§§§ IIIWHHUWIUJIHIIWIWUIHHII'HIJIHJHIWHHIHH,

Date

ll

MICHOGAN STATE UNI \

[Y
ll
ll ll

ITY LlBRARlES

\lllllll \\l\\\\lll\\\\\l\l

l\\\\\\\\\\l\\\\\ ll

3 1293 01405 6364

This is to certify that the

thesis entitled

Gaining Compliance From People In
Impaired States: How "Friends

Don't Let Friends Drive Drunk

presented by
Steve L. Robbins

has been accepted towards fulfillment
of the requirements for

M.A. Communication

degree in

 

 

Major professor

u—le-gu

 

MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

LIBRARY

Michigan State
Unlverslty

 

 

 

PLACE ll RETURN BOX to roman this checkout from your record.

TO AVOID FINES return on or baton data duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Afﬂnnulvo WM Opportunity Insulator!
WM!

 

GAINING COMPLIANCE FROM PEOPLE IN IMPAIRED STATES:
HOW "FRIENDS DON'T LET FRIENDS DRIVE DRUNK"

BY

Steve L. Robbins

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Communication

1 994

ABSTRACT

GAINING COMPLIANCE FROM PEOPLE IN IMPAIRED STATES:
HOW "FRIENDS DON'T LET FRIENDS DRIVE DRUNK"

BY

Steve L. Robbins

This paper examines the compliance-gaining strategies people use to prevent
drunk others from driving. Speciﬁcally, it investigates the strategies individUals might use
with acquaintances and close others who are intoxicated. A goal-driven approach is taken
using Dillard's framework of primary and secondary goals. It is believed that the level of
importance actors place on different goals affects the choice of compliance-gaining
strategies. Nine hypotheses are stated which predict relationships between goals,
situational variables, and message features. Findings suggest that perceived drunkenness
and relational intimacy signiﬁcantly affect ratings of goal importance. Moreover, results
indicate that goal importance is a good predictor of intervention and persistence. The data
also suggest that secondary goals may not be salient in a drunken driving intervention

situation. Implications are subsequently discussed.

TABLE OF CONTENTS

List of Tables ..................................................... IV
List of Figures ..................................................... V
Introduction ....................................................... 1

Theoretical Framework .............................................. 6

Method .......................................................... 16
Analyses ......................................................... 20
Results ........................................................... 24
Discussion ........................................................ 29
Appendix A ....................................................... 35
Appendix B ....................................................... 36
Appendix C ....................................................... 37
Appendix D ....................................................... 38
Appendix E ....................................................... 39
References ........................................................ 4O

LIST OF TABLES

Correlations Among Primary and Secondary Goals ........................ 37
Correlations Among Goals, Situational Variables and Message Variables ....... 37
Causal Model Scales ................................................ 38
Secondary Goals Scales .............................................. 39

iv

LIST OF FIGURES

Path Model ....................................................... 36

Introduction

A major public service message campaign extolling the dangers of drinking and
driving asserts "ﬁiends don't let friends drive drunk." While the message is appealing in
its simplicity and straightforwardness, it is curiously void of any suggestions as to how
individuals might actually go about preventing intoxicated friends from getting behind the
wheel. The problem of driving under the inﬂuence (DUI) is one with costly and often
tragic repercussions. Recent statistics show that 25-50% of all motor vehicle accidents in
the United States involve persons who were above the legal limit of intoxication (Jaccard
& Turrisi, 1-987; Ungerleider & Bloch, 1987). As much as 65% of single-vehicle fatal
crashes involved drivers who were intoxicated over the legal limit (Ungerleider & Bloch,
1987). Even more disheartening is the fact that a disproportionate amount of these

accidents involve young people between the ages of 15-24 (Swisher & Bibeau, 1987).

Rationale

Combating such a large problem as drunken driving is no easy task. One of the more
effective and popular methods of deterring drunken driving incorporates the use of a
designated driver. This method of preventing DUI calls for a pre-drinking strategy of

choosing a person who will abstain from drinking in order to drive home those who are

intoxicated. It necessitates planning and cooperation among the people involved.
However, there are many instances when people do not pre-plan or when people go alone
to occasions where alcohol is being served. These situations present a unique problem to
message sources; persuading an already intoxicated other to "hand-over his/her keys."
Persuading people who are drunk not to drive is an unenviable yet necessary task.
It is a difﬁcult task in light of the mental condition of the target. Intervening in such a
situation often requires getting a person to cognitively process a message coherently when
it is likely the case that s/he is unable to do so. That people see a need to attempt such
compliance-gaining is supported by previous studies. A study by Hernandez and Rabow
(1987) suggests that a large percentage of people have been in such a compliance-gaining
situation. Furthermore, an even greater percentage of people say they would intervene if
such a scenario confronts them (Hernandez & Rabow, 1987). Several studies suggest that
individuals are more likely to intervene if they know the target and if they have
previously spoken to the target (Newcomb, Rabow, Monto & Hernandez, 1991) Other
studies provide evidence that there is a greater likelihood of intervention in more formal
places like restaurants or bars (Rabow & Hernandez, 1987). Gaining a greater
understanding of how persuaders perceive drunk driving intervention situations and how

those perceptions affect compliance-gaining has pragmatic and theoretical value.

Study Purpose

This study will examine compliance-gaining strategies used by intervenors faced with
intoxicated targets. The study is interested in how intervenors generate
compliance—gaining messages with the goal of preventing drunken targets from driving. It
seeks to answer several questions: (1) What compliance-gaining strategies do people
consider and use to prevent others from drinking and driving? (2) What goals are salient in
this particular situation? (3) What factors might actors consider when deciding whether
to engage in compliance-gaining and to persist in the engagement?

The theoretical framework guiding this study is Dillard's notion of primary and
secondary goals. The paper will begin by discussing situational dimensions as relevant
factors in compliance-gaining contexts. Next, will then move to a brief explication of
Dillard's multiple goals framework. Based on this framework, nine hypotheses are stated.

We now move to the relevance of situational dimensions.

Situational Dimensions

In any compliance-gaining situation there are factors which the inﬂuencer must
take into consideration before taking action. Previous research by Cody and McLaughlin
(1980) indicates that there may be salient situational dimensions which affect
compliance-gaining behaviors. The researchers suggest six factors as relevant to college

students in compliance-gaining situations: Intimacy, Personal Beneﬁts, Resistance,

Dominance, Consequences, and Rights. Intimacy refers to the degree to which the source
perceives an interpersonal relationship with the target. Personal beneﬁts refers to any
gains and rewards which the persuader might receive by obtaining compliance from the
target. The resistance factor can be thought of as the degree of resistance which the
source expects the target to offer. Dominance refers to power differences in the
relationship between source and target. Consequences are said to be the negative short
and long-term effects which the source might anticipate as a result of the
compliance-gaining attempt. Finally, the degree to which the source perceives
compliance-gaining as legitimate action for the situation deﬁnes the Rights factor. The
researchers state that some dimensions may be more salient than others given the
situation. However, in a test of the Cody and McLaughlin dimensions, Dillard and
Burgoon (1985) found few effects attributable to the situation and small effect sizes when
effects were found. Their interpretation of the data did not dismiss the possibility that
actors use situational dimensions to formulate message strategies. Rather, it suggests that
the dimensions they tested were not relevant to their particular study. Although a useful
set of situational factors that might inﬂuence compliance-gaining attempts has yet to be
constructed, it is an intuitive notion that situational features impact compliance-gainers
and their message choices. This author believes it ﬁ'uitful to examine situational factors in
terms of their relationship with goals. It is likely the case that situational determinants

impact the evaluation of sources' goals, which in turn inﬂuence the choice of message

strategies. Subsequently, we now turn to a discussion of goals.

A Goal-Driven Approach

Previous research has suggested that a goal-planning-action (GPA) sequence
perspective is an appropriate way of examining how actors produce compliance-gaining
messages (Dillard, 1990; Dillard, Segrin & Harden, 1989). This perspective posits that
the goal of getting another to do something which s/he would not otherwise do motivates
planning, and that planning subsequently guides behavior or action. The present paper
will use the GPA perspective to investigate how individuals go about preventing drunken
others from getting behind the wheel. However, because previous drinking and driving
literature suggests that intervention is more akin to spontaneous rather than pre-planned
or consciously planned behavior (Newcomb, Rabow, Monto & Hernandez, 1991), the
planning step in the GPA sequence will not be discussed. The situation presented to
subjects in the present study severely limits the time between the recognition of a need
for intervention and the intervention itself. Therefore, it is appropriate for this study to
focus on the impact of goals on message selection and message choice while not explicitly

addressing issues relevant to the planning process.

Theoretical Framework

Goals are often deﬁned as future states of affairs which individuals seek to obtain
and/or maintain (Dillard, 1990; see Hobbs & Evans, 1980; see Klinger, 1985). Dillard,
Segrin, and Harden (1989) suggest that there are two general types of goals: primary and
secondary. Primary goals or inﬂuence goals are said to deﬁne the interpersonal situation.
That is, in any inﬂuence situation the primary goal is understood to be the altering of a
target's behavior; whatever that behavior may be. The desire to achieve the primary goal
acts as the compelling force which motivates the actor to action (Dillard, 1990). Dillard
suggests that the importance of the primary goal affects the likelihood that the source will
intervene and persist in intervention. Said differently, the greater the inﬂuencer believes it
important to bring about behavior change in the target, the greater the possibility the
inﬂuencer will act and persist to produce such behavior.

Secondary goals are considered to be goals which are common to individuals across
situations (Dillard, 1990). This general class of goals works to shape and constrain the
actor's compliance-gaining attempt. Because it is said that secondary goals mold
compliance-gaining behaviors, we may expect them to inﬂuence the likelihood of choosing
certain compliance-gaining strategies over others. Dillard (1990) suggests ﬁve categories
of secondary goals: identity, interaction, relational resource, personal resource & arousal
( see Dillard, Segrin, & Harden, 1989). Identity goals are objectives based on one's ethical

and moral principles, rules for living, and desires to act within the bounds of appropriate

personal conduct. They are said to be associated with one's self concept. Interaction
goals are goals which arise from concerns to act appropriately in social contexts.
Operationally, they can be thought of as desires to avoid actions which might threaten the
face of self and other and to communicate in a relevant and efﬁcacious manner. Sources
concerned with relational resource goals are moved to protect or enhance relational
assets. It is expected that such source concerns produce actions and messages which
avoid damaging the relationship between the compliance-gaining interactants. Personal
resource goals motivate actions which maintain and improve upon assets valued by the
source. These assets range from the physical and material to the mental and abstract.
Among these assets are physical safety, time, possessions, happiness, etc. Finally,
arousal management goals have to do with the assumed needs of individuals to maintain
a comfortable state of arousal. While the boundaries of what is meant by comfortable
differ across individuals, it is expected that individuals are concerned with not
over-stepping the bounds of comfortable arousal. These secondary goals are said to run
counter to the primary or inﬂuence goal and are expected to constrain the

compliance-gainer's range of possible inﬂuence actions.

The Present Primary Goal
In the present case, the primary or inﬂuence goal is preventing an intoxicated other

from driving. It is argued that the relative importance of this goal is judged by processing

certain situational information. In a drunken driving intervention attempt a rather salient
situational factor is the consequence of target compliance, or lack thereof. While Cody
and McLaughlin suggest consequences as a salient situational factor, they were referring
to relational consequences which result from a compliance-gaining attempt. There is no
discussion of what this author calls action/inaction consequences. These are
consequences incurred by the target as a result of non-compliance. In the present
situation, non-compliance has potentially lethal consequences: injury and death to self
and others.

Along those same lines, it is assumed that a majority of individuals are motivated
to protect friends or close others from actions that are perceived to be dangerous. Other
things being equal, the greater an individual believes the danger to be, the more important
it is to takes actions to protect the other. Partially because the media and other
organizations (MADD, SADD, The Ad Council) have diffused the consequences (death
and injury) of drinking and driving to much of the public, it is assumed that most people
believe drinking and driving to be dangerous. Perceptions of the degree of perceived
danger are probably a combination of many variables (number of [drinks consumed, time
of day, road conditions, etc.). However, it is believed that perceived drunkenness is the
factor most salient to persuaders in this inﬂuence situation. In the present situation,
perceived drunkenness is believed to be positively correlated with the actor's perception

of danger. Given that perceived danger is said to be positively correlated with importance

of intervention and that perceived drunkenness is positively correlated with perceived
danger, the more an individual believes another to be drunk, the greater the importance
placed on preventing the other from driving.

Another situational factor which is believed to affect ratings of inﬂuence goal
importance is relational intimacy. Simply put, the greater the perceived intimacy between
the source and the target, the more important it is for the source to alter the target's
behavior; in this case, to prevent the target from driving.

It has been argued that intervenors in the present situation place a level of
importance on preventing a drunken acquaintance from driving, and that this level of
importance is evaluated using certain salient situational features: perceived drunkenness
and relational intimacy. The degree to which importance is placed on the goal of
prevention is believed to impact actors' decisions to engage and persist in
compliance-gaining. Dillard (1990) suggests that inﬂuencers weigh the importance of
situational goals, both primary and secondary, to make decisions to engage in
compliance-gaining and to persist in the attempt. Thus, the ﬁrst hypotheses are stated.

H]: Source perceptions of target drunkenness are positively correlated with

importance of the inﬂuence goal.

H2: Source perceptions of relational intimacy with the target are positively

correlated with importance of the influence goal.

H3: As the importance of the inﬂuence goal increases, so does the likelihood

10

that the source will intervene, as well as persist.

Goal importance also affects the amount of cognitive and behavioral effort an actor
expends in interpersonal inﬂuence attempts to bring about desired results (see Dillard,
1990). As the importance of the primary goal increases so does the amount of cognitive
planning and behavioral actions increase. However, it was noted earlier that drunken
driving prevention is considered to be more spontaneous than planned behavior. As such,
the amount of time between the recognition of the need to act and the act itself is
relatively short. This short amount of time places a constraint on the actor. Waldron
(1990) suggests that constraints such as time and number of goals cognitively tax the
actor. Under cognitively taxing situations, actors are more likely to employ direct
compliance-gaining strategies (Waldron, 1990). Direct strategies are said to be those
which quickly and explicitly make clear what is wanted from the target; the quicker and
more explicit the request the greater the level of directness. Hinting and the use of
deception are considered to be low in directness. In the present situation, actors are
expected to be more cognitively taxed as relational intimacy increases. That is, it is likely
the case that the closer the actor feels to the target, the greater the need to attend to
secondary goals in the situation. In this case, concerns with secondary goals are expected
to tax cognitive processes more severely when close friends are involved than when
acquaintances are involved. Thus, hypothesis four is stated.

H4: Relational intimacy is positively correlated with the use of direct

11

compliance-gaining strategies.

The Present Secondary Goals

Persuaders attempting to prevent a drunken target from driving face multiple
competing goals which theoretically constrain and restrict their compliance-gaining
attempt. Under Dillard's framework, among the highly salient goals in this situation are
interaction goals. Asking or telling another not to drive may be perceived by the target as
an attack on his/her abilities to take care of him/herself. In other words, such a situation
is potentially face-threatening to the target and consequently inhibits intervention (see
Newcomb et al., 1991) or requires the actor to manage communication in a non-face
threatening manner. Since the present situation only involves people who know one
another to some degree, it is likely that the persuader is motivated to guard against actions
which might embarrass or degrade the target in the presence of others. Previous research
suggests that the greater the actor's perception of intimacy, the greater the degree of
importance is placed on protecting the face of the target (see Lim, 1990). Consequently it
is expected that there will be a negative relationship between desires to achieve interaction
goals and the use of anti-social compliance-gaining strategies. Anti-social strategies are
globally deﬁned as socially inappropriate behavior. Strategies which include the use of
profanity, threats, punishments, etc., are said to be high in anti-socialness. Strategies

which attempt to steer clear of controversy and avoid threatening or damaging the other's

12

face are said to be low in anti-socialness. Hypothesis ﬁve is thus stated:

H5: There is a negative relationship between the perceived importance of
interaction goals and the use of anti-social compliance-gaining strategies.

Relatedly, sources are believed to be concerned with relational maintenance.
These concerns are captured by the category of goals labeled relational resource goals. It
is likely the case that actors want to enhance or at least preserve the current relationship
they have with their drunken target. In seeking to achieve this goal, sources are expected
to resist using compliance-gaining tactics which might endanger the relationship.
Consequently, it is predicted that greater importance place on relational maintenance will
inhibit the use of anti-social strategies. Hypothesis six is stated:

H6: There is a negative relationship between the perceived importance of
relational resource goals and the use of anti-social compliance-gaining
strategies.

Similar to relational resource goals, another competing secondary goal which
persuaders might consider in this situation are personal resource goals. The present
situation is slightly different than compliance-gaining situations presented in previous
research. In past situations, successfully gaining compliance did not necessitate the
persuader to expend effort after gaining compliance (see Dillard, Segrin, & Harden, 1989).
Presently, if the compliance-gaining attempt is successful, it is likely that the persuader

may have to expend personal effort to get the drunken person home. This might involve

13

driving the target home, calling a taxi, allowing the target to spend the night, etc. All of
these actions may compromise available compliance-gaining strategies if the persuader is
unwilling or unable to carry out the actions. Desires to conserve such personal resources
work against the primary goal by placing greater importance on this secondary goal. This
is expected to affect the use of messages which involve what this author calls cooperative
or compromise strategies. These strategies are similar to the Schenck-Hamlin et al.
strategy of promise, but are not exactly the same. Promises usually involve offering
something which the target might value in exchange for compliance, e. g., "I'll give you ten
bucks if you let me drive you home." Cooperative/compromise strategies do not imply
that the target will receive anything s/he values. For example, the message, "I'll drive you
home if you hand over your keys," is cooperative/compromise because the target
probably does not value a ride home since s/he thinks that s/he can drive him/herself
home. Furthermore, receiving a ride home means that the target must make arrangements
to get his/her car home at a later point. The target probably does not value having to take
such actions. Therefore, it is expected that sources' desires to conserve personal
resources will inhibit the use of cooperative/compromise strategies. Hypothesis seven is
stated:

H 7: There is a negative relationship between the perceived importance of

personal resource goals and the use of cooperative/compromise

compliance-gaining strategies.

14

A possible identity goal that might arise in the present situation involves using
deceptive tactics. It is likely that many persuaders have moral and/or ethical principles
that inhibit the intentional employment of deception or lying as a means to an end,
especially when the target of the message is a friend. It is also likely the case that some of
these persuaders might see deception as the most effective means for gaining compliance.
However, deception runs counter to their moral and ethical standards for living, and as
such, they ﬁnd themselves in a cognitive tug-of-war with the primary goal and identity
goal as opponents. It is expected that the greater the importance placed on identity goals,
the less likely deceptive messages will be used to gain compliance. However, research
suggests that deceptive messages may be employed when (l) authentic communication
fails, (2) deceptive communication is believed to be efﬁcacious, (3) deception is not
expected to be detected, (4) detected deception will not result in unacceptable
consequences (see Booth-Butterﬁeld & Booth-Butterﬁeld, 1987). Furthermore, Neuliep
and Mattson (1990) suggest that deceitful persuaders tend to use rationale or
explanation-based strategies while truthful persuaders opt for positive or negative
sanctions. Consequently, it is predicted that when deception is used in this situation, we
would expect to ﬁnd those messages to be logically or rationally based. Hypothesis eight
is stated:

H8: There is a negative correlation between the perceived importance of identity

goals and the use of deceptive compliance-gaining strategies.

15

H8A: In addition, when deception is employed, deception will be positively
correlated with logically based strategies.

Arousal management goals are also said to affect the compliance-gaining attempt.

These goals suggest that individuals want to maintain a state of arousal with which they
are comfortable (Dillard, 1990). Desires to maintain an acceptable state of arousal might
inhibit inﬂuencers from becoming overly angry or at least overtly angry, and thus, inhibit
the use of anti-social compliance-gaining strategies. Such strategies are only expected to
be used when others have failed. In other words, the more important the inﬂuencer
believes the arousal management goals to be, the less likely s/he will use anti-social
strategies. Hypothesis nine is stated:

H9: There is a negative relationship between the perceived importance of
arousal management goals and the use of anti-social compliance-gaining
strategies.

It is apparent that persuaders in the present situation face multiple competing
goals. Furthermore, it has been suggested that the relative importance placed on these
goals affects their evaluation. Invariably, there is competition between the primary goal
and the secondary goals. Variations in the way individuals handle and resolve the
competition between goals, it is argued, ultimately affects the selection of
compliance-gaining messages. With this in mind, we now turn to the methods which will

be employed to test the previously stated hypotheses.

16

Method
Sample

Participants were undergraduate students enrolled in a communication class at a
large midwestern university. Those who participated received extra credit for their
involvement. After discarding instruments with considerable missing data and those
which suggested response bias or questionable involvement from the participant, N=187
usable instruments remained. Data from these completed instruments were used in
subsequent analyses. The mean age of the respondents was m=19.4 years; 65% were
females; mean years in college was m=2.3 years. Related to drunken driving intervention,
only 14% of participants had never tried to prevent another from driving drunk, and
approximately 17% had been in such a situation more than eight times. 40% of
respondents have had someone try to prevent them from driving drunk and 63% have
personally known someone who has been hurt or killed in a drunken driving accident.
Indeed, undergraduate students are an appropriate sample; they have many opportunities
to be in situations where they may be faced with preventing a drunken other from driving.
Furthermore, these students fall into the age bracket which has a disproportionate amount

of drunken driving related accidents.

Materials and Procedure

The instrument used in the study contained four major sections. In the ﬁrst

17

section participants were given a drunken driving prevention scenario, asked to imagine
themselves in that scenario, and asked to write down what they would do or say to the

person in the situation. The general situation is as stated:

Suppose you see an acquaintance/very good friend (Tim) of yours at a party.

You consider Tim to be an acquaintance because you only talk to him in class

and rarely do anything outside of class/a very good friend because you have

known him for years and do many things together. Near the end of the party

Tim says he needs to go home and pulls out his keys as he begins to leave.

From what you can tell, Tim shows some signs of being slightly/greatly

affected by the beers he has drank. He seems to be slurring his speech a

little/a lot, he's been slightly/extremely loud and obnoxious, and he has a

slightly/very difﬁcult time keeping his balance. What would you do and/or

say to Tim given these speciﬁc conditions?

Four scenarios crossing perceptions of target drunkenness (low = slightly, high =
extremely) and intimacy (low = acquaintance, high = very good friend) were used (see
Appendix A for exact wording of each scenario). Since the research was interested in the
types of messages that would be generated, the scenarios were developed to ensure
minimal "1 would not do anything," type responses. The scenarios were effective in that
sense; only two people indicated they would not take action in the given scenario. These
two respondents were left out of subsequent analyses. Participants were randomly given
one of the four scenarios and a series of items and questions to which to respond.
Obtaining constructed messages might give insight as to possible strategies used in this

particular situation, especially ones which may not be captured by previous

compliance-gaining studies.

18

After generating messages to the given scenario, participants responded to the
second major section. This part of the questionnaire asked participants to respond to a
series of 30 items. The items were designed to measure ﬁve hypothesized factors: 1)
perceived drunkenness, 2) relational intimacy, 3) goal importance, 4) likelihood of
intervention, and 5) likelihood of persistence. Following each item participants were
asked to indicate the extent to which they agreed or disagreed with the statement using a
7-point scale ranging from "very strongly agree" to "very strongly disagree." To offer
some protection from response bias and provide a slightly different measure of the ﬁve
hypothesized factors, the last ﬁve items of the section were 7-point Likert items used to
measure the factors.

The third major section of the questionnaire is similar to the second. Here,
respondents were given a list of 37 items designed to measure the importance of the
secondary goals, taken from Dillard's (1989) typology of goals, in the situation. All of
the items used a 7-point scale. The ﬁrst ﬁve items used a scale ranging from "No
Priority" to "Very High Priority." In the remaining 32 items, respondents rated each
item using "Very Strongly Agree" and "Very Strongly Disagree" as scale anchors.

The ﬁnal section of the questionnaire presented participants with the
Schenck-Hamlin, Wiseman, and Georgacarakos (1982) typology of compliance-gaining
messages adapted to the present situation. This particular typology is chosen because

there is empirical evidence to suggest its appropriateness over others (Dillard, Segrin, &

19

Harden, 1989; Neuliep & Hazleton, 1985). Each participant was asked to indicate on a
7-point scale, ranging from "Not Likely" to "Highly Likely," his/her likelihood of using

each of the strategies.

20

Analyses

T-tests were used as a check of the manipulation of perceived drunkenness and
relational intimacy. Indeed, the manipulation worked for each of the conditions. For
perceived drunkenness there was a signiﬁcant difference in the means between those in
the slightly drunk condition (m =5.82) and the extremely drunk condition (m =6.17), (t
=2.44, p <.05). For relational intimacy, the mean for those in the acquaintance condition
was m :30], and the mean for those in the very good friend condition was m =5.87 (t
=18.57, p <.001).

Constructed messages from section one were independently rated by four trained
assistants for the degree to which they are direct, positive, rational, and antisocial. A
5-point scale was used for these ratings where 1 = low and 5 = high. The use of deception
and the strategy of cooperative/compromise were coded for their absence or presence.
Only two coders were used to code these latter strategies.

Ratings of directness reﬂected the extent to which the message was clear or made
explicit what the source wanted the target to do. Messages in which the source made an
immediate request for compliance were considered high in directness. Messages which
didn't come out with an explicit request, but only hinted for compliance were rated low in
directness. Cronbach's alpha for directness was .91.

Positivity ratings were based on the degree to which the source mentioned or made

explicit the positive consequences of compliance or the negative consequences of

21

non-compliance. Messages which contained references to rewards or beneﬁts are
considered high in positivity. If the message referred to costs and negative outcomes, it
was considered low in positivity. When there was mention of both positive and negative
consequences, the consequences were compared and a global evaluation of positivity was
made. Cronbach's alpha for positivity was .87.

In determining ratings for logic, messages were evaluated for their inclusion of
evidence and compelling reasons. Messages which employ plausible and compelling
reasons for compliance were considered high in logic. The use of hints or directs requests
reﬂect a low degree of logic. Cronbach's alpha for logic was .92.

Ratings of anti-socialness reﬂected a global evaluation of whether the generated
message was socially appropriate or inappropriate behavior. Messages which used
profanity, threats, punishments, or were verbally aggressive were considered to be high in
anti-socialness. Messages which steered clear of controversy, contained few references of
profanity, threats, etc., and were not considered verbally aggressive were rated low in
anti-socialness. Cronbach's alpha for anti-socialness was .94.

Messages were rated as deceptive if the source was thought to have deliberately
lied, hid the truth, misrepresented himself/herself or the situation, or generally attempted
to mislead the target. Messages were not considered deceptive if the source just hints or
never comes around to asking for compliance. Cronbach's alpha for deception was .98.

Cohen's kappa was .92.

22

A cooperative/compromise strategy was considered to be used when the source
offers to do something to get the target home. That is, if the Source indicated willingness
to expend effort and energy in order to see that the target gets home safely then a
cooperative/compromise component was considered to be present. For example, the
source might offer to give the target a ride home or call a taxi for the target. Cronbach's
alpha for cooperative compromise was .87. Cohen's kappa was .81.

In sections two and three, conﬁrmatory factor analyses were performed to assess
the relative validity of the hypothesized factors. In conﬁrmatory factor analysis, the
researcher a priori hypothesizes the existence of factors, designs items to measure those
factors, and subjects the items to at least three separate tests. The ﬁrst test is intuitive,
assessing the face validity of each of the items. That is, do each of the items appear to be
tapping the hypothesized factor and no other factor? Secondly, each of the items is
tested for internal consistency. This statistical procedure tests for the similarity of the
item-to-total correlations of each of the within factor items. Finally, parallelism (external
consistency) is tested. To pass this test, each of the items within a factor must correlate
similarly with other factors. Items which pass these three tests are then considered to be
relatively valid indicators of the hypothesized factors. In general, a good number of the
items were compatible with their respective factors. A list of the factors and the items
which passed the content and statistical tests along with their loadings are presented in

Tables 3 (see Appendix D) and 4 (see Appendix B). These scales were used in

subsequent analyses.

23

24

Results

Hypothesis I predicted that as source perceptions of the target's drunkenness so
would the importance of the inﬂuence goal. In this case the inﬂuence or primary goal is to
prevent the target from driving. The data show strong support for hypothesis one (r
=.70, p <.001).

Hypothesis 2 predicted that the relationally closer or more intimate the source
feels the target to be, the more important the inﬂuence goal becomes. There is strong
support for this hypothesis as well (r =.58, p <.001).

Multiple regression analysis was also employed to test hypotheses l and 2. This
type of analysis allows the researcher to assess the impact of each independent variable
on the dependent variable while controlling for the other independent variables. The
regression of goal importance onto perceived drunkenness and relational intimacy was
found to be signiﬁcant (Multiple R =.78, R2 =.60, F(2, 185) =138.09, p <.001). The
standardized regression coefﬁcients are b =.56, and b =.36 for perceived drunkenness and
relational intimacy, respectively.

The third hypothesis predicted that as the importance of the inﬂuence goal
increases, so will the likelihood that the source will, 1) intervene, and 2) persist. Again ,
the data suggest strong support for these predictions. For goal importance and likelihood
of intervention the correlation is large and signiﬁcant (r = .71 , p <.001). Likewise, there is

strong support for the correlation between goal importance and persistence (r =.81, p

25

<.001).

Additionally, path analysis was performed using Path (Hunter & Hamilton, 1992)
to test the causal model implied in hypotheses one, two, and three. It was predicted in
these hypotheses that perceived drunkenness and relational intimacy have a direct,
positive effect on the judgment of goal importance (i.e., the importance of preventing Tim
from driving). In turn, goal importance directly effects one's likelihood for intervention
and likelihood of persistence within the intervention attempt. Results of the analysis
suggests a moderately good ﬁt to the data, X2 =4.34, df =6, p =.502. The path model
along with the corresponding path coefﬁcients are presented in Figure 1 (see Appendix
B).

Initially, simple linear regression was used to test each of the remaining
hypotheses. Subsequently, logistic stepwise regressions were employed to examine
which, if any, of the goals and situational variables inﬂuenced the message features (i.e.,
directness, positivity, logic, anti-socialness, deception). Stepwise regression allows the
researcher to examine the effect of different combinations of independent variables on the
dependent variable (Schroeder, Sjoquist, & Stephan, 1986). It searches for the "best"
regression equation accounting for variables which have been already entered in the
equation (Pedhazur, 1982). For each of the message variables, the goal scales as well as
the scales representing relational intimacy and perceived drunkenness were tested in the

regression equation as potential inﬂuencers of the dependent variable.

26

The fourth hypothesis predicts that as levels of relational intimacy increase so
will the use of direct compliance-gaining messages. Indeed, there is a signiﬁcant
relationship between the two variables (r =.43, p <.001).

Hypothesis ﬁve predicts that sources concerned with interaction goals are less
likely to use anti-social compliance-gaining strategies. This hypothesized relationship was
not supported by the data (r = .01, p >.05, n.s.).

Hypothesis six predicts that when sources are concerned with relational resource
goals, they will be less apt to employ anti-social strategies. This relationship was also
found to be not signiﬁcant (r = .05, p >.05, n.s.). In fact, the data suggest that none of the
goals (primary and secondary), perceived drunkenness, and relational intimacy have an
effect on the message feature of anti-socialness.

Hypothesis 7 predicts that there is a negative relationship between concern for
personal resource goals and the use of cooperative/compromise strategies. Data analysis
showed support for this hypothesis (r = -.23, p <.01).

Hypothesis eight predicts that sources concerned with identity goals are less
likely to use deceptive message strategies. The data did not fully support this hypothesis
(F -.11, p >.05). F urtherrnore, hypothesis 8A predicts that when deception is employed
there will a greater use of logic or rationally based strategies. The data suggest some
support for the hypothesis, however it is not signiﬁcant at traditional levels of

signiﬁcance (r = -.l4, p <.07, n.s.).

27

Hypothesis 9 predicts a negative relationship between arousal management goals
and the use of anti-social message strategies. The data suggest this relationship to be not
signiﬁcant (r =-.05, p >.05, n.s.).

The stepwise regression of the goal importance scale onto the secondary goals
scales indicates only one signiﬁcant relationship between the primary goal and secondary
goals. There was a strong, positive relationship between the primary goal and the
personal resource goal (r =.50, P < .001). Univariate analyses among the secondary goals
provided four statistically signiﬁcant relationships. Associations were indicated between
the interaction goal and the relational resource (r =.44, p <.001) and arousal management
goals (r =.30, P < .001). The arousal management goal also had signiﬁcant relationships
with the relational resource (r =.15, p <05) and personal resource goals (r =.21, p <.01).
Table 1 presents the correlations among the primary and secondary goals.

Correlations between the message variables (i.e. deception, directness, positivity,
logic, anti-socialness) and the goals and situational variables are presented in Table 2 (see
Appendix C). The data indicate that much of the impact of the goals and situational
variables is on the message variable of directness. Results show statistically signiﬁcant
correlations between directness and the primary goal (r =.32, p <.001), relational resource
goal (r =.17, p <.05), personal resource goal (r = -.30, p <.001), perceived drunkenness (r
=.23,'p <01), and relational intimacy (r =.43, p <.001). However, a stepwise regression

indicates that only relational intimacy should be entered in the regression equation (b =

28

.43, t = 6.4, p <.001, R2 = .18). The other goals and situational variable do not produce
any signiﬁcant increase in R2. It appears that these goals are only correlated with
directness because of their relationship with relational intimacy; that is, the primary goal,
relational resource goal, personal resource goal, and perceived drunkenness are spuriously
related to directness. Only two other signiﬁcant relationships are presented in the table.
Logic is positively correlated with both the primary goal (r =.22, p <01) and relational
intimacy (r =.19, p <.05). Again, stepwise regression indicates only the primary goal
should be included in the regression equation (b = .22, t = 3.00, p < .01, R2 = .05).
Relational intimacy, while its ﬁrst order correlation with logic is statistically signiﬁcant,
accounts for less than one percent of the variance in logic (b = .09, t = 1.02, p > .05). It
appears that most of the relationship between relational intimacy and the use of logic
based compliance-gaining messages is due to the relationship between relational intimacy

and goal importance (primary goal).

29

Discussion

This study examined the messages individual's produce in order to prevent another
from driving while intoxicated. Using Dillard's (1989) theoretical framework of primary
and secondary goals, nine hypotheses were presented as tests for that framework in the
present drunken driving intervention situation. Hypotheses one and two predicted that
the importance of the primary goal (preventing Tim from driving) would be positively
impacted by the degree to which the compliance-gainer perceives Tim to be drunk and the
degree to which he/she feels relationally close to Tim. These hypotheses were supported
by the data. Indeed, these two variables (perceived drunkenness and relational intimacy)
account for 60% of the variance in the primary goal (goal importance). Additionally,
Dillard suggests that as more importance is placed on the primary goal, the greater the
likelihood that the compliance-gainer will intervene and persist in the intervention. This
was the case in the present study. There was a strong, positive relationship between goal
importance and likelihood of intervention and persistence. The causal model represented
by these ﬁve variables was tested with a path analysis; support was found for the model.

The veriﬁed link between perceived drunkenness and goal importance has
important implications; especially in light of the indication that individuals will not only
attempt, but will persist in the attempt to prevent another from driving while impaired.
The data suggest that potential intervenors will take overt action if the goal of preventing

another is important enough. If it is the case that one's perception of another's

30

drunkenness plays large part in elevating the importance of the primary goal, then it
becomes considerably important to educate people about drunkenness cues. It is likely
that many drinkers who are obviously drunk and too impaired to drive will be prevented
from doing so. More dangerous are the drinkers who intoxicate themselves enough to
impair their driving skills, but not enough to make it obvious that they are too
drunk/impaired to drive. If we can educate individuals to be sensitive to subtle
impairedness cues, then the baseline for perceiving drunkenness might be lowered.
Consequently, we might expect to see more people stepping in to prevent an impaired
other from driving at earlier stages.

None of the hypotheses predicting relationships with anti-social message
strategies was supported. It’ is plausible that social desirability inﬂuenced subjects in
their construction of messages. Also, to produce anti-social messages might require a
state of arousal which the given scenarios did not induce. I

Hypotheses 8 and 8A concerning deceptive messages also were not supported.
The ﬁrst predicted that source's concerned with identity goals are less likely to employ
message which utilized deception. One plausible explanation for this ﬁnding has to do
with costs and beneﬁts. It may be the case that concern for violating one's beliefs, values,
ethics, etc., are not salient in a situation where people's lives are on the line. Intervenors
may judge the consequences (e. g., guilt) for deception in this case do not stack up to the

potential rewards (e.g., the saving of a friend's life).

31

The second part of hypothesis 8 predicted that when deception is used in a
message, then one might expect to ﬁnd the employment of logic or rationally based
compliance-gaining strategies. One possible reason for the non-signiﬁcant ﬁnding is the
fact that only 13% (N=23) of the participants constructed messages which utilized
deception. Such a small number of cases makes it difﬁcult to uncover any statistically
signiﬁcant relationships.

The results of this study suggest that, in this context, secondary goals are not
considered greatly or, if they are considered, have little effect on message production.
The primary goal seems to receive most of the attention. Indeed, it has a statistically
signiﬁcant with two of the message variables (i.e., directness, logic) and approaches
statistical signiﬁcance on two others. It is likely that the primary goal of preventing
another from driving drunk overshadows any of the secondary goals. That is, while
source's might consider one or more secondary goals in a drunk driving intervention
situation, the power of the secondary goals to inﬂuence a compliance-gaining message
strategy is small compared to the primary goal.

In sum, the data suggest that situational factors impact a drunken driving
compliance-gaining situation. Individuals, it appears, will intervene and persist in the
intervention in accordance to how important they judge it important to prevent a target
from driving drunk. Moreover, the data suggest that sources' perceptions of drunkenness

and how well they know the target signiﬁcantly impact how important it is to prevent the

32

target from driving.

These ﬁndings have important implications for drunk driving intervention
campaigns. Many of the well known drunk driving PSA's created by the Ad Council
suggest that "friends don't let friends drive drunk." In many respects the results of this
study support such a statement. However, it seems drunkenness it not the only concept
here worth focusing on. Impairedness is something future intervention campaigns may
want to focus on. For sure drunk individuals are impaired, but it is not necessarily the
case that impaired people are commonly deﬁned as drunk. Campaigns to educate people,
especially young people, need to put a focus on impairedness cues which are likely not
the same as drunken cues. The former are expected to be more subtle than the latter. To
the extent that the recognition of these impairedness cues will lead to a higher rating of
goal importance, we might expect people who have learned to recognize these cues will
intervene and persist at an earlier stage of "drunkenness."

As for the message features, it appears that relational intimacy plays the biggest
role in inﬂuencing I the degree to which a compliance-gaining message is direct. According
to this study, it is the only variable which should be included in the regression equation.
Other variables, while signiﬁcantly correlated with directness, appear only to be
correlated with directness spuriously. The degree to which logic is used in a
. compliance-gaining message is predicted somewhat by the level of importance placed on

the primary goal. However, only ﬁve percent of the variance in logic is accounted for by

33

goal importance. Apparently, there are other variables which affect the use of logic.
None of the goals and situational variables were found to impact the use of the other
message features.

Because of the inherent limitations of a paper and pencil study such as the present
one, it would be fruitful to go out in the ﬁeld to capture compliance-gaining messages in
real drunken driving intervention situations. Real situations would put constraints on
sources which a laboratory situation is incapable of doing. Sources would not only be
taxed cognitively, they would also likely experience physical arousal; two variables which
undoubtedly would affect the messages which are produced. While the call to the ﬁeld is
not new, it does seem like the logical next step for such a line of study. Results of future
studies of this sort might potentially save many lives which otherwise would have been

taken as a result of drinking and driving.

APPENDICES

34

Appendix A: Study Scenarios

Scenario 1

You see an acquaintance (Tim) of yours at a party. Tim is a male of average height and
weight. You consider Tim to be an acquaintance because you only talk to him in (classes
you have together and rarely do anything outside of class. Throughout the party, you
notice him drinking. Near the end of the party Tim says he needs to go home and pulls
out his car keys as he begins to leave. From what you can tell, Tim shows few if any
signs of being affected by the beers he has drank. What would you do or say to Tim
given these speciﬁc conditions? Use the space below to write down your response.
Please be as speciﬁc as you can about any actions you might take. For instance, if you

decide to say something to Tim, please write down exactly what you would say to him.

Scenario 2

You see an acquaintance (Tim) of yours at a party. Tim is a male of average height and
weight. You consider Tim to be an acquaintance because you only talk to him in classes
you have together and rarely do anything outside of class. Throughout the party, you
notice the him drinking. Near the end of the party Tim says he needs to go home and
pulls out his car keys as he begins to leave. From what you can tell, Tim shows some
signs of being affected by the beers he has drank. What would you do or say to Tim
given these speciﬁc conditions? Use the space below to write down your response.
Please be as speciﬁc as you can about any actions you might take. For instance, if you

decide to say something to Tim, please write down exactly what you would say to him.

35

Scenario 3

You see an acquaintance (Tim) of yours at a party. Tim is a male of average height and
weight. You consider Tim to be an acquaintance because you only talk to him in classes
you have together and rarely do anything outside of class. Throughout the party, you
notice Tim drinking. Near the end of the party Tim says he needs to go home and pulls
out his car keys as he begins to leave. From what you can tell, Tim shows signs of being
greatly affected by the beers he has drank. What would you do or say to Tim given these
speciﬁc conditions? Use the space below to write down your response. Please be as
speciﬁc as you can about any actions you might take. For instance, if you decide to say

something to Tim, please write down exactly what you would say to him.

Scenario 4

You see a very good friend (Tim) of yours at a party. Tim is a male of average height and
weight. You consider Tim a very good friend because you have known him for years and
do many things together. Throughout the party, you notice the him drinking. Near the
end of the party Tim says he needs to go home and pulls out his car keys as he begins to
leave. From what you can tell Tim shows few if any signs of being affected by the beers
he has drank. What would you do or say to Tim given these speciﬁc conditions? Use the
space below to write down your response. Please be as speciﬁc as you can about any
actions you might take. For instance, if you decide to say something to Tim, please write

down exactly what you would say to him.

 

36

Appendix B:

Figure 1
Path Model

 

I — —

Likelihood
of

Perceived

Drunkenness

Intervention

     

Goal
Importance

  
   
   
     

. Likelihood
Relational of
Intimacy Persistence

 

 

Note: The model provided a moderately good ﬁt to the data and was thus accepted, X2 = 4.34, df = 6, p =
.502.

37

Appendix C:

Table 1

 

Primary 1 .0 - - - - -
Interaction .06 1 .0 - - - _
Relational .01 .44“ * * 1 .0 - - ..
Personal" -.50*** -.02 .06 1.0 - -
Identity .00 .07 .01 .14 1.0 -

Arousal -.12 .30*** .15* .21” .10 1.0

   

 

Interaction .09 . 1 2 .00 .02 .01
Relational .00 . 17* .03 .03 .05
Personal .08 -.30*** .04 -.10 -.09
Identity .01 .04 -.04 .03 -. 12
Arousal -.02 -.01 -.04 .05 -.05
Per. Drunkenness .14 .23” -.05 .07 -.04
Rel. Intimacy -.03 43*" -.08 .19* .14

Note: I'p<.05; I""‘p<.01; “‘p<.001

38

 

Appendix D:
Table 3
Causal Model Scales

Loadings
Perceived Drunkenness (alpha = .78)
.78 I think Tim would be putting others in danger if he attempted to drive. (20)
.80 I think it would be unwise for Tim to drive. (22)
.64 How impaired from drinking do you think Tim is? (29)
Relational Intimacy (alpha = .84)
.88 If Tim were to move across the country, I would still try to keep in touch with him. (7)
.69 It would affect me deeply if something bad happened to Tim. 21)
.83 How close a relationship do you feel you have with Tim? (27)

Goal Importance (alpha = .72)

.43 I would not waste my time trying to stop Tim from driving home. (S-R)

.48 I think it is important for Tim to be prevented from driving. (10)

.47 Although Tim should not drive in his current state. it matters little to me if he does. (12-R)
.56 Keeping Tim from getting behind the wheel is of concern to me. (14)

.54 Someone should persuade Tim not to drive home. (23)

.85 How important is it for you to prevent Tim from driving? (28)

Intervention (alpha = .77)

.71 If Tim wants to drive, I would not do anything to stop him. (l-R)
.82 I would mind my own business and let Tim do what he wants to do. (3-R)
.67 How likely are to intervene to prevent Tim from driving home? (26)

Persistence (alpha = .91)

.76 I would show persistence in trying to prevent Tim from driving home. (8)

.88 I would not give up until I was successful at preventing Tim from driving home. (9)

.88 I would be willing to expend a lot of effort to prevent Tim from driving home. (19)

.85 If you were to intervene, how persistent would you be at trying to prevent Tim from driving
home? (30)

 

Note: Numbers in parentheses denote the position of the item in the questionnaire. The letter "R" indicates the
item was reverse scored.

39

 

Appendix E:
Table 4
Secondary Goals Scales
Factor
Loadings
Identity (alpha = .80) .
.66 I would be concerned with not violating my own ethical standards.(46)
.82 1 would be concerned with maintaining my own ethical standards. (48)
.70 acting in ways consistent with my values and beliefs is important. (54) '
.65 I would be concerned about being true to myself and my values. (65)
Interaction (alpha = .67)
.91 How high a priority is it for you to avoid embarrassing Tim in preventing him from driving? (39)
.55 Trying to avoid doing things which might make Tim look stupid is the least of my concerns. (47-R)
.48 I would avoid saying things which might embarrass Tim. (67)
Relational Resource (alpha = .64)
.75 How high a priority is it for you to avoid damaging your present and/or future relationship with Tim
in preventing him from driving? (40)
.80 I would be concerned about maintaining my relationship with Tim. (51)
.33 If I had to hurt my relationship with Tim to prevent him from driving, I would do it. (64-R)
Personal Resource (alpha = .67)
.52 How high a priority is it for you to limit the amount of time, energy, and/or effort you spend in
preventing Tim from driving home? (41)
.58 I would rather not have to give Tim a ride home as a means for preventing him from driving. (56)
.81 I would be willing to spend a lot of time and energy to prevent Tim from driving. (68-R)
.44 Protecting my own interests as I try to prevent Tim from driving is important to me. (75)
Arousal Management (alpha = .91)
.82 How high a priority is it for you to avoid getting mad, upset, nervous, angry, etc., in preventing Tim
from driving? (43)
.50 Maintaining control of my emotions is important as I attempt to stop Tim from driving. (45)
.45 I would not care if I got angry in trying to prevent Tim from driving. (70)

 

 

Note: Numbers in parentheses denote the position of the item in the questionnaire. The letter "R" indicates the
item was reverse scored.

40

References

Booth-Butterﬁeld, S, & Booth-Butterﬁeld, M. (1987). WWW;
WWW Paper presented at the annual

convention of the International Communication Association, Montreal Canada.

Cody, M. J., & McLaughlin, M. L. (1980). Perceptions of compliance-gaining situations:
A dimensional analysis. MW, 47, 132-148.

Dillard, J. P. (1990). A goal-driven model of interpersonal inﬂuence. In J. P. Dillard

(Ed) Seskmaﬂcmnhance. WWW (pp
41 -.56) Scottsdale, AZ: Gorsuch Scarrisbrick.

Dillard, J. P., & Burgoon, M. (1985). Situational inﬂuences on the selection of
compliance-gaining messages: Two tests of the predictive utility of the

Cody-McLaughlin typolOgY- Qammunicatichcncgranhs 52, 1985-

Dillard, J. P. Segrin, C, & Harden, J. M. (1989). Primary and secondary goalsinthe

production of interpersonal inﬂuence messages. MW, 56,
19- 38.

Fitzpatrick, M. A., & Winke, J. (1979). You always hurt the one you love: Strategies
and tactics in interpersonal conﬂict. WM, 27, 3-11.

Hernandez, A. C. R., & Rabow, J. (1987). Passive and assertive interventions in public

and private drunken driving situations. WM 48,
269-271.

Hobbs, J. R. & Evans, D. A. (1980). Conversation as planned behavior. Cognitive
m, 3, 275-310.

Jaccard, J. & Turrisi, R. (1987). Cognitive processes and individual differences in
judgments relevant to drunken driving. WW Bmhclch,
53, (1), 135-145.

Klinger, E. (1985). Missing links in action theory. In M. Frese & J. Sabini (Eds), Goal

 

41

We: (pp. 311-321). Hillsdale, NJ: Lawrence Erlbaum.

Lim, T. (1990). Politeness behavior in social inﬂuence situations. In J. P. Dillard (Ed),
“10' on at.” It"H.. no t In no. rift ‘U‘ -.°‘,(pp.
41- 56). Scottsdale, AZ: Gorsuch Scarrisbrick.

Neuliep, J. W. & Hazleton, V. (1985). A cross cultural comparison of Japanese and

American persuasive strategy selection. WW
W3, 9, 389-404.

Neuliep, J. W. & Mattson, M. (1990). The use of deception as a compliance-gaining
strategy W 16 (3) 409-421

Newcomb, M. D., Rabow, J., Monto, M., & Hernandez, A. C. R. (1991). Informal
drunk driving intervention. Psychosocial correlates among young adult women and

men W21 (24) 1988- 2006

Pedhaznr E J (1982) WWW Research..an Chicago:
Holt, Rinehart, and Winston, Inc.

Rabow, J. & Hernandez, A. C. R. (1986). College students do intervene in drunk driving
situations. SceiclcgJLaniSmialResearch 70. 224-225.

Schenck-Hamlin, W. J., Wiseman, R. L., & Georgacarakos, G. N. (1982). A model of
properties of compliance- -gaining strategies. WW1, 30, 92- 100.

Schroeder, L. D., Sjoquist, D. L., & Stephan, P. E. (1986). Understandingﬁegmssign
Analysis. Newbury Park: Sage Publications.

Swisher, I. D. & Bibeau, D. (1987). Who's driving who home?: Assessing adolescent
drinking and driving- .IaurnalniAlmhclmdDngducaticn, 32, (3), 25-37.

Ungerleider, S. & Bloch, S. A. (1987). Perceived effectiveness of driving—driving

countermeasures: An evaluation of MADD. WM 40, (2),
191-195.

Waldron, V. R. (1990). Constrained rationality: Situational inﬂuences on information
acquisition plans and tactics. WM, 57, 184- 201.

42

Wiseman, R. L. & Schenck-Hamlin, W. J. (1981). A multidimensional scaling validation
of an inductively-derived set of compliance- -gaining strategies. W

m, 48, 251 -.270

MICHIGAN Stars UNIV. LIBRARIES
lllWIlllllllllllllllllllllllllIlllllllllllllllll”HI
31293014056364