1.5. . , V V ,3: a 3....“ . .23.... 2.. V V . 1 En. , . . V flak; «rd... .1: V ... ... ,... .23.? 5...... 3 fiuwfiaa. : .fi .. , .... #31? y a. v . . $3.95 "me... $.43 .. 2.4. . .339. 45.4 .x aw gums.“ nu... iciih . . . 35.35. . . T 1. a... J... V . V . , .wfiiEu: .: : .: {.sz 5% .x 1.7;... : .3. . 2:34.33 . 2.. V Hymn?» at... . . \ .e I] I I’Ab i. ufifliwwwésfiz y. I. THEBlS l t } ' ‘ lIBRARY Mm Michigan State i. University _.._... _J This is to certify that the thesis entitled The Negativity Effect in a Dyadic Context: Effects on Attitude Toward Target Object, Likability, Credibility, Similarity and Uncertainty Reduction of Interactants presented by Jina Hyo-Jin Yoo has been accepted towards fulfillment of the requirements for M.A. degree in Communication MW Major professor Date I037 I“ 0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution _.._ _____.. -_ , __ _ _ , PLACE IN RETURN BOX to remove this checkout from your record. To AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE i \ a J 6/01 cJCIRC/DataDUGpGS-nm THE NEGATIVITY EFFECT IN A DYADIC CONTEXT: EFFECTS ON ATTITUDE TOWARD TARGET OBJECT, LIKABILITY, CREDIBILITY, SIMILARITY AND UNCERTAINTY REDUCTION OF INTERACTANTS By J ina Hyo-J in Yoo A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 2001 ABSTRACT THE NEGATIVITY EFFECT IN A DYADIC CONTEXT: EFFECTS ON ATTITUDE TOWARD TARGET OBJECT, LIKABILITY, CREDIBILITY,'SIMILARITY AND UNCERTAINTY REDUCTION OF INTERACTANTS By Jina Hyo-J in Yoo This study examined how the negativity effect functions in dyadic interpersonal communication in terms of attitude change and impression formation of the interactants. One hundred and fourteen participants were assigned to dyadic interactions with strangers and were asked to complete the given tasks: to find 10 positive aspects, 10 negative aspects, or 5 positive and 5 negative aspects of either Michigan State University or cell telephones. No differences emerged among the topics for any of the dependent variables. Findings indicated that (1) as. predicted, sharing negative information had significantly more capacity to change attitudes about the target object in a negative direction ( E (2, 111) = 7.58, p < .05), and (2) there was a constant, though nonsignificant, positive linear relationship across the impression formation variables such that perceived likability, perceived credibility, perceived similarity and uncertainty reduction were highest in the negative group, and lowest in the positive group. C0pyright by Jina Hyo-J in Yoo 2001 ACKNOWLEDGMENTS I would like to specially thank my advisor, Dr. Sandi Smith, for her inspiration and encouragement. Without her effort and affection, I would have not finished this. Also thank you to Dr. Frank Boster and Dr. Chuck Salmon for your precious time and guidance. I also thank you to Dr. Tim Levine who helped me throughout the thesis with his in—depth knowledge of statistics. I would also like to thank my family. Mom, Dad, Jae, and Brian. . .who always pray for me. I love you, I always have, and I will always. I thank God for letting me meet these wonderful people. iv TABLE OF CONTENTS LIST OF TABLES ................................................................................... vi LIST OF FIGURES ................................................................................. vii LIST OF FIGURES ................................................................................ viii INTRODUCTION .................................................................................... 1 CHAPTER 1: LITERATURE REVIEW .......................................................... 1 The Negativity Effect ..................................................................... 1 The Negativity Effect and Impression Formation ...................................... 6 Attribution Theory Hypothesis .......................................... .. ...6 Cost Orientation Theory ........................................................ 12 Perceived Similarity and Uncertainty Reduction Theory ............................ 16 CHAPTER 2: METHODS ......................................................................... 19 Overview of Design ....................................................................... 19 Variables of Interest ....................................................................... 20 Participants ................................................................................. 21 Procedure ................................................................................... 21 CHAPTER 3: RESULTS ........................................................................... 23 Independent Variables .................................................................... 23 Dependent Variables and Scale reliability ............................................. 23 Hypothesis 1 ............................................................................... 26 Hypothesis 2 ............................................................................... 27 Hypothesis 3 ............................................................................... 27 Hypothesis 4 ............................................................................... 28 Hypothesis 5 ............................................................................... 28 CHAPTER 4: DISCUSSION ..................................................................... 29 Limitation .................................................................................. 32 Conclusion ................................................................................. 34 BIBLIOGRAPHY ................................................................................... 36 TABLE & FIGURE ................................................................................. 43 APPENDICES ....................................................................................... 50 LIST OF TABLES Means, standard deviations, and t-test on different conversational topics ................... 44 Means, standard deviations, and coefficient alphas for the scales ............................ 45 Means, standard deviations, and coefficient alphas for the scales with items deleted. . ...46 Means of attitude change, perceived likability, perceived credibility, perceived similarity and uncertainty reduction ........................................................................... 47 Percentages of attitude change on message valence ............................................ 48 vi LIST OF FIGURE Trend Analysis of Valence of Judgment on Attitude Change in Hypothesis 1 ............. 49 vii Introduction Do not let any unwholesome talk come out of your mouths, but only what is helpful for building others up according to their needs, that it may benefit those who listen. Ephesians 4:29 Words have power. One word can save a person by enhancing self-esteem and self-worthiness, however one word also can destroy a life by harming one’s reputation. For example, the various studies on the negativity effect have shown widely that negative information is more pervasively powerful to change one’s attitude than positive information. Also negative information is more memorable than positive information due to its extremeness (Kellermann, 1984). The negativity effect and its extremity biases are commonly found. Studies have consistently indicated that impression formations are inordinately shaped by an actor’s more negative and more extreme attributes, traits, and behaviors. In other words, the extremeness of the negativity effect is more influential in impression formation than positive information (Anderson, 1965, 1968, 1974; Bimbaum, 1972, 1973, 1974; Carlston, 1980a, 1980b; Dreben, Fiske, & Hastie, 1979; Feldman, 1966; Fiske, 1980; Hamilton & Huffman, 1971; Hodges, 1974; Lampel & Anderson, 1968; Oden & Anderson, 1971; Warr, 1974; Wyer, 1974; Wyer & Watson, 1969, cited in Skowronski & Caelston, 1989). The negativity effect is generally applied to interpersonal communication in terms of evaluation in impression formation — how does negative information shape impressions and attitudes toward the target person or the target object? In various situations, negative information has been shown to have more extreme valence than positive information, especially in the formation of an impression of a target person (Kellermann, 1984), and from the evaluations of presidential candidates (Klein, 1991) to the evaluations of job applicants (Springbrett, 1958 cited in Allen, 2000). A considerable amount of research demonstrates that negative information is more powerful in terms of impression formation of the target person or target object, but what about the impression formation of the communicators in a dyadic context when sharing different valences of information? How would they perceive each other when sharing negative information in a dyadic situation as opposed to sharing positive information or combined information? Under the assumption of the negativity effect, this research paper will explore how sharing of negative information influences the communicators’ perceptions of each other in a dyadic interaction. First, this research will measure the magnitude of attitude change toward the target object when given a different valence of judgment. Next, it will examine communicators’ impression formation, especially likability and credibility judgments toward one another, while the communicators are generating positive, negative, or combined information about a target object. Finally, perceived similarity and uncertainty reduction in different message valence situations will be assessed. Chapter 1 Literature Review The Negativity Effect The negativity effect occurs when negative information is disproportionately weighted more heavily when compared to positive information about social stimuli, including objects, events, and people (Kellermann, 1984, 1989). Kellermann (1984) provided an excellent overview of this negativity effect. describing it as follows: “the tendency for negative information to be weighed more heavily than positive information when forming evaluations of social stimuli”(p. 37). Across widely varying objects, events, settings, and persons, positive experiences or positive features of stimuli have been found to be less influential in the formation of j udgments than are negative experiences or negative aspects of stimuli. Kellermann elaborates more on the negativity effect in terms of impression formation, “Not only is negative information weighted more heavily than positive information in the initial formation of impression, but negative information exhibits a greater capacity to alter an already existing impression” (pp. 37- 38). Taylor (1991) also has noted the tendency for negative information to result in a greater mobilization of an organism’s physiological, cognitive, emotional, and social responses. She calls these instances of greater sensitivity to negative information, the “negativity bias.” The negativity effect has been faithfully illustrated in research on impression formation. Numerous studies of impression formation have found that negative information is weighted more heavily than positive information as impressions of others are formed, since negative information tends to have stronger power in forming overall impressions of the target person or target object. In fact, according to Richey et al.’s findings (1975) on the power of negative information, one sentence of negative information could neutralize five sentences of positive information. Furthermore, Anderson (1965) found that negative adjectives more strongly influenced the formation of an impression than positive adjectives about the target object. According to his research, negative adjectives have a stronger modifying capacity than do positive adjectives in terms of evaluation. Moreover, he argues that observers’ impression should be neutral when the target object has the same number of positive and negative characteristics, but this is not the case in reality. Even if the relative number of positive and negative information is controlled and balanced, negative impressions tend to emerge toward the target object (e.g., Brinbaum, 1973; Fiske, 1980. cited in Wojciszke et al., 1993). Furthermore, negative information about the target object is more extreme which suggests that negative information is given more weight in impression and attitude formation toward the target object, thus yielding a negativity effect (Wojciszke et al., 1 993). In addition, negative information is accorded greater weight in the assessment of the likableness of the target person as well (Hamilton & Zanna, 1972). Negativity in information is further posited to be more resistant to impression change than is positive information in a variety of contexts. The political campaign is the most obvious example of testing the negativity effect in reality because it has pervasively verified that negative information is more important than comparable positive information in evaluations of presidential candidates (Klein, 1991). Generally, presidential candidates are expected to behave in a positive manner and to possess positive characteristics, including personal achievement, talent, and hard work which are the factors that produce the final crop of presidential contenders. However, the history of American electoral politics has various examples of campaign downfalls brought about by candidate scandals or wrongdoings. Therefore, Klein believes that voters usually have a positivity expectancy towards candidates, and, negative information about candidates might loom large in voters’ minds. Klein tested the hypothesis that the negativity effect characterized impressions of presidential candidates, focusing especially on National Election Study surveys from 1984 and 1988. He concluded that personality weaknesses were more important than personality strengths in determining the public’s evaluations of the candidates and the ultimate voting decisions in elections in 1984 and 1988. In summary, the fact that negative information is weighted more heavily than positive information when forming evaluations of social actors and events has been well established by various research studies. However, the effect has not been as well established with target objects, therefore, afier reviewing the aforementioned research on the negativity effect, the first hypothesis is presented: H1: In a dyadic communication context, participants who engage in exchanging only negative information about a target object will exhibit a greater attitude change about the target object than will participants who engage in exchanging either only positive information or combined (equal amounts of positive and negative) information about the target object. The negativity issue in interpersonal communication is viewed as an issue of person perception, which is how individuals form impressions and perceptions of a person as information becomes available to them (Allen & Burrell, 2000). Most of the negativity effect research focused on impression formation and person perception in a variety of contexts, such as the evaluation of applicants during employment interviews (Bolster & Springbett, 1961; Crissy & Regan, 1951; Kellermann, 1984), and attitudes toward law enforcement officers (Bayley & Mendelsohn, 1969; Jacob, 1971; Parks, 1976, cited in Kellermann, 1984), in terms of the information receiver’s perspectives. In this research, however, a different perspective on the negativity effect will be examined. This paper will inspect how the negativity effect influences forming mutual impressions of each communicator in a dyad, in terms of perceived source likability, credibility, similarity and uncertainty reduction, rather than the impression of the target person. An important question arises: How are the communicators’ impressions of one other affected differentially when they are exposed to other positive, negative, or equal amounts of positive and negative information? In order to explain further the process of how negativity effects might operate during impression formation of the other interactant, source factors, specifically, perceived source likability, credibility, similarity and uncertainty reduction incorporated with the negativity effect, need to be examined. The Negativity Effect and Impression Formation Theoretically, two approaches to the negativity effect are present: (a) the attribution theory hypothesis and (b) the cost orientation hypothesis (Allen & Burrell, 2000; Klein, 1991). The explanation based on attribution theory is rooted in the inforrnativeness of sources, whereas cost orientation is framed in terms of the motivation to account for the impact of negative information (Allen & Burrell, 2000). Attribution Theory Hypothesis Generally, attribution theory focuses on a target person’s behavior and reflects the attitude formation of the observer toward the target person. Although this research will basically focus on the attributions toward a target object, it is important to know how attribution theory has been developed in order to incorporate it with the negativity effect. Attribution theory suggests that any behavior that deviates from socially prescribed norms will be seen as indicative of “true” personality (Klein, 1991), or in the case of a target object, a true essence. Since social norms generally mandate and reward positive behaviors or attributes, negative actions or attributes are perceived as more reflective of dispositional characteristics (Klein, 1991), especially in the case of a target person. The discrepancy between social norms and perceived characteristics directly concerns the informativeness of the stimuli. For instance, if a target person behaves in a deviant manner, deviant behavior distinguishes from normal behavior, and it is perceived as more informative than normal behavior. As a result, this informativeness of the stimuli (a deviant behavior) functions as the major feature to weigh more heavily on the negative information about the target person’s behavior. A cognitive explanation of the negativity effect in the evaluation of social stimuli is available. This cognitive explanation relies on a common assumption concerning the frequency distribution of negative and positive outcomes, based on the Infonnativeness Hypothesis (II-I) by Fiske (1980). The TH suggests that since most life outcomes are perceived to be positive, the negativity effect in evaluation ironically originates from the pervasive existence of the positivity bias. According to DeCoto and Keuthe, cited in Kellermann (1984), positive interpersonal relationships are viewed to be more common than negative ones. As a result of the pervasive positivity bias, the 1H conceptually posits that when an individual actually receives negative information, it is weighed more heavily than positive information. Kanouse and Hanson (1971) suggested that the reason for the greater weighting of negative information is that it represents a “figure-ground” phenomenon. They continued that most evaluations and expectations are positive, thus negative information would be weighted more heavily as a result of contrast effect. Indeed, a great deal of evidence exists for a general positivity bias, or “Pollyanna Principle” that has been proposed as positive events are judged more likely than unpleasant events, even when the true possibilities actually appear to be equivalent (Matlin & Stang, 1987). The assumption of the TH has its theoretical base in this “Pollyanna Principle” that since extremely bad outcomes are less frequent than extremely good ones, the negative information becomes distinct in one’s conception. Various research confirms the assumption of the 1H, including pleasant interpersonal relationships are expected more often than negative relationships (DeSoto & Kuethe, 1959, cited in Amabile & Glazebrook, 1981); in a free association task, participants respond with pleasant words more frequently than with unpleasant words (Boucher & Osgood, 1969, cited in Amabile & Glazebrook, 1981); and people tend to think about positive events and items more often than negative ones (Bradburn, 1969, cited in Amabile & Glazebrook, 1981). Incorporated with this “Pollyanna Principle”, since the [H posits that the frequency of an event is inversely related to its informativeness, negative information is weighted more heavily in impression formation and decision making because negative events or outcomes are perceived as less frequent in daily lives. Given the explanation of the III, the negativity effect relates to the attribution of personality traits in terms of non-normative behavior as well. Individuals seem to be more stimulated by new and sensational information than by ordinary and uninteresting information. According to Fiske (1980), normative behavior of a third person is deficient in its informativeness to explain the attribution of personality disposition. In other words, it is unclear whether the behavior of the actor is due to situational forces, social influence, or a reflection of the social actor’s personality. However, non-nonnative behavior is defined as the extent to which the behavior fails to conform to social norms and consequently is viewed as being indicative of an individual’s personality. Given the assumption that normative behavior connotes positive information, such as socially desirable behavior, then negative information should be more informative than positive information in impression formation as it is viewed as an anomaly. Since non-normative information offers a more sensational and vivid experience for the receiver, individuals who receive negative information might have better recall and more concrete attitude change about the target person or object (Wojciszkwet a1, 1993). One of the reasons why the negative information has potentially more capacity to change one’s attitude about the target person or object is due to its characteristics of extremeness. Fiske (1980) postulated that extreme and negative events are the determinants of informativeness, and over 90% of the variance in one’s judgments about the target person or target object can be accounted for by characteristics of informativeness. The II-I posits a possibility of increasing the perceived similarity between interactants in a dyadic setting by agreeing on negative information about the target object. Suppose a person is giving negative information about the target object or person, affirmation by the listener validates the speaker’s subjective view of the target object or person (Alicke et al., 1992), which might create more perceived similarity between two interactacts due to the informative characteristics of negative information. Therefore, this paper posits the possibility that agreeing on negative information about the target object might bring out the perceived similarity between interactants more than agreeing on positive information, since negative information presents more sensational and vivid experience to the interactants. Attitudinal similarity has been repeatedly shown as a main aspect to enhance interpersonal attraction (See Bercheid, 1985; Byme, 1971). According to Byme, the higher perceived similarity level between two individuals, the greater their likability to each other will be (1961, 1969, 1971). Byme continued that attitudinal similarity is the major source of reward in dyadic encounters, especially in a first interaction, where people usually believe strangers are not similar to themselves (Kellermann, 1984), thus, perceived similarity after a certain social interaction can be a major determination of liking. Several consistent findings have emerged in the research on the positive linear relationship between perceived similarity and liking (Cherry, Byme, & Mitchell, 1976; Bond, Byme, & Diamond, 1968; Byme, Griffitt, & Golightly, 1966; Byme & Griffitt, 1966; Byme, Ervin, & Lamberth; 1970, Curran & Lippold, 1975; Griffitt, 1969, cited in Ah Yun, 1996). Burleson (1992) and his colleagues also have found that the perceived similarities in cognitive complexity are strongly related to the source likability. Burleson concluded that the results indicated that similar level of cognitive and communication ability might lead to enjoyable interaction. In this research, it is hypothesized that the interactants will perceive that by agreeing on negative information higher perceived similarity exists. Several studies have covered the receiver’s perceptions of negative evaluators, in terms of likability. For instance, Amabile found that negative evaluators were perceived as significantly less likable than positive evaluators, although negative evaluators were perceived as more intelligent, competent, and expert than positive evaluators (1981). Aronson and Worchel (1966) also found that people view negative evaluators more negatively than positive evaluators, when they themselves are the subject of those evaluations, although the suggested implication here is that when a third party was evaluated, an observer will view a negative evaluator as more intelligent than a positive evaluator. In summary, in terms of likability, the general tendency toward 10 negative evaluators is to dislike them but perceive greater credibility as a reverse tendency. However, this paper will present several counterarguments to the previous findings on the relationship between likability and the negativity effect. First of all, role- playing was involved in much of this research. In general, previous studies have followed the patterns of a participant playing the role of an observer or an information receiver, whereas the confederate plays the communicator, or an information giver. The discrepancy between an information giver and an information receiver can be explained by the actor-observer asymmetry (Jones & Nibett, 1972, cited in Malle & Knobe, 1997). Such actor-observer asymmetry is the classic approach of attribution research. It explains that actors and observers differ in how they explain a given negative behavior, with actors preferring situation causes and observers preferring person causes (Malle & Knobe, 1997). Suppose a clear distinction exists between observer and actor; then the actor’s behavior of expressing negativity can be viewed negatively by the observer, as if it is his/her personality disposition to complain or to criticize. However, in this study, a slightly different approach will be applied: no explicit and rigid distinction between the evaluator and the information receiver will be present, since both participants will spontaneously change their roles, actively taking each other’s perspectives in dyadic context with both of them generating valenced information about the target object. In this paper, it is hypothesized that when sharing negative information with another person about a target object, the perceived likability of the other should be increased. Thus, after reviewing source likability and the negativity effect, in terms of its informativeness, the following hypothesis is presented: 11 H2: In a dyadic communication context, participants who engage in exchanging only negative information about a target object will rate greater personal likability of the other than will participants who engage in exchanging either only positive information or combined (equal amounts of positive and negative) information about the target object. Cost orientation theory Another proposed source of the negativity effect is the cost orientation, or risk aversion, explanation, which provides an explanation for the negativity effect in terms of primary motivation (Klein, 1991). This hypothesis postulates that people are risk oriented and more strongly motivated to avoid potential costs than to approach potential gains. In other words, practically, the distress that people report in association with the loss of a given quantity of money typically exceeds the amount of pleasure associated with gain of that same amount (e.g., Kahneman & Tversky, 1984, cited in Ito et al., 1998). This explanation is quite different from the attribution model, which focuses on the impact of informativeness and the implications of attributions, as the cost orientation hypothesis provides a motivational explanation based on risk (Allen & Burrell, 2000). Overall, the desire to avoid costs is believed to increase motivation to change attitudes toward the target object more than the desire to obtain rewards when a stimulus is complex. Various risk-taking studies show that most of participants appear to concentrate on the negative consequences of losing rather than the positive consequences of winning. Thus, potential costs are regarded more importantly by people in determining an overall evaluation of a risk-taking situation than are potential gains (Kats, 1964; Retting & Pasamanick, 1964; Solvic & Lichtenstein, 1968, cited in Amabile & Glazebrook, 1981). Therefore, in summary, negative information is weighted more 12 heavily due to an increasing motivation to avoid costs. When applying this cost orientation hypothesis in interpersonal communication, the cost orientation hypothesis predicts the following: In forming a first impression of a stranger, the perceiver may be primarily concerned with determining whether or not the person poses a substantial threat for future interactions with him, rather than focusing on those desirable qualities that might be a basis for a positive relationship with the person. If this motive to avoid personal costs in interpersonal relations is characteristic of people as perceivers, then it would lead to giving disproportionately heavy weight to unfavorable information in the impression formation process (Hamilton & Zanna, 1972, p.205 cited in Kellermann, 1984). In general, individuals would be more sensitive about negative consequences others may bring than positive opinions of other people for the reason that people are risk averse; they are more concerned with the likelihood of negative outcomes than positive outcomes (Klein, 1991). Caccioppo (1994) proposes that the negativity effect has the tendency to cause more change in output in the negative motivational system as compared with the positive motivational system because of a tendency for the negative motivational system to respond more intensely than the positive motivational system to comparable amounts of activation. As a consequence of the negativity effect, attitudinal and behavioral expressions are more strongly influenced by negative input than positive input (Ito, et al., 1998). Cost orientation theory posits the possibility of increasing the perceived 13 credibility of the interactants in a dyadic context. This research paper will test the relationship between the negativity effect and the likelihood of “perceived credibility,” which refers to a judgment made by a perceiver pertaining to the believability of a communicator (O’Keefe, 1990). Furthermore, this research tests the notion that the communicator who gives more credible information (about the world) will be perceived as more credible in a dyadic context (Hamilton & Zanna, 1972). When testing the credibility of the message (e. g. world information), several studies consistently found that negative information is viewed as more credible than positive information (Leventhal & Singer, 1964 cited in Kellermann, 1984). Negative criticism is one of the forms of negative information. Negative criticism represents an objective and unfavorable observation about an object or event (Alicke et al., 1992). One of the reasons that perceivers see negative criticism as more intelligent than positive criticism is that negative evaluations normally make use of less frequent, more impressive vocabulary words or more refined sentence structures (Amabile, 1981). The unique selection of words used to describe negativity can emphasize its characteristics of intelligence and sophistication of the negative information. Furthermore, negative evaluators, who employ the unique selection of words, were also perceived as more intelligent, competent, and expert than positive evaluator. Amabile (1983) found that participants who are particularly concerned with others’ perceptions of their intelligence would tend toward negative criticism as a strategy of impression management. In other words, intellectually insecure individuals can use negative information to attempt to gain listener’s esteem. According to Hamilton & Zanna (1972), since negative information is perceived to be more informative than 14 positive information, when hearing negative information most people are more likely to activate their motivation and cognition to avoid the cost. Thus, the source seems to be perceived as more credible than the source of a message that helps to obtain a reward. The other account for the reason why the negative information is perceived to be more credible is that negative evaluations are relatively rare, therefore they will be regarded as more useful than positive evaluations. Thus, after reviewing source credibility and the negativity effect in terms of the cost orientation theory, the following hypothesis is presented: H3: In a dyadic communication context, participants who engage in exchanging only negative information about a target object will rate greater personal credibility of the other than will participants who engage in exchanging either only positive information or combined (equal amounts of positive and negative) information about the target object. These two theoretical discussions present the basic informational and motivational characteristics of the negativity effect, which lead to impression formation of the information receiver’s perception toward a target object as well as the impression formation of each interactant. In summary, the informativeness hypothesis posits the possible outcome of perceived source likability, since agreeing on negative information functions as a stimulus that increases perceived similarity between interactants. The cost orientation theory, which postulates the unconscious motivation activated to avoid costs, posits that negative information increases the perceived credibility of interactants in a dyad, since the negative information itself is perceived more credible than the positive information. Based on these possible outcomes, the following questions need to be addressed: When incorporated with the negativity effect and impression formation of 15 each interactant in a dyadic setting, especially regarding the characteristics of the source (in terms of perceived likability and perceived credibility), can sharing negative information function to reduce one’s uncertainty toward other in a dyad? Also, how does sharing certain types of information affect one’s perception of the other interactant in terms of perceived similarity? Perceived Similarity and Uncertainty Reduction Theory Perceived similarity - the degree to which people believe another’s characteristics are similar to their own - is sufficient to induce the attraction to others. Byme used the reinforcement theory to explain perceived likability through perceived similarity. In his experiment, a stranger with similar attitudes is liked better than one with dissimilar attitudes (1961). When people find similarities in a number of areas, including behavior, attitudes, beliefs, values, abilities, and personality, people generally find that positively rewarding (cited in Infante et al., 1997). Byme suggests that agreement on perceived similarity leads to likability because it enhances the individual’s confidence that his/her opinions or beliefs are correct, which is a significant factor in reducing uncertainty in an interpersonal communication. He favors reinforcement theory that proposes that similarity and agreement are rewarding, or reinforcing, and therefore, perceived likability increases. In other words, an interactant in a dyadic setting will end up liking the other interactant who agrees with a certain opinion, which leads to an increase of perceived similarity, because he/she is associated with the reward (Erber & Erber, 2001). Consequently, people are more attracted to the possible source of positive rewards, as perceived similarity and perceived likability are usually interdependent with one other. 16 Uncertainty Reduction Theory (URT) is often utilized in order to explain many of the research findings concerning the similarity-attraction relationship. One core assumption of this theory is that when strangers meet, they seek to reduce uncertainty about each other in order to increase their ability to predict their partner’s behavior in given situation (Berger, 1979). According to Berger and Calabrese’s URT, the greater the perceived similarity between communicators in a developing relationship, the more likely attraction will exist (cited in Infante et a1, 1997). In other words, to the extent that people are not sure about their views, agreement from the other interactant is able to provide “consensual validation” while reducing uncertainty toward each other (Goethals, 1986, cited in Erber & Erber, 2001). In this research, it is hypothesized that agreeing on negative information toward a target object will increase perceived similarity between both interactants, due to the informativeness hypothesis that agreeing on more vivid and informative information has more impact than agreeing on neutral information. Another theorem concerns the intimacy level of the communication content (Berger, 197 9), which states that as the communication content becomes more personal, perceived similarity and liking between communicators increases (cited in Infante et al, 1997). Sharing negative information toward a target object, in a form of complaining behavior, is usually considered to be more personal, as it might function as an emotional release from fi'ustration to “get it off one’s chest” (Alicke et al, 1992, p.287). Sharing negative information also has the intent of altering another’s behavior or changing an aversive situation, thereby serving an interpersonal function (Kowalski, 1996). In other words, sharing negative information is more personal than sharing positive information, and expressions of agreement on negative information reinforce the relationship 17 (Kowalski, 1996). Sharing negative information, or complaint behavior, may be used strategically as a tool of self-presentation (Kowalski, 1996). Complaining is defined in Webster’s Third International Dictionary as “expressing discontent, dissatisfaction, protest, resentment or regret” about the target object or person. However, Kowalski believes that people may complain, not because they are actually dissatisfied with a target object, but because the complaint serves as a way of creating an impression of similarity by conforming to the opinions and values around them (1996). Sharing negative information about a target object can enhance sympathy and understanding from the listener, as well as obtain value consensus for a person’s subjective interpretation of events (Alicke et a1, 1992). Alicke et al. continues that since the process of affirmation by the listener validates the complainer’s attitude about the target person or object (1992), it may serve as the interpersonal link between two interactants. People may complain not because they are dissatisfied but because the complaining behavior serves as way of creating an impression of similarity by conforming to the attitudes and values of the other interactants (Brehm, 1992). Goffinan (1959) also suggests that agreement satisfies conversational norms by avoiding embarrassing or annoying challenges to the complainer’s judgment about the target person or object. For example, if two people go to the same school and one begins to share negative information about the school, the other may also give negative information as a means of conforming to the other’s opinion. While confirming the other’s negative views of the school, the perceived agreement would be increased between two interactants, thereby inducing the perceived similarity ultimately leads to perceived liking. Therefore, in terms of complaint behavior, sharing 18 negative information is able to increase perceived similarity of each communicator in a dyadic interaction context. Sharing more personal information enhances perceived similarity between communicators, and the uncertainty level goes down, whereas the perceived likability goes up. Therefore, after reviewing the relationship between URT and perceived similarity, the fourth and fifth hypotheses are presented: H4: In a dyadic communication context, participants who engage in exchanging only negative information about a target object will rate more perceived similarity toward each other than will participants who engage in either positive information or combined (equal amounts of positive and negative) information about the target object. H5: In a dyadic communication context, participants who engage in exchanging only negative information about a target object will reduce more uncertainty toward each other than will participants who engage in either only positive information or combined (equal amounts of positive and negative) information about the target object. Chapter 2 Methods Overview of Design In this section of the paper, the methods used to test the proposed hypotheses are explained. The two independent variables, (1) valence of judgment (exchanging equal amount of positive and negative information vs. only positive information vs. only negative information), and (2) different conversational topics (Mid-Western University vs. cellular phone) were examined. The dependent variables included (1) attitude change toward the target object (See Appendix A and Appendix B), (2) perceived likability of 19 each communicator (See Appendix C), (3) perceived credibility of each communicator (See Appendix D), (4) perceived similarity, and (5) and uncertainty level of each communicator in a dyad-context (See Appendix E). Vflables of Interest Two independent variables were examined in this research. The first independent variable was the valence of judgment. It involved three levels: exchanging equal amount of positive and negative information, only positive information, or only negative information. The second independent variable was different conversational topic, either regarding a non-distant object (Mid-Westem University) or distant object (cellular phone). The first dependent variable is the perceived likability of interactants. The liking measure was developed by Rubin (1970, 1974a, 1974b) as a self-report measure originally used for the empirical study of attraction in romantic relationships. He differentiated loving from liking by referring to liking as “general, favorable evaluations of one person by another” (Tardy, 1988, p.272). The items in the scale were selected from a series of pretests and initial studies. A factor analysis of the scale indicated that it was unidimensional (Rubin, 1970, cited in Tardy, 1988). Various studies yielded good estimates of internal reliability. Reliability coefficient alphas ranged from .71 to .83 in previous research. Credibility scale (McCroskey, 1966) items are intended to assess whether the communicator is in a position “to know the truth, to know what’s right or correct” (O’Keefe, 1990, p.132). The reliability coefficient alpha for this scale has been .82 or greater. 20 The scale for perceived similarity and uncertainty level of each communicator comes from Parks and Adelman (1983). It was originally designed to assess the correlation between uncertainty and stability in premarital romantic relationships. Reliability coefficient alpha for the scale for perceived similarity has been .82, and for the scale for uncertainty has been .80 or greater in previous researches. The attitude toward a specific object was created and designed originally for this study, in order to measure the general attitude toward the target object. Ten items in the questionnaire are used to measure the audience’s general attitude about the Mid-Westem University or cellular phones. Participants One hundred and fourteen undergraduate students enrolled in two sections of an entry-level communication course at a large mid-westem university earned course credit for participating in this study. As an induction check, the experimenter asked the students in the dyads if they had not met before, and if they had not, they were classified as strangers. One hundred percent indicated they had not met before. The experiment included 38 males and 76 females. Procedure Permission and aid from teaching assistants was obtained. One week prior to the study, the teaching assistants distributed a survey to the target sample, assessing their attitudes toward the Mid-Westem University (See Appendix A) and cellular phones (See Appendix B). Each participant was then contacted to schedule a dyad-session in a laboratory room. At that time, the experimenter made sure that each individual was matched with a stranger of the same sex in order to minimize variability and to maximize 21 the experimental objective. In order to minimize the possible outcome of subject bias, the induction check on how well they knew one another was tested at the beginning of the experiment to make sure they were strangers to each other. When participants entered the laboratory room, the experimenter greeted them and provided an overview of the procedures, secured informed consent and then elaborated on the cover story. Also the experimenter explained that this experiment was designed to see how a given task enhances interpersonal communication. Participant pairs in the combined information group were asked to converse with their partner in order to make a list of 5 positive and 5 negative aspects of the given topic, either the Mid-Westem University or cellular phone, and the order was reversed in half of the sessions to minimize the possible outcome of the bias. Participant pairs who were in the positivity group were asked to make a list of 10 positive aspects of the given topic. By contrast, participant pairs in the negativity group were asked to make a list of 10 negative aspects of the given topic. Then the experimenter went outside, and participants began to converse with each other in order to complete the given task. After the task, the experimenter separated the pairs into two different rooms, and asked each participant to fill out (1) Rubin’s Liking Scale (See Appendix C), (2) McCroskey’s scale for the measurement of ethos (See Appendix D), (3) Parks and Adelman’s perceived similarity and Uncertainty scale (See Appendix E), and (4) either the scale measuring attitude toward the Mid-Westem University (this scale contains exactly the same contents as Appendix A, but had a different look, in order to prevent the participants from recalling the survey from one week before) or the scale measuring attitudes toward cellular phones (this scale also contains exactly the same contents as 22 Appendix B). After the session, the experimenter debriefed those participants involved in the experiment. Chapter 3 Results Independent Variables For independent variables, originally, two different conversational topics were randomly given to participants: the attitude toward the Mid-Western University as a non- distant topic, and the attitude toward cellular phone as a distant topic. T-tests produced a non-significant difference between two topics at an alpha level of .05: for the attitude change scale, (1 =-.567, df = 112, n.s), for the perceived likability scale, (1 = -.421, df = 111, n.s), for the perceived credibility scale, (t = -.122, (if = 112, n.s), for the perceived similarity scale, (I = -.3 77, df = 112, n.s), and lastly, for uncertainty reduction scale, (1 = .148, df = 112, n.s). As Table 1 summarizes, the effect of different conversational topics was not statistically significant. Also, a two-way ANOVA was conducted in order to determine if an interaction effect between two topics was found. The result suggested the lack of interaction effect between two topics, (E(2, 113) = 1.343, n.s). Therefore, these two topics were combined together for purposes of further analysis. Dependent Variables and Scale Reliability In order to make a specific prediction about how the given topic and condition might fit in the model, several methodological issues needed to be resolved. The factor structure of the items employed to measure individual attitudes about the attitude object and participants’ responses to the scale were first analyzed using the Confirmatory Factor Analysis (CF A) with Hunter and Lim’s PACKAGE. This computer program provides 23 factor loadings based on the centroid solution, computes predicted and obtained results based on factor loadings and the specified model, and then calculates deviations by subtracting predicted correlations from obtained correlations. CFA provides a means to estimate the parameters of a measurement model and to assess the unidimensionality of scales (Hunter & Gerbing, 1982). In order to evaluate the measurement model for unidimensionality, the items of a unidimensional cluster must pass statistical test of both internal consistency and parallelism. Then the magnitude of these deviations is tested against sampling error to determine the fit of the model. A detailed description of the procedure of CFA may be found in Hunter and Gerbing (1982). CFA affords a means to test internal consistency, or item homogeneity, by examining how individual scale items will cluster together in a single construct factors. Internal consistency in a single factor suggests that the items from a single construct will cluster together in a linear fashion as indicators of the specified underlying construct. A factor will be regarded as internally consistent when respondents’ responses to one item in the factor are similar to their responses made to the other items hypothesized to be a part of the same factor. A factor may be considered internally consistent if the deviations between the predicted correlation and observed correlation are not significant at the p = .05 level. CFA also tests for parallelism, or item heterogeneity, which is a test of external consistency. Parallelism is stringent test set which assesses the deviation between observed correlation and predicted correlation is at the p_ = .05 level of significance. The test of parallelism allows one to identify scale item that may demonstrate a significantly varied pattern of correlation with other measure. 24 The test of internal consistency for attitude change yielded one significant deviation at p< .05, with standard score coefficient alphas of .71. For the perceived likability scale, it yielded four significant deviations at p < .05 with standard score coefficient alphas of .91. Past research has shown that credibility has multiple dimensions, such as competence, trustworthiness, expertise, etc., however, the CF A here showed that the scale was unidimensional. The perceived credibility scale yielded two significant deviations at p < .05 with a standard score coefficient alpha of .79. The test of parallelism indicated that significant deviations at p < .05 were found across both of the items in 3 and 33. Therefore, in order to increase the coefficient alpha as well as have better fit of the model, item 3 and item 33 were deleted. Also there was one item (item 37) that was a weak indicator showing a distinctively low factor loading. It was also deleted to increase the reliability of the scale and a better fit of the model. After deleting item 3, item 33, and item 37, the coefficient alphas of each scale were increased as well as the fit of the model. The standard score coefficient alpha for the scale of attitude change increased from .71 to .72, the standard score coefficient alpha for the scale of the perceived similarity increased from .76 to .83, and uncertainty reduction increased from .74 to .84 after deleting three items in the study. See Table 2 for each of the scales’ means, standard deviations, and coefficient alphas with item 3, item 33, and item 37. Also see Table 3 for each of the scales’ means, standard deViations, and coefficient alphas after deleting items. All scales, including attitude change, perceived likability, perceived credibility, and uncertainty reduction were found to pass tests of internal consistency and parallelism after deleting three items. Therefore, given the sample size of 114, the results indicated the fit of the model. 25 Hypothesis 1 The data were consistent with the first hypothesis that posited that there would be greater attitude change about a target object in participants who engaged in exchanging only negative information than participants who engaged in exchanging either only positive information or combined (equal amounts of positive and negative) information. In order to test this, the valence of judgment variable was created. A one- sample t-test was conducted first, in order to confirm whether attitude change in the three different valence of judgment groups were significantly different from no attitude change. In both the positivity groups and combined groups, the attitude change was not significantly different from zero (2 > .05). However, in the negativity group, the data were consistent with hypothesis one in that the attitude change was significantly different from zero (t = 3.80, df= 37, p_ < .005). An ANOVA with contrasts was also conducted in order to determine whether attitude change was greater in the negative group. A significant main effect was found for the valence ofjudgment on attitude change ( E (2, 111) = 7.58, p_ < .05, n2 = 0.06). In other words, participants who engaged in exchanging only negative information in a dyadic interaction had significantly greater attitude change about a target object in a negative direction (M = -.33), than participants who engaged in exchanging either only positive information (M = -8.33E-02) or combined information (M? -.l7). Table 4 provides the overview summary of the experimental means and standard deviations. A trend analysis was also conducted to determine the presence of a linear trend in attitude change as negativity increased. A significant linear trend was found for the valence of judgment on attitude change (F (l , 113) = 7.61 , p<.05). Sharing negative 26 information had more capacity to change attitudes toward the target object than did sharing combined information or positive information (see Figure 1). Hypothesis 2 The second hypothesis stated that the participants who engaged in exchanging only negative information about a target object would rate greater personal likability of the other than would participants who engaged in exchanging either only positive information or combined (equal amounts of positive and negative) information about the target object. Testing this hypothesis was done by comparing the score of perceived likability in the three different valence groups. Oneway ANOVA with contrasts was used to test mean differences among these three different valence of judgment groups( E(2, 110) = 1.50, t = .97, n.s). The mean differences were not significantly different from other groups (t_= .96, p = .464). Although as the Table 4 shows, the mean differences were in the expected direction in that the negativity group has the highest mean (M = 5.81), the combined group has the second highest mean (M = 5.73), and the positivity group has the lowest mean (M = 5.56). Hypothesis 3 Hypothesis three stated that participants who engaged in exchanging only negative information about a target object would rate greater personal credibility of the other than would participants who engaged in exchanging either only positive information or combined (equal amounts of positive and negative) information about the target object. Results of the oneway ANOVA with contrasts indicated that the mean difference due to the valence of judgment were not significantly different ( E(2, 113) = 2.08, t = .95, n.s). However, as Table 4 showed, the mean differences were in the 27 expected direction in that the negativity group has a higher score on perceived credibility (M = 5.52), than either the combined group (M = 5.50) or the positivity group (M = 5.32). Hypothesis 4 The fourth hypothesis that claimed that participants who engaged in exchanging only negative information about a target object will increase more perceived similarity toward each other than would participants who engage in sharing either only positive information or combined (equal amounts of positive and negative) information about the target object. The oneway AN OVA with contrast results indicated that the data were not consistent with the hypothesis ( £12, 111) = .88, t = .69, n.s), although the means were in the expected direction in that the negativity group has a higher score on perceived similarity (M = 5.14), than either the combined group (M = 5.06) or the positivity group (M = 4.92). (See Table 4). Hypothesis 5 The final hypothesis that claimed that participants who engaged in exchanging only negative information about a target object will reduce more uncertainty toward each other than would participants who engage in either only positive information or combined (equal amounts of positive and negative) information about the target object. The oneway AN OVA with contrast results indicated that the data were not consistent with the hypothesis ( E(2, 111) = 1.55, t = 1.48, n.s), although the means were in the expected direction in that the negativity group has a higher score on uncertainty reduction (M =- 3.90), than either the combined group (M = 3.50) or the positivity group (M. = 3.49) (See Table 4). 28 Chapter 4 Discussion This study was designed to extend the application of the negativity effect to interpersonal communication in terms of attitude change and impression formation of the interactants in dyadic communication. In particular, efforts were made to learn the function of the negativity effect on the perception of ineractants’ likability, credibility, similarity, and uncertainty reduction of each other. First of all, no significant difference was reported between the two different conversational topics: the attitude toward the Mid-Western University as a non-distant topic, and the attitude toward the cellular phone as a distant topic. Since T-tests produced a non-significant difference between the two topics at an alpha level of .05 on all dependent variables and no interaction effect between them was detected, the two conversational topics were collapsed. It was interesting that similar outcomes were found in two very different topics of conversation. Based on the data, sharing negative information has significantly more influence on changing one’s attitude about the target object. These data were consistent with the negativity effect as Kellermann (1984) asserted as “the tendency for negative information to be weighed more heavily than positive information when forming evaluations of social stimuli”(p. 37). Therefore, it is plausible to speculate that valence of information can be one of the possible factors that are able to shape one’s attitude about the target object in a certain direction. These data were also consistent with Richey et al.’s finding that negative information has significantly more influence on forming an impression about the target object than positive information, since significant mean differences among three 29 valence of judgment conditions were evident. As figure 1 shows, a significant negative linear trend is present between valence of j udgment and attitude change. In other words, sharing negative information results in the greatest attitude change in a negative direction, as the negativity effect has proven. Since the data support the first hypothesis that sharing negative information results in significantly greater attitude change in a negative direction, it also supports Anderson’s findings (1965) that negative adjectives have a stronger influence on forming an impression than positive adjectives about the target object. According to statistical analyses and Anderson’s research, the reason why the negative information has more influence on the interactants is that the negative adjectives have a stronger modifying capacity than do positive adjectives in terms of evaluation. He reasoned that if the interactants share the same number of positive and negative information items about the target object, then the observers’ impression should be neutral, but the statistical analyses did not support the idea of being neutral when forming an impression. It should be noted that instead the analyses indicated that when sharing five equal amounts of positive and negative information, the attitude change about the target object is skewed to the negative direction as the negative mean value suggested as, M = -.l6, S_D '2 .63 (See Table 4, See also Figure 1). Also, table 5 summarizes the percentage of attitude change in the different groups in terms of positive, neutral and negative changes. Sharing negative information changed attitudes toward the target object in a negative direction for 63.2% of participants, while sharing positive information changed the attitude in positive direction in 50% of participants. It should be noted that sharing combined information changed the attitude toward the target object in a negative direction for 50% of these 30 participants, with 35% changing in a positive direction and 15% remaining consistent in their evaluation of the target object. These results suggest that although the relative amount of positive and negative information is controlled and balanced, the data showed that negative impressions tend to emerge toward the target object. This tendency was also expected by ample evidence including Ito et al.’s findings (1998) that, although the positive information and negative information were equally probable, negative information is evaluatively more extreme and more arousing. Caccioppo et al. (1997) also described the negativity bias as one’s tendency for activation to result in a greater change in output in the negative motivational system as compared with the positive motivational system. As a consequence of the negativity effect, attitudes and behaviors should be more strongly influenced by a negative stimulus than a positive stimulus. The second hypothesis posited that sharing negative information would increase the perceived likability of each interactant in a dyadic setting. The reasoning presented earlier was that the characteristics of informativeness of negative information enables interactants to increase perceived similarity while beginning to agree on the negative information. Although this study did not find significant mean differences for hypothesis two, the data indicated a positive linear relationship in the expected direction. The same expected mean direction was also shown in hypotheses three, four, and five. They showed that the negative group scored the highest in perceived credibility , perceives similarity, and uncertainty reduction, while the positive group had the lowest scores on these variables although lacking significant mean difference between the groups. This constant positive linear relationship suggests that a possibility exists that sharing negative information might function differently than sharing either positive or combined 31 information. Several points can be made in a summary of the main findings (a) The conversational topics are not significantly important to determine one’s attitude change, but the valence of judgment is a strong indicator to determine one’s attitude change in a certain direction. (b) Sharing negative information can significantly change one’s attitude about the target object or person in a negative direction. (c) When the same amount of positive and negative information is present, participants’ attitudes about the target object tend to be directed more toward negative information, since negative information weighs more heavily than neutral or positive information. ((1) Constant positive linear relationships are found for the negativity effect in impression formation, including perceived likability, perceived credibility, perceived similairty, and uncertainty reduction showing that negative group has the highest score on every scale, and the positive group has the lowest score on every scale. Limitations Although this study has found significant main effects in attitude change and the negativity effect, there are still areas where future research can improve. First of all, one of the reasons for the lack of significance in mean difference can be accounted for by the counterarguments of Byme’s reinforcement theory. While a number of Byme’s findings suggest one of the most established and well-known facts in interpersonal communication is that attitude similarity enhances the interpersonal attraction, an increasing amount of counterarguments suggest that this belief is unfounded (Sunnafrannk, 1992). Others can appear more interpersonally attractive by making attitudinally dissimilar statements during interaction processes (Sunnafrank, 1986; Sunnafrank, 1985; Sunnafrank, 1984). 32 The repulsion hypothesis proposed by Rosenbaum (1986) also suggests that attitude similarity does not heighten interpersonal attraction, since similarity is expected. In other words, the repulsion hypothesis suggests that the important relationship is not between attitude similarity and interpersonal attraction but between attitude dissimilarity and interpersonal repulsion (Rosenbaum, 1986). Future research must consider the problem of a control issue. For instance, the experiment session time varied group to group, and time is an important factor in impression management. According to Berger’s Uncertainty Reduction Theory, the more time strangers spend together, the more uncertainty reduction occurs (1979). Therefore, much variability occurred from group to group. Although, the conversational topic was given to each experimental group, no apparent control was maintained over every session, since the experimenter was not able to be present in every session. In other words, participants could possibly engage in a conversation that was not related to the actual experiment. The difference in experimental setting may affect the experiment. Although this difference may not have been an important factor of influencing the perception of the participants in any significant way, it is crucial to consider all possible variations in experiment. Therefore, for future research, efforts should also be made to obtain more control on the experiment of these addressed issues. In addition, future research should investigate a number of areas where the generality of the present research is limited: (a) Future research should measure the extremeness of each statement made in the experiment session, since the extremeness is one of the factors that determine the informativeness of the information. Future research should focus on how negative information is more extreme than positive or combined 33 information. (b) Future research should specifically look at where the behavior of agreeing on each interactant’s statement occurs, and discover whether the agreement on the negative information brings out more perceived similarity of each participant than positive or combined information. (o) Future research should consider the relationship between the negativity effect and complaining and giving negative criticism. Since these two behaviors both deal with the negativity effect in terms of affect, emotion, and cognition, it is important to find out whether the negativity effect is strongly related to these behaviors. In other words, future research should determine whether sharing negative information functions as complaining behavior or giving negative criticism about the target object is a good idea. ((1) The negativity effect also can have some implications in gossiping behavior. Most gossip consists of sharing negative information about the target person rather than sharing positive information about the person. How can the negativity effect be applied to gossiping behavior? (6) Is valence of judgment (sharing positive information, sharing negative information, or sharing combined information) the only factor that functions to form impressions of each other? In other words, it might be personality traits and physical appearance that determine one’s first impression rather than the valence of judgment. Therefore, it is essential to determine how these factors can be controlled in experiments. Conclusion In this research, the negativity effect was tested in dyadic interpersonal communication. The data showed that the valence of judgment is a strong indicator of changing one’s attitude about the target object. Negative information, in particular, has the greatest capacity to change one’s attitude in a negative direction. 34 Although the data did not provide significant evidence in the relationship between negativity effect and impression formation, including perceived likability, perceived credibility, perceived similarity, and uncertainty reduction, the constant positive linear relationship across the impression formation scales suggests that sharing negative information involves a somewhat unique process, differentiated from sharing positive or combined information. 35 Bibliography 36 Bibliography Abele, A. & Petzold, P. (1994). How does mood operate in an impression formation task? An information integration approach. European Journal of Social Psychology. 24, 173-187. Allen, M. & Burrell, N. (2000). The Negativity Effect in Political Advertising: A Meta-Analysis. Unpublished manuscript, University of Wisconsin-Milwaukee. Alicke, M., Braun, J ., Glor, J., Klotz, M., Magee, J., Sederholrn, H. & Siegel, R. (1992). Complaining Behavior in Social Interaction. Personality and Social Psychology Bulletin 18, 286-295. Amabile, T. (1981). Brilliant but Cruel: Perceptions of negative evaluators. Journal of Experimental 80%] Psychology, 19, 146-156. Amabile, T. & Glazebrook, A. (1981). A Negativity Bias in Interpersonal Evaluation. Journal of EitpLerimental Social Psychology/L18, 1-22. Anderson, N. (1965). Averaging vs. adding as a stimulus-combination rule in impression formation. Journal of Experimental Psychology. 70, 394-400. Aronson, E., & Worchel, S. (1966). Similarity versus liking as determinants of interpersonal attractiveness. Pchhonomic ScienceL5, 157-158. Au Yun, J. K. (1995). The effect of attitude similarity on interpersorgl att_raction: A meta-analysis. Unpublished doctoral dissertation, Michigan State University, East Lansing. Baumeister, R. F. & Leary, M. P. (1995). The Need to Belong: Desire for Interpersonal Attachments as a Fundamental Human Motivation. BMW M, 497-529. 37 Berger, C. R. (1979). Beyond initial interaction: Uncertainty, understanding, and the development of interpersonal relationships. In H. Giles and R. St. Clair (Eds), Language and Social Psycholog. Oxford: Basil Blackwell. Berscheid, E. (1966). Interpersonal attraction. In G. Lindzey & E. Aronson (Eds), Handbook of Social Psychology, 4, 670-680. Brehm, S. S. (1992). Intimate Relationships (2nd ed.). New York: McGraw-Hill. Burleson, B. Samter, W. & Lucchetti, A. E. (1992). Similarity in communication values as a predictor of fi'iendship choices: Studies of friends and best friends. Southern Communication Journalé7, 260-276. Byrne, D. (1961). The influence of propinquity and opportunities for interaction on classroom relationships. Human Relations, 14, 63-70. Byme, D. (1969). Attitudes and attraction. In L. Berkowitz (Ed.), Advanced in Experimental Social psychology, 4. New York: Academic Press. Byme, D. (1971). The Attraction Paradigm. New York and London: Academic Press. Cacioppo, J. T. & Bemtson, G. G. (1994). Relationship between attitudes and evaluative space: A critical review, with emphasis on the separability of positive and negative substrates. Psychological Bulletin, 115, 401-423. Cacioppo, J. T., Gardner, W. L., & Bemtson, G. C. (1997). Beyond bipolar conceptualizations and measures: The case of attitudes and evaluative space. Personality and Social Psychology Review, 1, 3-25. Chaiken, S., Liberman, A. & Eagly, A. H. (1989). Heuristic and systematic information processing within and beyond the persuasion context. Unintended Thought. 38 New York: The Guilford Press. Cusumano, D. & Richey, M. (1970). Negative salience in impressions of character: Effects of extremeness of salient information. Psychonomic Science, 20, 81- 83. Erber, R. & Erber, M. (2001). Intimate Relationships. Needham: A Pearson Education Company. Eagly, A. H. & Chaiken, S. (1993). Process theories of attitude formation and change: The Elaboration Likelihood and Heuristic-Systematic Models. The Psychology of Attitudes. F ortworth: Jovanovich College Publishes. F iske, S. T. (1980). Attention and weight in person perception: The impact of negative and extreme behavior. Journal of Personality and Social Psychology, 38, 889- 908. Goffrnan, E. (1959). The Presentaflm of Self in Everdeay Life. Garden City: Doubleday. Hamilton, D. L., & Zanna, M. P. (1972). Differential weighting of favorable and unfavorable attributes in impression formation. Journal of Experimental Research in Personality. 6. 204-212. Hunter, J. E., & Gerbing, D. W. (1982). Unidimensional measurement, second order factor analysis, and causal models. Research in Organiaaaional Beflvior, 4, 267- 320. Huston, T. L. (1974). Foundations of Interpersonal Attraction. New York: Academic Press. Ito, T. A., Larson, J. T., Smith, N. K. & Caccioppo, J. T. (1998). Negative 39 information weighs more heavily on the brain: The negativity bias in evaluative categorizations. Jourrgl of Personality and Social Psycholagy, 75, 887—900. Johnson-Cartee, K. S. & Copeland, G. A. (1991). Negative Political Advertising: Coming of Agg. Hillsdale: Lawrence Erlbaum Association. Kanouse, D., & Hanson, L. (1971). Negativity in evaluations. In E. Jones et al. (Ed.), Attribution: Perceiving the cause of behavior. Morristown, NJ: General Learning Press. Kellermann, K. (1984). The Negativity Effect and its Implications for Initial Interaction. Communication Monographs, 51, 37-55. Kellermann, K. (1989). The Negativity Effects in Interaction: It’s all in Your Point of View. Human Communication Research 16, 147-183. Kline, J. G. (1991). Negativity Effects in Impression Formation: A Test in the Political Arena. Personality and Social Psychology Bulletin, 17, 412-418. Kline, J. G. (1996). Negativity in impressions of presidential candidates revisited: The 1992 election. Personam and Social Psychology Bulletin, 22, 288-295. McCroskey, J. C. (1966). Scales for the measurement of ethos. _S_pe_egh Monoggapha, 33, 65-72. Malle, B. F. & Knobe, J. (1997). Which Behaviors Do People Explain? A Basic Actor-Observer Asymmetry. Journal of Personality and Social Psychology, 72, 288-304. Matlin, M., & Stang, D.(1978). The Pollyanna Principle. Cambridge: Schenkrnan. O’Keefe D. J. (1990). Persuasion: Theory and Research. Thousand Oaks: Sage. Parks M. R. & Adelman, M. B. (1983). Communication Networks and the 40 Development of Romantic Relationships: An Expansion of Uncertainty Reduction Theory. Human Communication Research, 10, 55-79. Petty, R. E., Cacioppo, J. T., Sedikides, C., & Strathman, A. J. (1988). Affect and Persuasion. American Behavioral Scienctist, 31, 355-371. Richey, M. H., Bono, F. S., Lewis, H. V., & Fortin, R. (1975). Negative Salience in Impressions of character: Effects of Unequal Proportion of Positive and Negative Information. Journal of Social Pchhology, 97, 233-241. Rosenbaum, M. E. (1986). The repulsion hypothesis: On the nondevelopment of relationships. Journal of Personality and Social Psychology, 51, 57-65. Skowronski, J. J. & Carlston, D. E (1989). Negativity and extremity biases in impression formation: A review of explanations. Psychological Bulletin, 105, 131-142. Sunnafrank, M. (1984). A communication-based perspective on attitude similarity and interpersonal attraction in early acquaintance. Communication Monographs. 51. 372-380. Sunnafrank, M. (1985). Attitude similarity and interpersonal attraction during early communicative relationship: A research note on the generalizability of findings to opposite-sex relationships. Western Journal of Speech Communication, 49, 73-80. Sunnafrank, M. (1986). Communicative influences on perceived similarity and attraction: An expansion of the interpersonal goals perspective. Western Journal of Speech Communication, 50, 158-170. Sunnafrank, M. (1992). On debunking the attitude similarity myth. Communication Monographs, 59, 164-179. Tardy, C. H. (Ed.). (1988). For the Study of Human Communication: Methods 41 and Instruments for Observing, Meaning, gd Assessing Communication Process. Norwood: Ablex Publishing Corporation. Taylor, S. E. (1991). Asymmetrical effects of positive and negative events: The mobilization-minimization hypothesis. Psychological Bulletin, 110, 67-85. Wilson, P. O. (1987). Presidential advertising in 1988. Election Politics, 4, 17- 21. Wojciszke, B., Brycz, H. & Borkenau, P. (1993). Effects of Information Content and Evaluative Extremity on Positivity and Negativity Biases. Journal of Personality and Social Pchhology, 64, 327-335. 42 TABLES AND FIGURE 43 Table 1. Means, standard deviations, and t-test on different conversational topics Scale M _SL) af 1 P Attitude Change 112 -.567 .572 Non-distant -.1 1 .65 Distant -.18 .68 Perceived Likability 111 -.421 .675 Non-distant 5.74 .82 Distant 5.67 .92 Perceived Credibility 112 -.122 .903 Non-distant 5.46 .59 Distant 5.44 .64 Perceived Similarity .707 Non-distant 5.08 .94 112 -.377 Distant 5.00 l .08 Uncertainty .883 Reduction 3.61 1.27 112 .148 Non-distant 3.65 1 .55 Distant 44 Table 2 . Means, standard deviations, and coefficient Alphas for the scales Scale N g M S_D Attitude Change 114 .71 -.15 .62 Perceived Likability 113 .91 5.71 .87 Perceived Credibility 114 .79 5.45 .61 Perceived Similarity 114 .76 5.20 .90 Uncertainty Reduction 1 14 .74 4.07 1.15 45 Table 3. Means, standard deviations, and coefficient Alphas for the scales with items deleted Scale N g M SQ Attitude Change 114 .72 -.14 .67 Perceived Likability 113 .91 5.71 .87 Perceived Credibility 114 .79 5.45 .61 Perceived Similarity 114 .83 5.04 1.01 Uncertaintj Reduction 114 .84 3.63 1.41 46 Table 4. Means of Attitude Change, Perceived Likablity, Perceived Credibility, Perceived Similarity, and Uncertainty Reduction by different Valence of Judgment Dependent Variables Group a Change Likability Credibility Similarity Uncertainty Positivity 36 M 0.05 5.56 5.32 4.92 3.49 S_D 0.77 0.94 0.57 0.76 1.34 Combined 40 M -0.16 5.73 5.50 5.06 3.50 SD 0.63 0.83 0.65 1.01 1.53 Negativity 38 M -0.34 5.81 5.52 5.14 3.90 SD 0.54 0.85 0.61 1.21 1.35 47 Table 5. Percentages of Attitude Change on Message Valence Valence of Information Negative Combined Positive Total Positive 23.7% 35.0% 50.0% 36.0% Attitude Change Neutral 13.2% 15.0% 5.6% 11.4% Negative 63.2% 50.0% 44.4% 52.6% Total 100% 100% l 00% 100% 48 Attitude Change -.4 positive combined negative Valence of Judgment Figure 1. Trend Analysis of Valence of Judgment on Attitude Change in Hypothesis 1. 49 APPENDICES 50 Appendix A Assessing General Attitude Toward MSU 1. Igenerally like MSU. notatall l 2 3 4 5 6 7 verymuch 2. If I had a chance to reapply to college, I would apply to MSU. not at all 1 2 3 4 5 6 7 verymuch 3. I like my friends and professors here at MSU. not at all 1 2 3 4 5 6 7 verymuch 4. I am NOT satisfied with MSU’s reputation.* not atall 1 2 3 4 5 6 7 verymuch 5. I am happy that I chose to attend to MSU. notatall l 2 3 4 5 6 7 verymuch 6. I DO NOT like the general atmosphere at MSU. * notatall 1 2 3 4 5 6 7 verymuch 7. I consider MSU as the best university for me. not atall 1 2 3 4 5 6 7 verymuch 8. I will NOT encourage people to apply to MSU. * notatall l 2 3 4 5 6 7 verymuch 9. I am proud of being an MSU student. not at all 1 2 3 4 5 6 7 verymuch 10. I have good feelings toward MSU. notatall 1 2 3 4 5 6 7 verymuch *reflected items 51 Appendix B Assessing General Attitudes Toward Cell Phones 1. I think using cellular phone is good. disagree 1 2 3 4 5 6 7 agree 2. If I have a chance to use cellular phone (or already have one), I will use one (or keep using one). . disagree 1 2 3 4 5 6 7 agree 3. It is convenient to have cellular phone when there is no coin or no public phone. disagree 1 2 3 4 5 6 7 agree 4. Cellular phones are expensive for college students.* disagree 1 2 3 4 5 6 7 agree 5. It is easy to assess much useful information when you have a cellular phone (e. g. e- mail, stock, news etc.) disagree 1 2 3 4 5 6 7 agree 6. Having a cellular phone can invade your privacy. * disagree 1 2 3 4 5 6 7 agree 7. You don't have to worry about missing a phone call when you are away from your home or your office when you have a cellular phone. disagree 1 2 3 4 5 6 7 agree 8. I will NOT encourage people to buy a cellular phone. * notatall l 2 3 4 5 6 7 verymuch 9. There is a potential danger of radiation when using a cellular phone.* not at all 1 2 3 4 5 6 7 verymuch 10. I have good feelings toward using a cellular phone. notatall 1 2 3 4 5 6 7 verymuch *reflected items 52 Appendix C Assessing Perceived Likability of the Partner 1. I think my partner was not well adjusted in our session.* notat all 1 2 3 4 5 6 7 verymuch 2. I would highly recommend my partner for a responsible job. not at all 1 2 3 4 5 6 7 verymuch 3. In my opinion, my partner is an exceptionally mature person. notatall 1 2 3 4 5 6 7 verymuch 4. I have great confidence in my partner’s good judgment. notatall l 2 3 4 5 6 7 verymuch 5. Most people would react very favorably to my partner after a brief acquaintance. not at all 1 2 3 4 5 6 7 verymuch 6. I think that my partner and I were quite similar to each other in the session. notat all 1 2 3 4 5 6 7 verymuch 7. I think that my partner is one of those people who quickly wins respect. notat all 1 2 3 4 5 6 7 verymuch 8. My partner is one of the most likable people I know. not at all 1 2 3 4 5 6 7 verymuch 9. My partner is the sort of person who I myself would like to avoid. * notatall 1 2 3 4 5 6 7 verymuch 10. It seems to me that it would be very hard for my partner to gain admiration from others. * not at all 1 2 3 4 5 6 7 verymuch *reflected items 53 Appendix D Assessing Perceived Credibility of the Partner 1. 8. My partner is a reliable source of information on the given topic. not at all 1 2 3 4 5 6 7 verymuch . I believe that my partner is quite intelligent. notatall 1 2 3 4 5 6 7 verymuch . My partner is basically honest. notatall 1 2 3 4 5 6 7 verymuch My partner has good character. notatall 1 2 3 4 5 6 7 verymuch . I would like to have my partner as a personal friend. notatall 1 2 3 4 5 6 7 verymuch I did not trust my partner to tell the truth while in our session. * not at all 1 2 3 4 5 6 7 verymuch . My partner seems to be an honorable person. notatall 1 2 3 4 5 6 7 verymuch I did not believe that my partner would be concerned with my well-being. * not at all 1 2 3 4 5 6 7 verymuch 9. Under most circumstances I would be likely to believe what my partner says about this topic. notatall 1 2 3 4 5 6 7 verymuch 10. I would prefer to have nothing at all to do with my partner. * notatall 1 2 3 4 5 6 7 verymuch *reflected items 54 Appendix E Assessing Perceived Similarity and Uncertainty Reduction 1. 9. My partner and I seem to like a lot of the same things. not at all 1 2 3 4 5 6 7 verymuch . We shared a lot of the same attitudes about the given topic in our session. notatall 1 2 3 4 5 6 7 verymuch . We had very different values toward the given in our session. * notatall 1 2 3 4 5 6 7 verymuch . I believe we are very similar. notatall 1 2 3 4 5 6 7 verymuch . We seem to have a similar outlook on life. notatall 1 2 3 4 5 6 7 verymuch . I feel like I know my partner pretty well after the session. notatall 1 2 3 4 5 6 7 verymuch . My partner often did or said things which surprised me in our session. *. notatall 1 2 3 4 5 6 7 verymuch In the future, if I keep having a relationship with my partner, I will be confident of my ability to accurately predict my partner’s behavior. notatall 1 2 3 4 5 6 7 verymuch I felt I could predict my partner’s behavior in our session. not at all 1 2 3 4 5 6 7 verymuch 10. I felt I could tell what my partner was feeling inside in our session. notatall 1 2 3 4 5 6 7 verymuch * reflected items 55 Appendix F Factor After Loading items deleted Attitude Change 1 (M = -.15, S_D = .62, a = .71) 1. I generally like MSU. .42 .37 2. If I had a chance to reapply to college, I would apply to MSU. .63 .61 3. I like my friends and professors here at MSU. .19* deleted 4. I am NOT satisfied with MSU’s reputation. .29 .27 5. I am happy that I chose to attend to MSU. .47 .53 6. I DO NOT like the general atmosphere at MSU. .36 .34 7. I consider MSU as the best university for me. .52 .52 8. I will NOT encourage people to apply to MSU. .35 .36 9. lam proud of being an MSU student. .60 .61 10. I have good feelings toward MSU. .62 .64 Attitude Change (M = -.15, S_D = .62, a = .71) 1. I think using cellular phone is good. .42 .37 2. If I have a chance to use cellular phone (or already have one), .63 .61 I will use one (or keep using one). 3. It is convenient to have cellular phone when there is no coin .19* deleted or no public phone. 4. Cellular phones are expensive for college students. .29 .27 5. It is easy to assess much useful information when you have a .47 .53 cellular phone (e. g. e-mail, stock, news etc.) 6. Having a cellular phone can invade your privacy. .36 .34 7. You don't have to worry about missing a phone call when you .52 .52 are away from your home or your office when you have a cellular phone. 8. I will NOT encourage people to buy a cellular phone. .35 .36 9. There is a potential danger of radiation when using a cellular .60 .61 phone. 10. I have good feelings toward using a cellular phone. .62 .64 Perceived Likability (M = 5.71, S1; = .87, g = .91) 1. I think my partner was not well adjusted in our session. .35 .35 2. I would highly recommend my partner for a responsible job. .77 .77 3. In my opinion, my partner is an exceptionally mature person. .80 .80 4. I have great confidehce in my partner’s good judgment. .86 .86 5. Most people would react very favorably to my partner after a .81 .81 brief acquaintance. 6. I think that my partner and I were quite similar to each other .68 .68 in the session. 7. I think that my partner is one of those people who quickly .86 .86 wins respect. 56 8. My partner is one of the most likable people I know. .75 .75 9. My partner is the sort of person who I myself would like to .58 .58 avoid. 10. It seems to me that it would be very hard for my partner to .56 .56 gain admiration from others. Perceived Credibility (M = 5.45, S_D = .61, g = .79) 1. My partner is a reliable source of information on the given .54 .54 topic. 2. I believe that my partner is quite intelligent. .70 .70 3. My partner is basically honest. .83 .83 4. My partner has good character. .86 .86 5. I would like to have my partner as a personal friend. .65 .64 6. I did not trust my partner to tell the truth while in our session. .25 .25 7. My partner seems to be an honorable person. .85 .85 8. I did not believe that my partner would be concerned with my well—being. -.48 -.48 9. Under most circumstances I would be likely to believe what .62 .61 my partner says about this topic. 10. I would prefer to have nothing at all to do with my partner. .55 .55 Perceived Similarity (M = 4.64, SD = .91, a = .93) 1. My partner and I seem to like a lot of the same things. .76 .75 2. We shared a lot of the same attitudes about the given topic in .47 .40 our session. 3. We had very different values toward the given topic in our .14* deleted session. 4. I believe we are very similar. .82 .82 5. We seem to have a similar outlook on life. .84 .83 Uncertainty Reduction (M = 3.63, S_D = 1.41, g = .84) 1. I feel like I know my partner pretty well after the session. .73 .76 2. My partner often did or said things which surprised me in our .14* deleted session. 3. In the future, if I keep having a relationship with my partner, I .65 .70 will be confident of my ability to accurately predict my partner’s behavior. 4. I felt I could predict my partner’s behavior in our session. .63 .66 5. I felt I could tell what my partner was feeling inside in our .51 .54 session. * Item was deleted due to internal consistency and parallelism. 57