Mehrabian & Diamond, 1971; Patterson & Schaeffer,

1975). In opposite sex dyads, the results are not so clear cut.

Although Willis (1966) and Giesen and McClaren (1976) report that

individuals adopt distances closer to females than to males, a

number of researchers report that males tend to approach females

closer than females approach males (Dosey & Meisels, 1969; Mehrabian

& Friar, 1969; Hartnett, Bailey & Gibson, 1970). Patterson (1978)

suggests that this effect is due to the expression of intimacy

being traditionally more acceptable for females than males.

Recently, Tipton, Bailey and Obenchair (1975) reported that non-

traditional ("feminist") females approached and maintained closer

distances to males than did traditional females but that traditional

and nontraditional females did not differ in their distancing

when interacting with females.

Aae as a determinant of norms seems to interact with sex.

subcultural background and status. For example, Lomranz, Shapira,

Choresa and Yitzchak (1975) reported that three-year-old children,

regardless of sex, maintained closer distances than did five-

and seven-year-old children; but when approached by other children,

females were allowed to approach closer than males. Doll place-

ment research suggests that between the ages of six and eleven,

children of both sexes prefer closer distances with opposite

sex children than same sex children (Pederson, 1973). As noted

earlier, Jones and Aiello (1973) found age by race differences

in distancing preferences in first and fifth-grade children.

Research cn older individuals suggests that distance is a function

of both the age and status of the interactants. Generally, an

individual is approached closer by someone his/her own age than

someone older or younger (Willis, 1966; Pederson, 1973) and high

status individuals are allowed greater distances in both live

interaction and doll placement tests (Lott & Sommer, 1967;

Mehrabian, 1968; Mehrabian & Williams, 1969; White, 1975; Edwards,


Personality. Personality variables represent idiosyncratic

behaviors which influence an interactant's normative distance

(Burgoon & Jones, 1976). Porter, Argyle and Salter (1970)

point out the importance of such variables by noting that 90

percent of the variance in their initial study was unaccounted

for, which they attributed to idiosyncratic behavior. Williams

(1963, 1971) found that extroverts establish closer distances '

than do introverts. Some research has substantiated this effect

for females but not for males (Patterson & Holmes, 1966;

Patterson & Sechrest, 1970); other research has failed to

support the relationship (Meisels & Canter, 1970; Porter, Argyle

& Salter, 1970). Another variable affecting an individual's

spacing norm is anxiety. Limited support for the inclusion of

this variable comes from Williams (1963), who found that perceived

interpersonal distances are judged as being closer by highly

anxious females than less anxious ones. Tangentially, Long,

Calhoun and Selby (1977) found that neuroticism (partially

produced by feelings of anxiety) was the strongest predictor of

cross-situational seating in a projective test. Aberrant and/or

violent personality also has been demonstrated to influence

distancing norms. Fisher (1967) found that emotionally disturbed

children place greater distances between stimulus figures than do

normal children; abnormal seating and interaction patterns by

mental patients have been reported (Sommer, 1959; Winnick & Holt,

1961); and approach distances for mental patients and nonmental

patients were found to be closer when approaching an object than

a human (Horowitz, Duff & Stratton, 1964). Research on violent

and nonviolent individuals indicate that violent prisoners need

two to two and a-half times the distance that normal, nonviolent

prisoners need (Roger & Schalekamp, 1976).

Coorientation. The degree of coorientation between any two

individuals, which includes such factors as acquaintanceship,

mutual liking and attraction and desire for approval, also

determines distancing norms (Burgoon & Jones, 1976). The degree

of acquaintance has been demonstrated to affect distancing norms

such that greater distances are maintained between strangers than

acquaintances, who, in turn, maintain greater distances than

between friends (Little, 1965; Willis, 1966; Russo, 1967; Edwards, 1977).

Little (1965) found that men across all situations set a normative

distance between self and acquaintances and self and friends while

women only set a standard distance between self and friends.

Edwards (1972) obtained similar results with a projective (figure

placement) technique for women; although not significant, males

also approached closer to friends than to acquaintances and strangers.

Research on attraction and liking suggests that individuals who

are attracted to each other or wish to demonstrate liking estab-

lish closer distances (King, 1966; Lott & Sommer, 1967; Norum,

Russo & Sommer, 1967; Mehrabian, 1968; Kleck, et al., 1968;

Allgeier & Byrne, 1973; Lerner, 1973; Power & Dabbs, 1976;

Storms & Thomas, 1977). Powell and Dabbs (1976), for example, found

that individuals approached a physically attractive person closer

than an unattractive person and that females approached attractive

members of either sex closer than unattractive ones; males approached

attractive females closer than unattractive females but approached

attractive and unattractive males about equally. Similarily, appro-

val seeking females have been found to adopt closer distances than

approval-avoiding females (Rosenfeld, 1965; Mehrabian & Ksionzky,


Interactional Variables

This set of variables concerns the interaction in which the

individuals find themselves. As noted earlier, Hall's distance

zones are based on perceived intimacy or formality, with a great

degree of variability within each zone. Scherer and Schiff (1973)

and Connolly (1975) found that perceived intimacy varies with

distance much as Hall described. In each case, subjects who

viewed slides of people interacting at different distances (and

at various arrangements in the former study) perceived closer

interaction distances as more intimate than farther distances and

corner seating as more intimate than side-by-side seating.

The purpose of the interaction also mediates the normative

distances adopted by individuals. When the purpose is casual

or conversational, seating preferences are generally corner-to-

corner or adjacent. Side-by-side seating is usually found in

cooperative task situations, while opposite seating arrangement

is normally found in competitive task situations. In coaction

situations, farther distances are normally found (Sommer, 1965;

Norum, Russo & Sommer, 1967; Cook, 1970; Long, Calhoun & Selby, 1977).

Environmental Variables

The final variables which determine spatial norms concern the

environmentsin which the interaction takes place. Within the

environment, type of territory, location, familiarity with the

location, number of people within the location and arrangement of

furniture all function to determine distancing norms. The

location of an interaction is contingent upon the type of territory

the individual interactant is in since different distancing

patterns are expected in public through private interactions.

Lyman and Scott (1967) have delineated four types of territory

in terms of accessibility: public, which is open to all; inter-

actional, which is used for social purposes and has social bounda-

ries to limit access; home, where the individual controls access

and is allowed a sense of intimacy and freedom within its bounda-

ries; and body, which is the most highly restricted in terms of

accessibility. Although research has not compared distancing

norms across all four types of territories, the type of territory

does determine to a degree the normative distances interactants

will adopt and maintain.

In terms of location Edney and Jordan-Edney (1974) found, for

example, that while Personal distances decreased on a beach (public

territory) in interacting groups the group's physical size did not

increase as more people joined the group. Similarily, Fried

and DeFazio (1974) reported smaller normative distances in public

transportation than in automobiles, public transportation being

viewed as interactional territory and the auto as home territory.

Generally, it seems to be that the greater the physical area, the

more closely the interactants will space themselves (Sommer, 1962).

In addition to location, familiarity with the environment

also influences the normative distances adopted in interaction.

Edney (1972) found that, with increased familiarity with an

environment, distances between individuals interacting within it

decreased. Similarily, Vine (1975),Aschoff (1965) and Coleman

(1968) also found that familiarity with an environment produced

similar distancing norms.

A third variable affecting distancing norms is density.

Stokols (1973; 1976) notes that density is more closely related

to interaction distance than crowding (a subjective, perceived

environmental restriction cued by contextual cues such as hostil-

ity). Generally, the greater the density within an environment,

the closer will be the normative distances and a more even distri-

bution of space among interactants will occur (Sommer, 1969).

Cne final variable exerts its influence on distancing norms.

Arrangement of the environment plays an important role in estab-

lishing distancing norms in that furniture and seating arrangements

typically determine interactional distances. Sommer (1962) notes,

for example, that the "comfortable conversational distance" is an

arc of about eight feet in a home. Further, individuals generally

prefer across to side-by-side seating arrangements up to a point

that a comfortable (as determined by interactant and interactional

variables) distance is exceeded and then side-by-side seating will

be preferred. As noted earlier, when seated around a table, corner-

to-corner or adjacent seating is preferred for casual interactions.

Effects of Violating Interpersonal Distancing Expectations

Theoretical Models

Several theoretical models have been proposed to explain the

effects of distance on communication outcomes. The earliest, by

Argyle and Dean (1965) proposed that an equilibrium point is

reached whereby increases or decreases in one or more nonverbal

behaviors produce compensatory behaviors leading to a state of

equilibrium. The compensatory behaviors act as approach and

avoidance factors which balance to reach a steady state of behav-

ior. Two more recent models predict distancing effects to be a

curvilinear function of the amount of expected distance between

two individuals and the characteristics of one or both of the

interactants (Burgoon & Jones, 19T6; Sundstrom & Altman, 1976).

Of the two, the Sundstrom and Altman model is more restrictive as

it only considers the effect of degree of acquaintanceship between

individuals (i.e., friends, strangers and strangers who do not

expect to interact) on comfort levels in an interaction. The

Burgoon and Jones model, on the other hand, considers a wide range

of interactant characteristics, which are used to classify initiatorss"

as rewarding or punishing (the intitialor being the interactant

whose deviations are examined for their effect on a reactant).

As such, degree of acquaintanceship is but one variable which can

reflect on the reward/punishment power of the initiator. Other

potential factors might include physical attractiveness, sex

composition, racial composition, status, power and positive and

negative feedback. The Burgoon and Jones model also considers

a wide range of dependent variables, including attraction, credi-

bility, power, liking, comprehension and persuasion. Partial

support for the model has been offered by Burgoon (1978) and

Burgoon, Stacks and Woodall (1977), with physical attractiveness

and positive or negative feedback affecting reactants' perceptions

of an initiator's attraction and credibility.

The Burgoon and Jones model also predicts that the relation-

ship between reward, distance and outcome is curvilinear in nature.

The shape of the curve is based on the amount and direction of

deviation from the expected distancing norm, the reward/punishment

power of the initiator of the deviation and the "threat threshold"

of the reactant (Burgoon R Jones, 1976). As such, the predicted

curves should differ in two ways: first, the rewarding initiator

should achieve more favorable communication outcomes than the

punishing initiator; second, the apex of the curve for the rewarding

initiator should occur at a distance closer than the normative

distance, while the apex for the punishing initiator should occur

at the normative distance. Burgoon, Stacks and Woodall (1977)

argue that, for the rewarding initiator, any deviation, except

within the threat threshold, will produce better outcomes.

Research Paradigms

At least three different research paradigms are identifiable

in the research investigating the effects of normative distancing

violations. The majority of research in personal spacing has

centered on two paradigms: the use of preestablished distances,

or structured interaction distances, and invasions of personal

space. The structured interaction design does not allow for

normative distancing to occur naturally, but instead artifactually

imposes a norm for that interaction. Most of the research in this

paradigm has relied upon Hall's personal and/or social distance

zones (18 inches to 12 feet) for their distance manipulation.

As such, it is questionable whether these distances truly represent

interactional distances as they fail to consider the effects of

the interactant, interaction and environment which contribute to

true normative distances. Research utilizing structured interac-

tion distances, however, has been supportive of an effect whereby

more "immediate" distances (closer distances between individuals)

tend to increase perceptions of an individual's perceived credi-

bility and attractiveness (depending upon the reward or punishment

power of the source) and, in some cases, a persuader's perceived

influence (e.g., Mehrabian & Williams, 1969; Albert & Dabbs, 1970;

Garner, 1972; Rosegrant & McCroskey, 1975; Imada & Hake], 1977;

Riess, 1977). Research also has demonstrated an effect on vio-

lations farther than the norm (Lassen, 1973). Lassen found that

anxiety was increased at great distances (nine feet or more) as

well as close distances. Patterson and Sechrest (1967), however,

found that compensatory reactions decreased at far distances (10

feet), but that ratings of aggressiveness, dominance, extraversion

and friendliness decreased as distances increased (Patterson &

Sechrest, 1970). Invasion research, on the other hand, has

relied on the effect of extremely close distance violations on

the individual and has been demonstrated to produce compensatory

reactions (e.g., reactions of anxiety, nerviousness and flight)

and strong physiological reactance from the individual whose space

has been violated (e.g., Garfinkel, 1964; McBride, King & James,

1965; Felipe & Sommer, 1966; Middlemist, Knowles & Mather, 1976).

The third paradigm is identifiable by the use of actual

interaction distances adopted by the interactants. In this

respect the research representative of this approach is interested

in allowing interactant, interactional and environmental variables

to function to establish a more "natural" interaction distance to

be adopted. Burgoon, Stacks and Woodall's (1977) research falls

into this paradigm where subjects are allowed to establish their

own normative distances from a confederate who then either maintains

that norm or deviates closer or farther than the norm. At least

two other studies fall between the structured interaction and

subject setting norm approaches. Burgoon (1978) identified the

prevailing normative distances for her subject population by

pilot test and then had confederates systematically deviate from

the preestablished norms. In this way she operated her distance

manipulation with more precision than studies based on cross-

cultural norms. Albert and Dabbs (1970) also attempted to at

least identify distances judged as "uncomfortably close,"

"average," and "uncomfortably far." Unfortunately, their average

and far distances (five to six feet and fifteen feet, respectively)

fall outside the personal space zone and include both the social

and public zones as defined by Hall (1966).

Effects of Violating Normative Distancing Expectations

Compensatory and Physiological Reactions. As noted earlier,

most compensatory and physiological reactions to spatial invasion

are the result of extremely close distances between two individ-

uals. These behaviors have been attributed to be the result of

anxiety reactions or discomfort arising from the invasion (Argyle

& Dean, 1965, p. 293). The behaviors manifested by violations of

personal space include changes in eye contact, body lean, increasing

distances between violator and subject and actual flight. Probably

the most often cited study of personal space invasion is that of

Felipe and Sommer (1966) who found such a series of behaviors to

occur when an experimenter violated subjects' personal space. More

recently, Krail and Leventhal (1976) found an effect for sex on

compensatory reactions. Using Felipe and Sommer's method they found

shorter latency periods in compensatory behavior the closer the

invader was to the subject. For same sex interactions, males had

shorter latency periods than did females and a significant sex of

subject by sex of invader interaction was found where a female

invader/male subject interaction elicited longer latency periods

than did a male invader/female subject interaction. Other research-

ers have found that extremely close spatial invasions produce

anxiety on several indices (Horowitz, Duff & Stratton, 1964;

McBride, King & James, 1965; Sommer, 1969; Baxter & Deanovich, 1970;

Patterson & Sechrest, 1970; Patterson, Mullens & Romano, 1971;

Middlemist, Knowles & Mather, 1976); embarrassment (Garfinkel, 1964);

changes in eye contact (Argyle & Dean, 1965; Mehrabian, 1968;

Goldberg, Keisler & Collins, 1974); and changes in body orienta-

tion and self-manipulations (Argyle & Dean, 1966; Felipe & Sommer,

1966; Patterson, Mullens & Romano, 1971). These effects offer

strong support for the existence of a threat threshold as posited

by Burgoon and Jones (1976) and suggest that extremely close

distances between individuals produce responses which the individual

might label as negative.

Attraction and Credibility. The effect of different distances

between two individuals has been demonstrated to produce differing

perceptions of attraction and credibility. Rosegrant and McCroskey

(1975) found that an "invasion" of six inches produced better

ratings of credibility on the dimensions of sociability, competence

and composure. Their interviewers were graduate students and, if

one assumes their status was higher, the results are not inconsistent

with the prediction that rewarding (i.e., high status individuals)

interviewers will be perceived as higher in credibility at a close

distance. Similarily, Garner (1972) found that an invasion to

within six inches of the subject increased ratings of a female

interviewer on the credibility dimensions of competence, composure

and dynamism. Given that the interviewer was moderately attractive

in dress and appearance, it may be inferred that subjects viewed

the interaction as rewarding. Moreover, males may have viewed

the invasion as having sexual overtones, thus increasing the

reward value of the interaction. The same type of results have

been found by Patterson and Sechrest (1970) in an interviewer-

interviewee situation, Heston (1974) with spatial invasion and

Imada and Hakel (1971) in mock employment interviews. Burgoon (1978)

also utilizing an interviewer-interviewee situation, found that

when rewarding and punishing interviewers (giving positive and

negative feedback, respectively) were perceived differently when

they sat at differing distances from the reactant. Rewarding inter-

viewers were perceived as more attractive and credible at closer

distances. (The apex of the curves peaked at a distance about

18 inches closer than the preestablished normative distance.) The

punishing interviewers, however, obtained the highest ratings at

the preestablished normative distance. These studies, however,

utilized structured interaction distances and, thus, did not allow

natural spacing norms to occur. Burgoon, Stacks and Woodall (1977),

however, did allow subjects to establish their own normative dis-

tances (a range of one to forty-nine inches) in an interview setting.

Testing a number of reward/punishment operationalizations, they

found physical attractiveness of the interviewer and positive regard

to be the best reward definitions. Again, the reward interviewer

obtained higher ratings on attraction and credibility by deviating

from the established normative distance while the punishing inter-

viewer obtained highest ratings by maintaining the observed norm.

Rewarding interviewers obtained higher ratings in either the close

or far distance conditions (deviations of 18 inches closer or farther

than the normative distance, respectively) on some outcome variables.

Whether the threat threshold (operationalized as a distance of less

than six inches between interviewer and interviewee) was reached is

questionable. Although there is strong implicit support for such

an effect in the literature, it may be that in these experiments
the distance was not sufficiently close to generate a feeling of

threat or that the laboratory atmosphere and explanation of the

task reduced subjects' reactions to extreme physical closeness.

Persuasion. A review of the literature on nonverbal communi-

cation and persuasion reveals that very little research has been

undertaken in the area of interpersonal distancing and none with

violations of distancing expectations. For the most part, the

research dealing with spatial effects has found it best to explain

hypothesized results on a post hoc basis. Mehrabian and Williams

(1969), for example, used the concept of immediacy and expected

that more immediate (i.e., closer) distance cues should be directly

related to the intended persuasiveness of encoders and perceived

persuasiveness of decoders. Albert and Dabbs (1970) explained

their findings of distance effects in terms of reactance theory,

while Garner (1972) suggested that derogation of topic or source

took place when subjects' space was encroached. Riess (1977),

integrating research on communicator attractiveness, expertise,

status and distance, hypothesized that distance between a speaker

and target would affect perceived persuasiveness when the target

was focused on source attractiveness and status. He concluded

that distancing cues had no effect on status and attraction but

when the focus was on status a trend was obtained for persuasion.

Since all four studies used structured interaction distances,

natural distancing norms were not operational; however, the results

suggest that distancing between individuals does affect persua-


Mehrabian and Williams (1969) found that subjects, when asked

to encode messages of varying degrees of persuasiveness (highly

persuasive, moderately persuasive or purely informational), adopted

the same approximate distances. In a second study, they employed

a decoding methodology, where subjects viewed a video-taped, but

silent, communication at two distances (immediate, four feet, and

nonimmediate, 12 feet). A significant main effect was found where

the immediate distance between persuader and persuade was perceived

as more persuasive than the nonimmediate distance. Further, signif-

icant sex, eye contact and distance interactions were found such

that at the far distance males were less persuasive with 90 percent

eye contact than 50 percent eye contact; females were less persuasive

with 50 percent eye contact than with 90 percent; and at the close

distance, eye contact was nonsignificant for either sex. The effect

of eye contact, however, may be questioned since the persuaders

wore masks with slits for eyes which would make the eyes the

most salient physical feature available to the subject. Another

factor troubling the Mehrabian and Williams study was the opera-

tionalization of distance. Although the immediate distance seems

to fall within Hall's personal space zone, the nonimmediate distance

lies at the outer boundary of social space and far exceeds Sommer's

arc of comfortable conversation. It is plausable that the far

distance was perceived by the subjects viewing the interaction as

impersonal to the persuadees (c.f., Jones& Davis, 1965) who thus

cued in on the mask for curiosity's sake. This might explain the

higher order interactions. The sex difference at the far distances

may be attributed to normative eye contact patterns which indicates

that females maintain more eye contact while speaking than do males

(Exline, 1963).

Albert and Dabbs (1970) found that as distance between a hostile

persuader and persuade decreased so did attitude change. They

used pretested distances of comfort (uncomfortably close, one to

two feet; average, five to six feet; uncomfortably far, fourteen

to fifteen feet) between subject and confederate and, while increased

distance tended to increase persuasion, attention at both the close

and far distances was found to be focused on the speaker's physical

appearance. Two interpretations are possible from this study.

First, if status is an important characteristic in a persuasive

interaction, then the increasing distances adopted by the speaker

may have indicated increased status (e.g., Willis, 1966) and subjects

who perceived the speaker to be of higher status changed their

attitudes in the speaker's direction. Secondly, since the attitude

change for the hostile speaker was nonsignificant between the
"average" and far distances, one might reinterpret these distances

on two grounds. The "average" distance may have approximated the

normative expected distance for a punishing (in this case, hostile)

source, and the far distance (beyond the outer boundary for Hall's

social zone) may have conveyed a status demand so strong as to

override the otherwise negative distance effect.

A second study which found decreased attitude change at extremely

close distances (Garner, 1972) might have reached subjects' threat

threshold. Since the extremely close distance was less than six

inches, heightened arousal caused by an attractive interviewer could

have caused subjects to ignore the topic and focus on their physio-

logical reactions to inadequate space. A third study (Riess, 1977)

manipulated the seated distance and focus between persuader and

persuade. Distance wasoperationalized as falling within Hall's

social-consultative zone (3.5 to 7 feet) and, although no signifi-

cant results were found for distance, a trend was obtained whereby

perceived persuasiveness increased as distance increased, but only

in the status orientation. His results are comparable to Albert

and Dabbs' when interpreted from the perspective that higher

status individuals are accorded more distance. Again, however,

the use of structured interaction distances may be responsible for

the lack of significance on distance.

Several potential explanations may be advanced to account for

these results which are not completely inconsistent with each other.

That spatial invasions (i.e., distances of six inches or less)

produce less persuasion should be expected from the nature of the

invasion alone; invasions of this sort have been documented to pro-

duce anxiety, embarrassment and reactions reflecting general discomfort

on the part of the individual whose space has been invaded (e.g.,

Garfinkel, 1964; Horowitz, Duff & Stratton, 1964; Argyle & Dean,

1965, 1966; Patterson, Mullens & Romano, 1971), as well as strongly

analogous behavior in subhuman species (e.g., Calhoun, 1962;

Christian & Davis, 1964). Mehrabian and Williams' findings that

more immediate distances produce more persuasion could be explained

by noting that the immediate distances which was not close enough

to constitute an invasion, may have approximated that an average

individual adopts in delivering a persuasive message, but close

enough to indicate some degree of attraction between the two indi-

viduals. As such, closer distances would indicate both involvement

in the task and attraction within the interaction. Albert and

Dabbs' find that increased distance produced increased acceptance

of the persuasive message and attention to physical appearance

is best explained from the status effect. Additionally, the only

significant difference was found between the close and average

distances. The decreased message acceptance at the close distance

suggests that it may have been too close and possibly threatening

to the subject. Since the close distance was measured as one to

two feet, face-to-face,with the interactants seated facing each

other, the actual knee-to-knee distance would necessarily be less

than one inch; a distance which should be threatening to the

subject. What the results would have been if normative distances

had been allowed to be established and then violated has not been

answered. A potential rationale for the effects of violating

distancing expectations lies in the distractive effect such

violations may have on a persuader's attraction, credibility and


Distraction and Violations of Interpersonal Distancing Norms

Recent research suggests that distraction created by other

people or environmental events may mediate an individual's suscep-

tibility to a persuasive message (e.g.,Baron, Baron & Miller, 1973).

In much of this research the process of counterarguing has been

used as an explanatory concept. It has been argued that the

ability of a message to be influential is dependent upon the

probability of its being received and understood by a receiver,

its probability of being discredited and the validity of the

arguments in the message which are yielded to (Weyer, 1974).

Baron, Baron and Miller (1973) argue that distractive stimuli may

facilitate persuasion by dividing a receiver's attention when

exposed to counterattitudinal messages, by disrupting the screening

of information and inhibiting the production of counterarguments.

It is this disruption of counterargument production which leads to

increased susceptibility of accepting the message.

Festinger and Maccoby (1964) first reported that persuasion was

enhanced by distracting receivers of counterattitudinal messages

through the simultaneous presentation of disruptive stimuli. They

suggested that distraction interfered with counterarguing, disrupting

a subvocal (psychological) process whereby the receiver is prevented

from subvocally counterarguing against the message. Along this line

of thought, Brock (1967) found that where there was an increase in

message discrepancy, there was also an increase in receiver counter-

argument production. Further, Brock's data suggested an inverse

relationship between the production of counterarguments and yielding

to the message; as counterargument production increased or decreased,

yielding to the message decreased or increased.

Since counterarguing is a subvocal, psychological variable, the

question arises as to whether distraction interferes with the produc-

tion of counterarguments. Several studies suggest that it does

(Brietrose, 1966; Dorris, 1967; Osterhouse & Brock, 1970; Zimbardo,

et al., 1970; Keating & Brock, 1974). Decreased counterarguing

was strongly associated with counterargument production where

subjects who were distracted had greater acceptance of the message's

position and less counterarguing, and, conversely, nondistracted

subjects were more resistant to the message and produced more

counterarguments. Other researchers (Freedman & Sears, 1965;

Kiesler & Mathog, 1968), although not measuring counterarguing, have

relied upon the distraction-counterargument production relationship

to explain their findings of increased persuasion with distraction.

Miller and Baron (1973) suggest that the counterarguing process

consists of two dimensions, one being purely cognitive and the

other motivational. The purely cognitive dimension taps activity

consisting of content-oriented information being generated by an

individual to refute a persuasive message. The motivational aspect

reflects an individual's desire to counterargue and resist the

persuader's message. High motivation is indicated when an individ-

ual increases counterargument production and is reflected in the

presentation of a large number of reasons to reject or discard the

persuader's arguments. Brandt, DinKlecker and Stoyanoff (1977)

argue that individuals differ in their propensity to counterargue;

that just as some people tend to focus on the communication source

when processing a persuasive message, others are motivated to focus

on the content of the message, testing, attacking and refuting

arguments contained in the message. They found support for the

hypothesis that individuals with a high propensity to counterargue

would be more prone to counterargument production and more resistant

to the communicator's influence than individuals with a low propen-

sity to counterargue. Further, they suggested that perceptions

of source credibility are also affected by an individual's propen-

sity to counterargue. Festinger and Maccoby (1964) suggested that,

other than discrediting the points a communicator makes in a

persuasive presentation, counterarguing also takes the form of

source derogation. Brandt, et al. (1977) found that subjects with

a high propensity to counterargue perceived the message source as

significantly less credible than did subjects with a low propensity

to counterargue. This result is similar to Cook's (1969) finding

that decreased counterarguing is associated with highly competent


Interactional Distance and Distraction. Distractive stimuli

have traditionally taken the form of environmentally produced

distractors. Distractors employed have ranged from irrevelant

movies and slides, monitoring tasks, and heckling to eating

while reading evaluation instructions. Little research has dealt

with distracting stimuli which are inherent to a social interaction

aside from environmental or outside distraction. In cases where

outside stimuli have been employed, audio- or video-taped presen-

tations of messages were employed (e.g., Miller & Baron, 1968;

Brandt, et al., 1977; Burgoon, et al., 1978) and have dealt with

interactions not specifically of an interpersonal dyadicc) nature.

Of interest to communication researchers might be stimuli which

are available to a persuader in a social interaction which would

increase his/her persuasiveness, credibility and attraction.

Violations of interpersonal distancing expectations provide a

subtle, yet potentially distracting stimulus, which could be manip-

ulated by a persuader.

The relationship between distraction and the violation of dis-

tancing expectations has not been specifically addressed in the

literature, nor has a rationale for potential distractive stimuli

been advanced. It seems reasonable that a rationale can be derived

based on the disruption of an individual's motivation to counter-

argue a persuasive message due to (1) the distance the persuader

maintains or does not maintain from the subject and (2) the charac-

teristics of the persuader. Additionally, a third variable, pro-

pensity to counterargue, may contribute to the distancing-persua-

sion process.

Based on the assumption that individuals, when presented a

counterattitudinal message, screen and counterargue information

and arguments advanced by a persuader, it is expected that the

distance the persuader adopts in the interaction will either facili-

tate or inhibit the subject's motivation to counterargue. It is

suggested that this effect takes place due to either anxiety

from extreme spatial invasion or from attributions concerning

hedonic reward and personalism at other distances.

Almost all theories of interpersonal attraction predict that

a source's attractiveness enhances the efficacy of his or her

persuasive message (Tedeschi, Schlenker & Bonoma, 1973), attractive

(rewarding) sources should gain greater compliance to their messages

than unattractive sources. Support for source attractiveness is

offered by Mills and Aronson (1965) who found that a female who was

made to appear physically attractive was more effective in influ-

encing a male audience (when she made clear her intent to influence)

than was the same female made physically unattractive.

While reward-punishment power may be a characteristic of the

source, the perceived personal involvement indicated by the sit-

uation should be affected by the distances adopted and deviated

from in the interaction. Proxemic research suggests that individ-

uals who are attracted to one another establish closer normative

distances, which should indicate a degree of personalism within

that interaction. If one individual deviates from the norm, the

degree of personalism should be proportional to the hedonic rele-

vance of that individual and the direction of the deviation. If

the initiator of the deviation is rewarding, perceived personalism

should increase with a deviation closer than the norm. If the

initiator is punishing (low in hedonic relevance), the degree of

personalism should decrease proportionally with the deviation since

punishing initiators are not allowed deviation from established


The Burgoon and Jones (1976) model proposes that violations

of personal spacing expectations should lead to positive and negative

effects depending upon the characteristics of the initiator of the

deviation from the distance expectation and the direction and amount

of that deviation. As such, four distances can be conveniently,

if arbitrarily, identified: the normative distance established at

the onset of the interaction, a close distance which is a deviation

closer to the individual than the norm, a far distance which is a

deviation farther than the norm and a threat threshold which is an

extreme spatial invasion. Burgoon, Stacks and Woodall (1977)

suggest that the four distances may be operationalized as points

on a continuous line set not by preestablished distances but by

the actual normative distances selected by the subject and confed-

erate. As such, instead of referring to four distances as structured

interactional distances, the distances take on a more natural mean-

ing, except for the threat threshold which is defined as a few inches

from the subject.

The Burgoon and Jones model proposes that extreme spatial

invasion constitutes a violation of an individual's threat threshold.

Effects of such close distancing have been noted to produce both

behavioral and psychological reactions which have been labeled as

anxiety or embarrassment. These reactions should cause individuals

to divert energy from concentrating upon message content to iden-

tifying the source of the reactance. This focus might lead to two

complementary outcomes. First, the individual could identify the

anxiety or embarrassment with the initiator of the invasion. The

equating of anxiety with the initiator should cause the individual

to derogate the source in terms of perceived credibility and attrac-

tion which, in turn, should lead to less acceptance of the initia-

tor's message. Secondly, the extremeness of the invasion might

cause the individual to ignore the message completely; that is, in

order to overcome or understand the reactance, the individual should

be so distracted as to miss the message and its arguments.

These effects could explain Garner's (1972) finding that extreme

invasion produced less favorable attitudes toward the topic and Albert

and Dabbs' (1970) finding of less attitude change with decreased

distance. That subjects engage in source derogation and counter-

argument production would seem to intuitively make sense at

extremely close distances. If increased counter-arguing is associa-

ted with increased source derogation then spatial invasion should

facilitate both, as an apparent cause for discomfort is readily

available to the subject. Garner, however, reported that perceptions

of competence, composure and dynamism were higher with invasion than

noninvasion subjects. As noted earlier, the attractiveness of the

confederate in the case of male subjects and the perception of

female-female distancing norms in the case of female subjects may

have facilitated ratings of credibility. However, given the

moderately attractive source, the distance distraction may have

interfered with subjects' motivation to derogate the source yet may

not have been sufficient to interrupt the subjects' motivation to

counterargue the message. It may be that the invasion produced

more attention pertinent to source characteristics such that the

invasion subjects interpreted the distance as an indication of the

source's attraction to them, yet not sufficiently enough to override

counterarguing the message. Another explanation may be that stand-

ing distances are interpreted differently than seated distances.

An additional finding by Albert and Dabbs (1970) might clarify

what happens at "close" distances. They found that attention was

focused on the source's physical appearance at the "close" condition

and not on the message. Although their "close" distance does not

approximate an extreme invasion distance as they operationalize it,

reinterpreting the face-to-face distance (12 to 24 inches seated

distance) suggests that knee-to-knee distance of less than one inch

might be defined as an invasion of personal space. Albert and Dabbs

also found that source characteristics affected persuasion. Specifi-

cally, the friendly persuader produced more attitude change across

distances than did the hostile persuader. If friendly and hostile

characteristics can be equated with rewarding and punishing charac-

teristics, these results suggest that rewarding individuals produce

better outcomes than do punishing individuals, which is consistent

with the Burgoon and Jones model. In terms of credibility and attrac-

tiveness of the initiator of an invasion, Burgoon (1978) and Burgoon,

Stacks and Woodall's (1977) findings suggest that deviation produce

lower ratings of initiators' credibility and attraction at the

threat threshold than the close or normative distances.

In accord with Burgoon and Burgoon, Stacks and Woodall's

suggestion that the initial model of violations of personal space

expectations be modified in the case of rewarding initiators, any

deviation, up to the point of the threat threshold, should produce

better communication outcomes for that source. A deviation closer

than the norm should indicate attraction and increased liking toward

the reactant. Since people are more susceptible to persuasive

messages from sources they perceive as liking them (Mills, 1966;

Rosnow & Robinson, 1967), their motivation to counterargue a counter-

attitudinal message should be disrupted and greater yielding to

the message should result. Additionally, a deviation closer by

the rewarding source would allow the reactant to engage in quasi-

courtship behaviors, which should also decrease motivation to

counterargue. With the quasi-courtship behaviors and attraction

there should be an increase in perceived personal involvement in the

interaction. This increased involvement reflects upon the reactant's

feeling that the influence is mutual, directly motivating the behavior

of the initiator.

For the punishing source, however, the effect of distancing

violations is to facilitate counterarguing. This facilitation

should be coupled with an increased motivation to derogate the

source. Although research indicates that, when focused in on

negative characteristics of a source, distraction inhibits the

counterarguing process (Burgoon, et al., 1978), the distance

distraction should be more subtle. Burgoon, et al., has subjects

assigned to "critical sets" whereby the subject was instructed to

attend to positive or negative source characteristics while the

distance manipulation should produce first negative attributions

of the source and then a transfer of these attributions to the

message. As such, the increased motivation to derogate the source

should lead to an increased motivation to counterargue the message.

A deviation farther than the normative distance established

between the persuader and reactant should decrease outcomes for the

punishing source and increase outcomes for the rewarding source.

Aside from indicating less attraction, research suggests that a

farther distance is indicative of antipathy (Sommer, 1971), incompat-

ibility (Altman, 1971), aggression and dominance (Patterson & Sechrest,

1967) and anxiety (Lassen, 1973). As such, the reactant should be

motivated to produce counterarguments and engage in source

derogation. Since Albert and Dabbs found attention at their

far distance also to be centered on the physical appearance of the

source, it is reasonable that the farther distance heightens the

reactants'perception of the source's attractiveness and provides

more motivation to both generate counterarguments and to derogate

the source, if that source is punishing. Since the rewarding source

is allowed greater deviation from normative behavior the deviation

farther than the norm would not appear counternormative but may

actually increase his/her attractiveness and status (e.g.,

Burgoon, Stacks & Woodall, 1977). This would result in less

source derogation and counterarguing.

At the normative distance, the rewarding source should pro-

duce lower outcomes than at the far or close distances. This

effect is reflected in that by maintaining the normative distance

the source has done nothing to distract the reactant from counter-

arguing the message and derogating the source. For the punishing

source, the normative distance should produce the best results

as deviations from the norm should facilitate reactants' motivation

to produce counterarguments and engage in source derogation.

An additional variable, propensity to counterargue, suggests

that individuals differ in their ability or proneness to engage in

counterarguing. Research suggests that individuals with high

propensities to counterargue are not affected by the distractive

stimuli (Brandt, et al., 1977). High propensity individuals would

seem to overcome the distraction and continue to focus on the

content of the message while still derogating the message source.

This suggests that, whatever the distraction, high propensity

individuals should be more resistant to a counterattitudinal message,

at least for environmental or outside distractive stimuli. The

question which has not been addressed is whether the kind and/or

variation of similar kinds of stimuli produce the same result. If,

as Festinger and Maccoby (1964) and Brandt, et al., (1977) suggest,

counterarguing also takes the form of source derogation, it is rea-

sonable to conclude that if a persuader can manipulate the distrac-

tion so as to enhance his or her perceived credibility and attrac-

tion, it should inhibit the receiver's motivation to engage in counter-

arguing. In other words, a distraction which increases attraction

and credibility should result in less counterarguing. This should

hold true for both high and low prepensity-to-counterargue indi-

viduals. The amount of counterarguing, however, should still be

higher for the individual with a high propensity to counterargue.

It should be expected, then, that violations of interpersonal

distancing norms, in terms of propensity to counterargue, would

produce results similar to thosediscussed earlier. High propensity

subjects, when interacting with a rewarding source who presents a

persuasive message, should be most susceptable to influence at devia-

tions closer or farther than the expected' normative distance and

less influenced and more critical of the source at extremely

close (threatening) and normative distances. When the persuader

is punishing, the most influence should occur at the established

normative distance. For low propensity subjects, the same result

should occur, regardless of the reward/punishment power of the

persuader, except that in terms of perceived attraction and

credibility and susceptibility to the persuader's influence, low

propensity subjects should yield more to the persuader's position

and perceive him/her more credible than their high propensity coun-


Derivation Of Hypotheses

Based on the preceding rationale, the first four hypotheses

parallel those of the modified Burgoon and Jones model. Rewarding

(i.e., physically attractive) sources should receive higher

ratings of attraction and credibility and be more persuasive than

punishing (i.e., physically unattractive) sources. It is also

predicted that the rewarding and punishing characteristics of the

source will interact with the distance and that two separate curves

will result from that interaction.

H1: Physically attractive sources, who present subjects
with a counterattitudinal message, will receive higher
ratings of credibility and attraction and will be more
persuasive than physically unattractive sources.

H2: The effects of distance on attraction, credibility and
persuasion are dependent upon the nature of the rewarding
or punishing characteristics of the source, i.e., the
effects are interactive rather than additive.

H3: For physically attractive sources, the relationship
between distance and attraction, credibility and persua-
sion is curvilinear in nature such that a close or far
distance will produce higher ratings of credibility and
attraction and more acceptance to the persuasive message
than a threat or norm distance.

H4: For physically unattractive sources, the relationship
between distance and attraction, credibility and persua-
sion is curvilinear in nature such that a normative
distance will produce higher ratings of credibility and
attraction and more acceptance to the persuasive message
than a threat, close or far distance.

From the research that suggests that individuals differ in

their propensity to counterargue, subjects with a low propensity

to counterargue should be less resistant to the persuasive message

than subjects with a high propensity to counterargue. Additionally,

low propensity subjects should rate the message as more credible

than their high propensity counterparts.

H5: Subjects with a low propensity to counterargue will be
less resistant to a source's counterattitudinal message
and will perceive the message source as more credible
than will high propensity subjects.

It is expected that subjects who interact with physically

attractive message sources should yield more to the source's position

than to physically unattractive sources. Further, both subjects

with a high propensity to counterargue or low propensity to counter-

argue should be more resistant to a message from an unattractive

source than an attractive source, with high propensity subjects

more resistant to the message and critical of the source than low

propensity subjects.

H6: Subjects will be less resistant to counterattitudinal
messages from an attractive source than an unattractive
source, with high propensity-to-counterargue subjects
more critical of the source and resistant to the message
than will be low propensity subjects.

High propensity-to-counterargue subjects who are not distracted

should be most resistant to the message, while distracted low

propensity subjects interacting with an attractive source should

be least resistant to the message.

H7: Nondistracted, high propensity-to-counterargue subjects
will be most resistant to a counterattitudinal message,
while distracted low propensity subjects will be the
least resistant to the message from an attractive source.

The effect of distance is dependent upon the reward/punishment

power of the source. Whereas distance and reward interact to produce

two different curves, propensity is additive and should produce two

parallel curves for both rewarding and punishing initiators. From

the propensity main effect and the hypothesized curvilinear rela-

tionship between reward and distance, it is expected that high

propensity-to-counterargue subjects will be more resistant to the

source's message and rate the source as less attractive and credible

than low propensity-to-counterargue subjects. For the physically

attractive source, the apex of both curves should occur at the

close and far distances, for the physically unattractive source,

the apex should be at the established normative distance.


H,: Given high propensity subjects, the relationship between
reward and distance and attraction, credibility and
persuasion is curvilinear in nature such that a close or
far distance for the physically attractive source and a
normative distance for the physically unattractive source
will produce the highest ratings of attraction and credi-
bility and the most acceptance of the message.

Hg: Given low propensity subjects, a curve parallel to that
of high propensity subjects will be produced but yield
higher ratings of attraction and credibility and more
acceptance of the message across all distance conditions.



The experiment was carried out during the first three weeks

of the Spring Quarter, 1978, at the University of Florida. Subjects

were initially pretested for propensity to counteragrue and assigned

to either high or low propensity conditions. Subjects then interacted

with a confederate who either maintained or deviated from an estab-

lished interpersonal distancing norm before presenting an unexpected

counterattitudinal message. Following the interaction subjects

rated the confederate on measures of credibility, attraction and

their attitude toward the topic of the persuasive message. In

addition they responded to scales and an open-ended questionnaire

designed to measure distraction.

Subjects and Confederates. Subjects were 406 undergraduate

students enrolled in speech courses at the University of Florida.

However, 56 were excluded from the analysis: six because they knew

the purpose of the experiment, five because the confederate either

forgot the message or began over again; thirteen because they did

not remain seated; and 32 for failure to complete the post measures.

Of the 32 excluded from the analyses, ten failed to complete the

attitude measures and 22 for failure to complete either the attrac-

tion or credibility scales. Participation was voluntary in some

cases or for class credit.

The confederates were two male and two female volunteers

enrolled in nonverbal communication courses. Initially, twelve

volunteers were chosen by the experimenter on the basis of their

physical attractiveness. They were then rated for physical

attraction by 35 students enrolled in an upper division speech

course who did not participate in the actual experiment. Subjects

rated each volunteer on a series of subscales developed by McCroskey

and McCain (1974) reflecting perceived physical attraction. From

the 12 volunteers rated, four were finally selected to participate

as confederates in the experiment. Criteria for final selection

was based on the ratings of physical attractiveness. The ratings

then submitted to an Analysis of Variance and comparison of mean

attractiveness scores was used to select the final four confederates.

Confederate ratings across the individuals ranged from 22.26 to 12.49

(p .05). Reward was treated two ways: as a categorical variable

with four levels and as a continuous variable. Because reward was

found to deviate from linearity, a third operationalization was


Independent Measures. There were three independent variables

in this investigation: reward, distance and propensity to counter-

argue. Reward was operationalized as a message source's physical

attractiveness as rated by McCroskey and McCain's (1974) subscales

for physical attractiveness. Confederates' scores were 22.26, 18.60,

15.77 and 12.49; all significantly different from each other (p .05).

Reward was operationalized as a simple dichotomy (Burgoon, 1978;

Burgoon, Stacks & Woodall, 1977). In these analyses, the two

highest rated confederates' outcomes wre collapsed and tested against

the collapsed outcomes for the two lower rated confederates. The

dichotomy produced significantly different levels of reward

(F = 12.56, p .05).

Distance was operationalized as the thigh-to-thigh seated dis-

tance between the confederate and subject. Instead of using structured

interaction distances (e.g., Albert & Dabbs, 1970; Riess, 1977),

subjects were allowed to establish their own normative distances

from the confederates. The distance manipulation was accomplished

in the following manner: once the normative distance was established,

a deviation of 18 inches closer than the norm constituted the close

distance, whereas a deviation of the same distance farther than the

norm constituted the far distance. For the threat distance, the

confederate's deviation was less than three inches from the sub-

ject. The normative distance was operationalized as no deviation

from the established normative distance. Distance was then

operationalized in two ways, first, as a categorical variable with

four levels (threat, close, norm and far) and, second, as points on

a continuous line set by the deviations from the actual norm

adopted by the subject. The treatment of distance as four levels

was used to arbitrarily set the four distance conditions. The

continuous treatment allowed distancing to take on a more

natural form. The normative distances adopted by subjects

ranged from one to forty-seven inches, indicating that subjects'
did not feel constrained to adopt any particular distance. The

advantages of treating distance as a continuous variable were

that it avoided the problems of artificially defining distance

as rewarding or punishing and it provided a more sensitive test

of the hypotheses.

Subjects' propensity to counterargue was measured by scores

obtained from a modified technique reported by Brandt, et al.

(1977) which taps the content dimension of counterarguing.

Additionally, items were added to measure subjects' motivation

to derogate the message source. Thirty propositional statements

were pretested, each relating to a different topic, to determine

which were most discrepant from held beliefs. Sixty-five under-

graduates from the Winter Quarter served as pretest subjects.

Each responded to the statements on three sets of scales:

(1) a seven-point interval scale bounded by "strongly agree" and

"strongly disagree;" (2) a set of four semantic differential scales

suggested by Osgood, Suci and Tannenbaum (1957) bounded by the

adjectives "good-bad," "valuable-worthless," "pleasant-unpleasant,"

and "wise-foolish;" and the Known Interval Scale, developed by

Burgoon (Burgoon, Burgoon & Vaughn, 1976). The first set of scales

represent those used by Brandt, et al., while the second and third

sets were identical to those used in the experiment. Based on

subjects' responses, the following six topics were chosen as most

belief discrepant: (1) "The State University System should redesignate

areas of study to different universities, i.e., UF should only be

allowed to offer hard sciences, FSU only social sciences, USF only

management and marketing, etc.;" (2) "24-hour visitation in

dormitories should be abolished at this university;" (3) "Search

and seizure limits on police should be abolished;" (4) "The purpose

of the university should be primarily research-oriented with teaching

at best a secondary orientation;" (5) "There should be an immediate

tuition increase at this university;" and (6) "The sale of heroin

in the United States should be legalized." The six topics were

chosen because they yielded means of 2.1 or less on the belief

scale (on a one to seven scale), responses were highly skewed in

the direction of opposition and the three sets of scales were

highly correlated. Five of the topics were then selected for in-

clusion in the propensity measure and the remaining topic was used

for the actual message (legalization of heroin). Supportive arguments

were then generated from various sources (e.g., debate files, speeches,

newspapers) for each of the five topics. Additionally, phrases

were added which represented derogatory or defamatory statements of

a source using these arguments. The arguments and phrases were

then randomly ordered in a "check-list" format under the appro-

priate topic, with the source derogation list appearing after the

content list (see Appendix A).

Because Brandt, et al., (1977) made no assumptions concerning

the dimensionality of their instrument and because a second set of

items were added, estimates of internal reliability were computed

for each topic and derogation scale. The estimates were computed by

the Kuder-Richardson Formula 20, as suggested by Guilford (1954),

for dichotomous data. Reliabilities obtained were .59, .72, .54,

.75 and .75 for each of the five topics, respectively. Reliabilities

for the source derogation scales were .59, .75, .80, .72 and .77

respectively. Chronbach's Coefficient Alpha was then computed

to determine subjects' frequency of acceptance-rejection across

topics for each set of items (Chronbach, 1951). Coefficient Alphas

of .69 and .54 were obtained for the content and source derogation

measures, respectively. Reliability estimates were lower than

expected; consequently the interpretation of results may be suspect.

Additionally, correlations between marking behavior for high and low

propensity subjects for content and source-oriented item sets were

computed. Since it was assumed that subjects with a high propensity

to counterargue would also derogate a message source more than

their low propensity to counterargue counterparts, the two sets

would be correlated. The correlation between content and derogation

scales for high propensity subjects was -.12 (p < .10) and .02 (nsd)

for low propensity subjects. The correlations were not in the

expected direction for high propensity nor as strong as desired.

The validity of the instrument was treated in three ways. First,

in the same manner as Brandt, et al., (1977), it was expected that

an inverse relationship between propensity to counterargue and communi-

cation acceptance would occur and, to the extent that subjects'

responses to the measure corresponded to that expectation, some

measure of validity would be obtained. Secondly, if the results

obtained were similar to those of Brandt, et al.'s, then some measure

of validity could be expected. Finally, a separate sample of

subjects (n = 29) was administered the propensity-to-counterargue

measure, exposed to a randomly selected recording of the

experimental message and then, requested to react to the message in

essay form. Statements were then categorized according to whether

they were content- or source-oriented and were counted. The number

of statements were then correlated with the subjects' scores on the

propensity measures. Because the experimental message was heard

before the essay was written it was expected that correlation between

the number of counterarguments and derogatory statements would be

high for the essay measure. The obtained correlation (r = .82,

p < .05) indicate the statements were related. Because the message

provided subjects with a "live" source and arguments stated in a

logical order, any positive correlation with the propensity measure

was taken as a measure of concurrent validity. A correlation of .05

(nsd) was obtained between the two measures for propensity and .09

(nsd) for derogation, indicating the two measures were not correlated.

Message. The topic of the persuasive message argued in favor of

the legalization of heroin in the United States (see Appendix B).

The message was written by graduate students for a previous study

(Burgoon, et al., 1978) and found to be extremely persuasive. The

message contained 491 total words with an average sentence length of

21.20 words. Comprehension was rated as extremely high on the Con-

tingency Index (Becker, Bavelas & Braden, 1961), with an index of

6.65. The Gunning-Fog Readability Index (Gunning, 1952) indicated

the readability of the message was comprehendable to those with

over 12 years of formal education. Since the intensity of the language

used in a message has been demonstrated to effect attitude change

(Burgoon & Miller, 1971), the message was created to be moderately

intense with highly intense metaphors and extreme adverbial quali-

fication not used.

The message was memorized by each confederate. It was rehearsed

until a verbatim delivery of approximately the same rate was reached.

The message took approximately three and one half minutes to deliver.

To ensure that the messages did not differ across distance and

reward conditions, all messages were taped via a two-way speaker

system without the knowledge of the confederate, and one message from

each confederate was randomly selected. These messages were rated by

20 undergraduates on credibility and attitude scales. Mean ratings

were then compared to determine if any systematic variation had

occurred. No significant differences were found. Additionally, six

tapes per confederate were randomly selected and timed. No signi-

ficant differences were found for length of message either between

or within confederates.

Dependent Measures. There were three dependent measures in

this investigation: source credibility, source attraction and atti-

tude. Additionally, two dependent measures were added as manipulation

checks. Subjects rated the confederates on credibility via a series

of semantic differential scales recommended by McCroskey, Jensen

and Valencia (1973) which measure peer credibility. Altogether

five dimensions were tapped: competence, character, composure,

extroversion and sociability. Attraction was measured by a series

of twelve Likert-type statements that reflect physical, social

and task attraction (McCroskey & McCain, 1974). Separate dimension

scores were used as the dependent measures for attraction and

credibility. Attitude was measured two ways. First, by using the

evaluative dimension suggested by Osgood, Suci and Tannenbaum

(1957) which consisted of four semantic differential items and,

second, by the Known Interval Scale (KIS) (Burgoon, Burgoon & Vaughn,

1976). The KIS measure was used to help gain support for the

measure and as the first test of the eleven point scale. The

correlation between the two attitude scales was .86 (p < .05).

The two distraction self-report measures were used to establish

if the distance deviations were distracting. The first check

consisted of a series of adjectives bounding a seven-point semantic

differential-type scale. Scale boundaries were selected to represent

dimensions of distraction which should have occurred in the experimen-

tal setting. As such, the items bounding the scales were: "calm-

anxious," "comfortable-uncomfortable," "distracted-not distracted,"
"relaxed-tense" and "attentive-inattentive." The scales were

submitted to factor analysis with varimax rotation which produced

one factor with all five scales loading greater than .60.

A second manipulation check was administered after the subjects

had completed all experimental conditions and consisted of having

subjects list any factors which made them pay more or less atten-

tion to the persuasive message.

Experimental Procedures. Two weeks prior to the experiment,

students enrolled in sophomore and junior level speech classes

completed the propensity to counterargue measure. The instrument

was introduced as part of on-going research being conducted by the

Center for Human Communication Research. Results indicated a mean

of 65.7 percent rejection, with a standard deviation of 14.45. The

sample median (66%) was used to establish high and low groups.

Following training of the confederates, subjects were asked to

report to a room to participate in a study on coverbal word use.

The room, which was equipped with a two-way speaker system that per-

mitted audio-taping, had the appearance of a clinic waiting room.

It was carpeted and arranged with a row of armless, padded chrome

chairs which, when placed side-by-side, formed a bench-like surface

that occupied most of the wall. Three identically furnished rooms

were counterbalanced within the design.

Prior to entering the room, subjects were paired with a confede-

rate and informed that a short delay had occurred and that they

would be placed in a "waiting room". In each case the confederate

entered the room first and sat at the far end of the bench, approxi-

mately 20 inches from the end. The subject then took a seat and an

assistant handed each a "subject informed consent form" to be

completed. The subject's form indicated the nature of the coverbal

interview which was to take place and asked that he/she sign the

form. The confederate's form accomplished two tasks: (1) it allowed

him/her to record the normative distance adopted by the subject

on the form and (2) it contained the assigned distance condition he/

she was to adopt. Once the forms were completed the assistant

collected them and informed the pair that he/she would be back

shortly. After a few seconds, the confederate stood up, crossed

the room and closed the door, which the assistant had "accidently"

left ajar. Upon returning, he/she adopted the assigned distance

condition. Once seated, the confederate inquired whether or not

the subject was enrolled in a speech course and whether he/she

was taking public speaking. The confederate then stated that

he/she had to deliver a memorized speech in class (indicating

that the topic he/she had chosen was from a survey filled out

earlier in the course) and asked if he/she might practice it on

the subject. All but three subjects agreed to hear the speech

(the three who did not were in threat conditions and immediately

crossed the room, refusing to cooperate).

During the message presentation all other nonverbal behaviors

were kept constant. Eye contact was practiced and maintained at

fifty percent and the same body angle and orientation was maintained

by each confederate at all distance conditions.

Following the message the subject and confederate were taken to

separate rooms. While the subject reported to the next room, the

confederate exited from the area by means of a back stairway and

returned to the original starting position. Once in the room,

the subject was informed that since the study was designed to explore

in the difference between human-human and human-machine coverbal

behavior, the experimenters needed some information about the subject

and the person with whom he/she had been waiting. Subjects were led

to believe that they might be paired with the other person and were


Since part of our study deals with the interface
between humans and machines versus interface between
humans and humans, we need you to fill out some
additional information about the individual with whom
you were waiting. Additionally, we need some
information about you.

A copy of the questionnaire appears in Appendix D. It contains

1) the source credibility scales and 2) the attraction scales,

3) the first distraction manipulation check and 4) the attitude

statement, which is among five other statements. It was explained

to subjects that the attitude section was not part of the current

experiment, but was included for someone else collecting survey

data on campus. It was hoped that the subject would equate this

section with the survey mentioned by the confederate earlier.

Additionally, subjects were informed that they did not have to com-

plete the survey, but that it would help the other researchers if

they did.


Following completion of the dependent measures, the

subject then reported to a second room, was partially debriefed and

asked to complete the second manipulation check. All subjects were

then debriefed and sworn to secrecy. The second manipulation check

and debriefing statement also appear in Appendix D.



Multiple regression analysis was used to test the nine

hypotheses. The dependent variables of attitude, credibility and

attraction were regressed on reward, distance, propensity to

counterargue and the resultant interactions. The experimentally

manipulated distance was treated as both categorical and contin-

uous data. Reward was manipulated in three ways: first, as a

simple dichotomy; second, as four levels, with each confederate

representing a separate level of reward; and third, as a contin-

uous variable, with each confederate's score on the physical

attraction subscales as a point on a continuum.

Reliabilities for all dependent measures, with the exception

of the Known Interval Scale (KIS) and the second manipulation

check were computed with Cronbach's (1951) Coefficient Alpha.

Reliability coefficients for source credibility measures were:

.91 for composure, .69 for character, .65 for competence, .70 for

sociability and .73 for extroversion. Reliability coefficients

for attraction measures were: .88 for physical attraction, .89 for

task attraction and .79 for social attraction. Reliability

coefficients for the scalar distraction measure and the semantic

differential measure of attitude were .77 and .96, respectively.


Source df SS MS F 2
Due to Model 1 103.19 103.19 15.08 <.05
Due to Error 319 2183.23 6.84
Total 320 2286.43

Treatment Mean D t p1

Experimental 3.99 2.04 4.00 <.05
Control 1.95

1Critical value t.95 = 1.96


Source df SS MS F
Due to Model 1 555.06 555.06 12.24 < .05
Due to Error 319 14460.95 45.33
Total 320 15016.01

Treatment Mean D t p1

Experimental 10.71 4.27 3.55 < .05
Control 6.44

1Critical Value t.95 = 1.96



Source df SS MS F R

Due to Model 4 148.09 37.95 5.28 <.05

Due to Error 315 2216.59 7.01

Total 320 2364.58

Treatment Mean D t p1

12 4.07 2.09 3.80 < .05

2 4.02 2.01 3.60 < .05

3 4.33 2.35 4.29 < .05

4 3.52 1.50 2.88 < .05

1Critical Value t = 2.44

21 = most rewarding, 4 = least rewarding



Source df SS MS F o

Due to Model 4 774.17 193.54 4.29 <.05

Due to Error 316 14241.84 45.06

Total 320 15016.01

Treatment Mean D t pl

12 11.13 4.96 3.38 < .05

2 10.39 3.95 2.78 < .05

3 11.77 5.33 3.87 < .05

4 9.45 3.01 2.14 NSD

Control 6.44

1Critical Value t 95 = 2.44
21 = most rewarding, 4 = least rewarding

Manipulation Checks

As a check on whether the message was persuasive and to

establish a baseline for determining message acceptance, the

attitude scores of the no message control group were compared to

the means of the experimental group.

Analyses of variance, summarized in Tables 1 and 2, yielded

significant F-values for the KIS measure (15.08, p < .05) and

the semantic differential measure of attitude (12.24, p < .05).

Dunnett's t-tests (Winer, 1973) also yielded significant t's for

the KIS (4.00, p A .05) and the semantic differential measure

(3.55, p < .05).

A check on whether the message was persuasive across the four

levels of reward, when compared to the no message control group,

indicated that three of the four confederates differed from the

control. Significant F-values were obtained for the KIS measure

(5.28, p < .05) and the semantic differential measure (4.29, p < .05).

The first three levels differed significantly from the control but

the fourth, and lowest, did not for the semantic differential

measure of attitude (see Tables 3 and 4). Because reward was

found to deviate from linearity and because the least attractive

confederate did not differ from the control, an additional manipu-

lation check was computed with reward treated as a dichotomous

variable. An analysis of variance, summarized in Tables 5 and 6,

yielded significant F-values for the KIS (8.61, p < .05) and the

Semantic Differential measure of attitude (6.12, p < .05) for

reward treatments.




Source df SS MS F p

Due to Model 2 121.47 60.73 8.61 < .05
Due to Error 318 2243.11 7.05
Total 320 2364.59


Source df SS MS F p

Due to Model 2 556.38 278.19 6.12 < .05
Due to Error 318 14459.63 45.47
Total 320 15016.01

All comparisons were significantly different from the control

beyond the .05 level of confidence by Dunnett's t-test (Winer,

1971). Table 7 summarizes the results of these comparisons. The

message had some impact on all experimental subjects. A check on

whether the message was persuasive across distances when compared

to the no message control group indicated that all subjects were


Treatment Mean D t pl

Reward 10.782 4.34 3.37 < .05
3.403 2.05 4.04 < .05
Punishment 10.642 4.21 3.29 < .05
3.883 1.93 3.83 < .05
Control 6.442
1Critical Value t.95 = 2.21

2Semantic Differential measure


significantly different from the control at each distance. Analy-

ses of variance, summarized in Tables 8 and 9, yielded significant

F-values for the KIS (4.46, p < .05) and the semantic differential

measure of attitude (3.29, p <..05). All comparisons were signifi-


Source df SS MS F p

Due to Model 4 126.47 31.61 4.46 < .05
Due to Error 316 2238.11 7.08
Total 320 2364.59


Source df SS MS F p

Due to Model 4 600.73 150.18 3.29 < .05
Due to Error 316 14415.28 45.61
Total 320 15016.01

cantly different from the control beyond the .05 level of confi-

dence by Dunnett's t-test. Table 10 summarizes the results of

these comparisons. The message had some impact on all experimental



Treatment Mean D t p

Threat 11.272 4.83 3.45 < .05
4.103 2.48 4.50 < .05
Close 10.512 4.07 2.88 < .05
3.933 1.98 3.54 < .05
Norm 10.222 3.78 2.70 < .05
3.723 1.77 3.21 < .05
Far 10.862 4.42 3.13 < .05
4.023 2.07 3.77 < .05
Control 6.442
1Critical Value t.95 = 2.44
2Semantic Differential measure

As a check on distraction, two self-report measures were

utilized to test whether the violation of distancing expectations

was distracting. It was expected that deviations from the

established normative distance would be distracting, with the

deviation into the threat threshold yielding the most distraction.

The results of the semantic differential measure failed to find

any significant differences for distraction between distances.

The means also indicated that the close deviation was most

distracting. The analysis of variance and means are presented

in Table 11.


Source df SS MS F p

Due to Model 3 152.24 50.74 1.67 .17
Due to Error 283 8515.63 30.33
Total 286 8737.87

N Distance Mean Distraction
Condition Score

73 Threat 27.39
69 Close 29.07
74 Norm 28.36
71 Far 27.28

The second manipulation check asked subjects to list factors which

made them pay more or less attention to the persuasive message.

This check was administered after subjects had completed the

experiment. No significant differences were obtained between

subjects who listed more or less (X2 = 1.60, df = 1) or between

distances (X2 = 2.13, df = 3).

A final manipulation check found that subjects rated the

confederates'in terms of physical attraction in the expected rank

ordering. From most to least rewarding, confederates ratings were:

22.62, 18.47, 16.58 and 14.08.

Hypothesis 1

Hypothesis 1, which posited a main effect for reward such

that rewarding initiators would produce better outcomes than would

punishing initiators, was partially supported for credibility

and attraction. Significant F-values (df = 3, 271, p < .05) were

obtained for competence (6.79), composure (20.48), sociability

(17.16) extroversion (15.71), social attraction (7.31), task

attraction (7.28) and physical attraction (51.30). For both

measures of attitude, F-values were nonsignificant. (See Appendix

E, Tables 1 through 10, for analysis of variance summary tables.)

It was originally planned to treat reward as a continuous variable,

assuming that it was linear. An analysis of the means, however,

found that some deviation from linearity existed (see Table 12).

Reward, therefore, was treated as a categorical variable. (Power

< .90.)



1 2 3 4

Competence 22.10 22.39 21.66 19.53
Composure 19.10 19.91 17.07 13.28
Character 21.34 21.98 21.50 20.29
Sociability 23.19 25.12 24.22 21.06
Extroversion 22.73 24.27 22.92 20.00
Social Attraction 23.21 22.97 22.58 20.44
Physical Attraction 22.62 18.47 16.58 14.08
Task Attraction 21.60 21.59 21.09 19.39
KIS 4.04 3.96 4.29 3.45
Semantic Differential 11.13 10.39 11.78 9.46

*1 = most attractive; 4 = least attractive

Because later analysis required that reward be treated as a

dichotomy, the two highest rated confederates were collapsed into

a reward condition and the two lowest rated confederates into a

punishment condition. Main effects were obtained on the same

variables as in the previous analysis (see Appendix E, Table 11

through 20). An examination of the means (see Table 13) indicated

that on all variables, rewarding initiators produced better results

than did punishing initiators (power < .90).

Hypotheses 2, 3 and 4

Hypothesis 2. The second hypothesis, which predicted an

interaction between distance and reward, was not supported. One

significant interaction was obtained for extroversion with distance


Reward Punishment

Competence 24.24 20.62
Composure 19.48 15.22
Character 21.65 20.91
Sociability 24.11 22.68
Extroversion 23.46 21.50
Social Attraction 23.09 21.54
Physical Attraction 20.65 15.36
Task Attraction 21.50 20.26
KIS 3.99 3.88
Semantic Differential 10.78 10.65

treated as a continuous variable (F = 8.25, df = 1, 283, p <.05).

The summary tables for the analyses of variance are reported in

Appendix E, Tables 11 through 30. The means for the reward by

distance interactions are reported in Table 14. (Power analyses

ranged from .10 on the nonsignificant interaction to < .90 on the

significant interaction.

Hypotheses 3 and 4. The third and fourth hypotheses predicted

that the relationship among distance, reward and the communication

outcomes of credibility, attraction and persuasion would be

curvilinear in nature, with a close or far distance producing the

most positive outcomes for the rewarding initiator and the normative

distance for the punishing initiator. The hypotheses were analyzed

by treating reward as a dichotomous variable and treating distance

as a continuous variable. Polynomial equations were generated by


Threat Close Norm Far

Rewarding Initiator 22.66 21.83 21.51 23.07
Punishing Initiator 20.53 21.65 20.46 19.97
Rewarding Initiator 19.71 19.62 21.23 21.25
Punishing Initiator 15.63 14.35 16.35 14.51
Rewarding Initiator 21.77 22.23 21.32 21.25
Punishing Initiator 20.68 21.85 20.70 20.51
Rewarding Initiator 23.34 24.17 24.70 24.19
Punishing Initiator 22.18 24.08 21.76 22.82
Rewarding Initiator 23.06 23.37 23.14 24.38
Punishing Initiator 21.87 22.82 20.51 20.92
Rewarding Initiator 23.20 23.00 23.24 22.90
Punishing Initiator 20.16 22.03 21.38 22.62
Rewarding Initiator 20.54 20.40 20.76 20.91
Punishing Initiator 14.63 15.85 15.16 15.85
Rewarding Initiator 21.14 22.17 21.41 21.69
Punishing Initiator 19.28 20.68 20.03 20.59
Rewarding Initiator 3.91 4.19 3.91 3.99
Punishing Initiator 4.27 3.67 3.52 4.04
Rewarding Initiator 10.63 10.63 10.89 11.00
Punishing Initiator 11.87 10.38 9.54 10.74

*With distance treated as a categorical variable

using the linear, quadratic, cubic and quartic terms in the

regression analyses. For rewarding initiators, the cubic or

quartic terms were selected as representative of the Burgoon (1978)

model as each could fit the predicted relationship. The means

were then examined to determine the best fit. For punishing

initiators, the cubic term was selected as best representing the

model. Additionally, significant quadratic terms would also be

indicative of curvilinear relationships. As such, two separate

analyses were computed for the rewarding and punishing initiators.

For each analysis, the regression equation was plotted to determine

if the regression line conformed to the predictions.

Rewarding initiators produced significant curvilinear

relationships for extroversion, competence and task attraction

(df = 1, 134, p < .05). For extroversion, significant partial

were obtained for the linear (F = 3.15), quadratic (F = 5.86),

cubic, (F = 6.88) and quartic (sequential and partial F = 7.39)

terms. For competence, the quartic sequential and partial terms

were significant (F = 4.10). For task attraction, the cubic

(partial F = 3.92) and the sequential and partial terms obtained

significance (F = 4.29). The regression summary tables are reported

in Appendix E, Tables 31 through 40. When plotted, extroversion

and task attraction appeared to conform to the predicted curves

(see Table 14).

The expected relationship between punishing initiators and

distance was not supported (see Appendix E, Tables 41 through 50).

When plotted, all means deviated from the expected pattern for

punishing initiators and seemed to conform instead to the predicted

reward and distance relationships (see Table 14).

Given the lack of significant relationships for both punishing

and rewarding initiators with distance, hypotheses 3 and 4 were

only partially supported.

Supplemental Analyses on Hypotheses 1 4

An examination of the variances associated with the confed-

erates suggested that two other factors may have been operating

to affect perceptions of reward: sex of the confederate and sex

composition of the dyad. Previous research (Burgoon, Stacks &

Woodall, 1977) had defined these variables as rewarding and punish-

ing where females were seen as more rewarding than males and

opposite-sex interactions more rewarding than same-sex interactions.

The analyses were performed in the same manner as before.

Confederate sex. Significant main effects for confederate

sex were obtained for composure (F = 5.14, df = 1, 279, p <.05),

physical attraction (F = 32.07) and Social Attraction (F = 6.67).

Tables summarizing the analysis of variance for each dependent

variable are reported in Appendix E, Tables 51 through 70. The

means are reported in Table 15. The interaction between confed-

erate sex and distance was obtained only for extroversion, with


Outcome Female Male

Competence 21.87 20.90
Composure 18.06 16.45
Character 21.41 21.10
Sociability 23.72 23.00
Extroversion 22.83 22.04
Physical Attraction 19.54 16.18
Task Attraction 21.34 20.44
Social Attraction 22.89 21.65
KIS 4.17 3.69
Semantic Differential 11.46 9.91

distance treated as a continuous variable (F = 4.90, df = 1, 283,

p < .05). The means are reported in Table 16. An examination of

the R2 statistic found no substantial increase in variance for

most dependent variables when taking confederate sex into account.

This, coupled with the lack of significance for main effects and

interactions, made further analysis fruitless. (Power for all non-

significant effects was < .25; power for significant effects was

< .90).

Sex composition of the dyad. Only one significant main

effect was obtained whereby opposite-sex initiators were rated

as more physically attractive than same-sex initiators. No

significant interactions were obtained (see Appendix E, Tables

71 through 90). An examination of the R2 statistic found no

substantial increase in the variance for most dependent variables


Threat Close Norm Far

Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate
Female Confederate
Male Confederate

22.10 21.83 22.13 21.39
20.91 21.65 19.71 21.34

18.97 16.63 19.31 17.11
16.00 17.44 15.77 16.60

21.51 21.91 21.49 20.78
20.85 22.18 20.49 20.91

22.97 24.40 24.05 23.50
22.47 23.85 22.31 23.37

22.31 23.26 22.54 23.28
22.59 22.94 21.03 21.66

22.05 23.60 23.21 22.75
21.12 21.41 21.31 22.74

19.38 19.83 20.13 18.78
15.26 16.44 15.54 17.46

20.97 22.03 21.54 20.86
19.82 20.82 19.80 21.31

4.64 4.37 3.67 3.99
3.47 3.47 3.77 4.05

12.44 11.80 10.08 11.58
9.94 9.18 10.37 10.11

*With distance treated as a categorical variable

when taking the sex composition of the dyad into account. Further

analyses were not attempted. The means are reported in Tables 17

and 18. (Power for all significant effects was < .90; power for

all nonsignificant effects was < .20.)


Outcome Same-sex Opposite-sex

Competence 21.82 21.11
Composure 17.46 17.16
Character 21.51 21.10
Sociability 23.22 23.49
Extroversion 22.49 22.42
Physical Attraction 16.87 16.68
Task Attraction 20.71 21.05
Social Attraction 22.39 22.22
KIS 3.83 4.02
Semantic Differential 10.60 10.80

Hypothesis 5

The predicted main effect for propensity on source credibility,

attraction and message acceptance was not supported. Low propen-

sity subjects, on all outcomes except extroversion, rated confe-

derates as less attractive and credible than did high propensity

subjects. Significant differences in the opposite direction

were obtained for competence (F = 4.95, df = 1, 279, p .05),

composure (F = 11.27) and task attraction (F = 5.05). The analysis

of variance summary tables are reported in Appendix E, Tables 91


Threat Close Norm Far

Opposite Sex 21.04 22.18 20.13 21.13
Same Sex 22.65 21.19 21.94 21.68
Opposite Sex 17.76 17.24 17.00 16.50
Same Sex 17.22 16.77 18.34 17.32
Opposite Sex 20.76 22.55 20.36 20.85
Same Sex 22.17 21.42 21.74 20.84
Opposite Sex 22.66 23.74 23.28 23.53
Same Sex 22.91 23.38 23.17 23.32
Opposite Sex 22.24 23.24 21.97 22.30
Same Sex 22.87 22.94 21.66 22.71
Opposite Sex 21.66 28.55 21.56 22.30
Same Sex 21.52 21.66 23.14 23.32
Opposite Sex 18.34 19.50 18.26 18.75
Same Sex 15.57 16.52 17.63 17.32
Opposite Sex 20.54 22.24 20.46 21.15
Same Sex 20.22 20.45 21.00 21.00
Opposite Sex 4.34 3.73 3.94 3.97
Same Sex 3.56 4.17 3.47 4.08
Opposite Sex 11.88 9.74 10.51 10.73
Same Sex 9.96 11.45 9.89 11.03

*With distance treated as a categorical variable

through 100. For message acceptance, low propensity-to-counter-

araue subjects aDDeared to be more resistant to the message than

hich Drooensity-to-counterargue subjects. The means are reported

in Table 19. (Power analyses ranged from < .10 on nonsignificant

effects to < .90 on significant effects.)


High Low

Competence 22.01 20.90
Composure 18.47 16.31
Character 21.63 20.97
Sociability 23.68 23.11
Extroversion 22.41 22.48
Social Attraction 22.56 22.07
Physical Attraction 18.35 17.57
Task Attraction 21.51 20.41
KIS 4.11 3.80
Semantic Differential 10.94 10.53

Hypothesis 6

The predicted interaction between reward and propensity to

counterargue was not supported when reward was treated as four

levels for credibility, attraction or message acceptance (see

Appendix E, Tables 91 through 100). The means are reported in

Table 20. When reward was treated as two levels (i.e., rewarding

and punishing) a significant interaction was obtained for the KIS

measure of attitude (F = 5.15, df = 1, 283, p < .05). When analyzed,


1 2 3 4

High Propensity 22.63 22.36 22.33 20.39
Low Propensity 21.68 22.42 20.97 18.98
High Propensity 20.18 20.70 18.27 14.04
Low Propensity 18.17 19.12 15.92 12.80
High Propensity 21.44 22.18 21.59 20.79
Low Propensity 21.27 21.79 21.08 19.98
High Propensity 23.16 25.12 24.41 21.64
Low Propensity 23.22 25.12 24.05 20.68
High Propensity 22.75 24.03 22.54 19.93
Low Propensity 22.71 24.52 22.28 20.05
High Propensity 22.97 23.55 22.78 20.64
Low Propensity 23.39 22.39 22.38 20.32
High Propensity 22.38 13.88 17.14 14.71
Low Propensity 22.80 18.06 16.05 13.68
High Propensity 21.66 22.09 21.68 20.43
Low Propensity 21.56 21.09 20.53 18.73
High Propensity 3.60 3.93 4.92 3.82
Low Propensity 4.38 3.98 3.70 3.21
High Propensity 10.16 10.27 12.41 10.68
Low Propensity 11.90 10.52 11.18 8.68

*1 = most attractive; 4 = least attractive

however, the pattern of means deviated from the predicted order.

As predicted, high and low propensity subjects did not differ

significantly across reward (t = 1.44, df = 260, p < .05), while

low propensity subjects differed significantly between rewarding

and punishing sources (t = 1.84, df = 260, p < .05). For punishing

sources, however, low propensity subjects were significantly more

resistant to the message than were high propensity subjects (t =

2.24, df = 260, p < .05). Since high propensity subjects were

predicted to be more resistant to the message than low propensity

subjects, regardless of the reward power of the source, Hypothesis

6 was not confirmed for attitude. All other interactions were

nonsignificant (see Appendix E, Tables 101 through 110). The

means are presented in Table 21. (Power on the analyses ranges

from less than .50 on the nonsignificant analyses to greater than

.90 on the significant KIS analysis.)

Hypothesis 7

The key to hypothesis 7 was a two-way interaction between

propensity and distraction where high or low propensity would raise

or lower attitude scores depending on whether subjects were

distracted or not distracted. Because of the predicted interaction

between reward and propensity, the hypothesized interaction had

to be three-way. The interaction between reward, propensity and

the distance distraction was not supported (KIS F < 1; semantic

differential F < 1). For this analysis, the threat, close and far


Reward Punish

High Propensity 22.49 22.52
Low Propensity 22.01 19.92
High Propensity 20.49 16.45
Low Propensity 18.59 14.27
High Propensity 21.82 21.45
Low Propensity 21.50 20.43
High Propensity 24.15 23.22
Low Propensity 24.07 22.27
High Propensity 23.40 21.42
Low Propensity 23.51 21.57
High Propensity 23.26 21.86
Low Propensity 21.95 21.29
High Propensity 20.60 16.09
Low Propensity 20.69 14.80
High Propensity 21.88 21.14
Low Propensity 21.35 19.58
High Propensity 3.77 4.45
Low Propensity 4.20 3.44
High Propensity 10.22 11.66
Low Propensity 11.28 9.86

were collapsed and treated as distracting and the normative distance

as not distracting. An examination of the means indicated that

nondistracted low propensity subjects at the fourth, and lowest,

level of reward were most resistant to the message while nondis-

tracted high propensity subjects at the second level of reward

were least resistant. The analysis of variance and means summary

table for the KIS measure is reported in Table 22 and the semantic

differential measure in Table 23.

Hypotheses 8 and 9

In order to confirm Hypotheses 8 and 9, the curvilinear

relationships between distance and reward had to conform to the

predicted shaoes and main effects for Drooensitv had to be obtained.

The expected curvilinear relationships were obtained only for

extroversion and task attraction. For each outcome, however,

the propensity main effect was not significant. Hence, any signi-

ficant interaction among distance, propensity, reward and the

communication outcomes of credibility, attraction and attitude could

not be interpreted as supporting the hypothesized shape or

parallelism of the curves (see Appendix E, Tables 111 through 130).

The means are reported in Tables 24 and 25 and significant poly-

nomial terms are summarized in Table 26.



Source df SS MS F P R2

Due to Model 15 114.08 7.60 1.04 NSD .05
Due to Error 271 1974.23 7.28
Total 286 2088.31

Source df SS MS F P

Propensity (A) 1 6.73 6.73 0.92 NSD
Distraction (B) 1 5.07 5.07 0.70 NSD
Reward (C) 3 25.96 8.65 1.19 NSD
A x B 1 17.85 17.85 2.45 NSD
A x C 3 42.16 14.05 1.93 NSD
B x C 3 10.40 3.47 0.48 NSD
A x B x C 3 5.93 1.97 0.27 NSD

Propensity Propensity
Hijh Low High Low

1 3.56 3.56 3.61 4.27
2 5.08 3.47 3.49 4.15

3 4.75 2.96 4.98 3.96

4 3.80 2.79 3.84 3.33

Not Distracted Distracted

*1 = most attractive; 4 = least attractive


Source df SS MS F p R2

Due to Model 15 654.89 43.65 0.92 NSD .05
Due to Error 271 12917.68 47.66
Total 286 13572.57

Source df SS MS F D

Propensity (A) 1 11.94 11.94 0.25 NSD
Distraction (B) 1 25.00 25.00 0.52 NSD
Reward (C) 3 214.80 71.60 1.50 NSD
A x B 1 101.30 101.30 2.13 NSD
A x C 3 152.72 50.90 1.07 NSD
B x C 3 114.35 38.11 0.80 NSD
A x B x C 3 34.75 11.58 0.24 NSD

Propensity Propensity
High Low High Low

1 10.13 10.58 10.17 12.45
2 13.22 9.60 9.17 10.84

3 11.44 8.20 12.71 12.21

4 10.86 8.10 10.60 8.85

Not Distracted Distracted

*1 = Most attractive; 4 = least attractive


Threat Close Norm Far

High Propensity
Low Propensity
High Propensity
Low Prooensitv
High Propensity
Low Propensity
High Propensity
Low Propensity
High Propensity
Low Propensity
High Propensity
Low Propensity
High Propensity
Low Propensity
High Propensity
Low Propensity
High Propensity
Low Propensity
High Propensity
Low Propensity

22.33 22.50
18.90 21.05



21.56 22.50
19.90 21.40

23.72 24.57
20.80 23.75













20.94 20.38
20.05 19.70

17.94 13.13
15.00 15.48

21.41 20.44
20.10 20.57

22.06 22.69
21.50 22.91









4.28 4.57
2.88 3.68

11.18 12.43
8.15 9.57

as a categorical variable

*With distance treated


Threat Close Norm Far

High Propensity 23.07 21.50 22.06 23.60
Low Propensity 23.35 22.18 21.05 22.58
High Propensity 19.00 21.11 19.41 22.47
Low Propensity 20.25 18.06 18.50 17.29
High Propensity 20.87 22.06 21.82 22.47
Low Propensity 22.45 22.41 20.90 20.18
High Propensity 22.53 23.67 24.88 25.53
Low Propensity 23.95 24.71 24.55 23.00
High Propensity 22.80 23.11 23.12 24.67
Low Propensity 23.35 23.64 23.15 24.12
High Propensity 23.00 22.78 22.65 24.80
Low Propensity 23.35 23.34 23.75 21.24
High Propensity 20.07 20.39 20.35 21.67
Low Propensity 20.90 20.41 21.10 20.24
High Propensity 20.93 21.83 21.18 23.67
Low Propensity 21.30 22.53 21.60 19.94
High Propensity 3.49 4.10 4.36 2.96
Low Propensity 4.22 4.29 3.52 4.91
High Propensity 9.80 10.72 11.76 8.27
Low Propensity 11.25 10.53 10.15 13.41

*With distance treated as a categorical variable


Linear Quadratic Cubic Quartic

High Propensity/Rewarding1 3.77 4.57 5.08 6.01

High Propensity/Rewarding1 7.73 7.26 7.73 7.78

High Propensity/Rewarding1 7.89 7.57 7.81 8.04
High Propensity/Punishing2 3.01 4.30 4.93

High Propensity/Rewarding1 5.99 6.94 7.66 8.00

High Propensity/Rewarding1 4.85 6.29 7.35 7.95

High Propensity/Rewarding1 7.70 6.87 5.77
High Propensity/Punishing3 4.07 4.18

High Propensity/Punishing2 4.49 4.04
*p .05
Idf = 1, 60
2df = 1, 60
3df = 1, 61



The underlying rationale for the effects of distraction in a

communicative setting is based upon the disruption of an individ-

ual's ability to counterargue. Several studies have demonstrated

a distraction effect upon subjects' acceptance of a persuasive

message and their perceptions of the source's credibility but

with mixed results. It seems that the nature of the distractor

acts to intervene in the counterarguing process, producing more

resistance or less resistance to the message, depending upon the

nature of the distractor. Additionally, research has attempted

to examine an individual's propensity to counterargue a message

when distracted and the results of the distraction on the source's

credibility. Such research predicts that high propensity-to-

counterargue subjects will be more resistant to both the message

and the source than will low propensity-to-counterargue subjects.

This study attempted to: (1) establish the relationship between

violations of personal spacing expectations as a distractor during

a social influence attempt; (2) to test the relationship between

propensity to counterargue and message acceptance, source credi-

bility and attraction as mediated by both the distractor and the

subject's propensity to counterargue; and (3) to expand previous

research on violations of personal space expectations to persuasive


The results of multiple regression analyses failed to

support the predicted relationships among reward, distance,

propensity and the communication outcomes of credibility, attrac-

tion and oersuasion. For the test of the Burgoon (1978) model,

only two outcomes conformed to the predicted relationships; for

extroversion and task attraction, subjects interacting with rewarding

initiators conformed to the predicted curvilinear relationship.

For subjects interacting with punishing initiators, the obtained

outcomes appeared to conform more to the expected pattern of means

for rewarding initiators. For propensity, where significant

effects were obtained, the effects were in the opposite direction;

low propensity subjects rated the source less credible and attrac-

tive than did high propensity subjects. The relationship between

distance and distraction was not supported. Subjects appeared

to be equally distracted at any distance. The predicted parallelism

of high and low propensity curves for rewarding and punishing

initiators across distance was not found, instead high and low

propensity curves intersected at one or more distances. In the

case where parallelism was obtained, the interaction term failed

to obtain significance and the two curves may have been a chance


Propensity to Counterargue

One major reason for the lack of support for the hypothesized

relationships was the propensity measure. The correlations between

source derogation and counterarguing were not significant, indica-

ting that the relationship between counterarguing was not reliable.

The reliabilities for each scale also indicated that the measure

was not highly reliable (.69 for counterarguing and .54 for source


A second consideration on the propensity measure is its

validity. Validity was analyzed in three ways: first, a correla-

tion was obtained for counterarguing and source derogation between

the propensity measure and another measure; second, the results

of this investigation were compared to those of the Brandt et al.

(1977) investigation, which used a similar measure; and third, the

results themselves were considered a source of validity: if the

predicted relationships obtained, they would be considered support

for the validity of the measure. The three criteria were not met.

The correlations between the essay measure of counterarguing and

source derogation and the propensity measure were nonsignificant.

The scores for the propensity measure did not reflect the results;

high propensity subjects were more susceptible to the message

and source than were low propensity subjects.

The second criterion, that the Brandt et al. results be

comparable to the results obtained in this investigation, might

be an artifact of the method. In the Brandt et al. study, the

interaction was tape recorded, played back and attributed to a

high status source. In the present study, the message was delivered

by a live source with equal status. Brandt et al.'s study utilized

a message created from arguments used in the propensity measure

which their subjects had been previously exposed to via the propen-

sity measure. In their study, they may have inadvertently

inoculated subjects to the message by pretest. High propensity

subjects may have been threatened by the arguments on the pretest

and engaged in more counterarguing upon receipt of the message.

Low propensity subjects, on the other hand, may not have been

threatened if, by nature, they do not key on the arguments and

became more susceptible to the later message. In the present

study, subjects were not exposed to the arguments contained in

the persuasive message.

In terms of credibility results, there are no good explana-

tions for the differences between the two studies. However, it

may have been that the use of a live source in the present inves-

tigation mediated credibility results. High propensity subjects

may have evaluated the sources as poor communicators and not

engaged in counterarguing while low propensity subjects may have

been threatened by the sources and engaged in increased counter-



It was proposed that violations of personal spacing expecta-

tions would be distracting. This effect was not supported.

Subjects reported being almost equally distracted at all four

distances on the semantic differential-like measure used as a

manipulation check. The second manipulation check indicated that

subjects did not differ in the amount of attention paid to the

message. Since this check was obtained after the experiment was

completed, its reliability may be questionable. Assistants also

reported that a majority of subjects did not know they had received

a persuasive message. This finding also casts doubt on the relia-

bility of the measure.

That the distance deviations were not distracting is surprising,

especially at the threat distance. Previous research (e.g.,

McBride, King & James, 1965) had established that extremely close

physical distances produced physiological responses in people.

A deviation to within three inches of the subject should have

been distracting in that the subject should have felt discomfort.

Subjects indicated, however, that the threat distance was slightly

less distracting than the close or norm distances. A more sensitive

measure of distraction is needed in future research which utilizes

distance deviations as a distractor.

Violations of Personal Space Expectations

The test of the modified Burgoon model (1978) of personal

space violations offers limited support for the model. Upon closer

examination, however, stronger support is indicated. The lack of

distance by reward interactions may have been due in part to the

reward variable. It was originally planned to treat reward as a

continuous variable, assuming that it was linear. Because it

deviated from linearity, reward was treated as a dichotomous

variable by collapsing the two highest rated confederates into a

reward condition and the two lowest rated confederates into a

punishing condition. Whether or not this produced punishing

intitators is questionable. The punishing initiators' attraction

ratings for one confederate were found to be greater than the

attraction ratings of Burgoon, Stacks and Woodall's (1977)

punishing initiators, thus more in the range of rewarding initia-

tors. When the means are reinterpreted from this perspective, the

punishing initiators' outcomes conform to the model's predicted

curvilinear relationship between distance and rewarding initiators

for the credibility and attraction outcomes. For the attitude

outcomes, however, the most message acceptance occurred at any

deviation from the norm. Although it was originally posited that

any deviation from the norm for the less rewarding initiator would

cause the subject to derogate the source and produce more counter-

arguments resulting in less message acceptance, subjects apparently

cued in on the positive characteristics of the source and enhanced

the effect of the message.

For rewarding initiators, the results are not easily inter-

pretable. Of the two outcomes which confromed to predictions, task

attraction conformed to earlier findings (e.g., Burgoon, 1978;

Burgoon, Stacks & Woodall, 1977) but extroversion produced a

curve where rewarding initiators were rated most extroverted at

the far and close distances, rather than the inverted U-shaped

curve found by Burgoon, Stacks and Woodall. In terms of message

acceptance, although significant results were not obtained, plots

of the rewarding initiators' curves appeared to conform for the

KIS measure, with the greatest message at the close and far

distances. For the semantic differential measure, however, the

most message acceptance was obtained at the far distance, although

the differences across distances were minimal.

The relationships among distance, reward, propensity and

the communication outcomes of credibility, attraction and

message acceptance are not explanable given the lack of reliability

for the propensity measure. Where significant interactions were

obtained, the high and low propensity curves were not parallel

but intersected at one or more distance. Uhen the individual

curves were plotted, the configuration of means for several out-

comes did appear to be curvilinear; the punishing initiators'

curves again conformed more to the expected relationships for

rewarding initiators.

The finding that two other factors might interact with

distance to affect communication outcomes was unexpected. An

inspection of R2's indicated, however, that in this investigation

they did not improve the amount of variance accounted for.

Additionally, both sex of the confederate and sex composition of

the dyad failed to produce reward main effects on significant

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs