In J milllllllllllllllllllfilllllll ' g,.. 3 1293 00829 0896 news :fiflvfijiflfif Myriam t: ‘ n ';_£i§p§iell This is to certify that the thesis entitled THE EFFECTS OF SELF AND SOCIAL MONITORING ON DECEPTION AND DETECTION presented by Carra Sleight has been accepted towards fulfillment of the requirements for M.A. degreeianmmnnJ'mtion 4%ng Major professor DateMemhazJ 5 ’ 0-7639 MS U is an Aflinnaa'vc Action/Equal Opportunity Institution MSU RETURNING MATERIALS: Place in book drop to LJBRARJES remove this checkout from -_ your record. FINES will be charged if book is returned after the date stamped below. ‘ ' ”6190'? m". , -- v ' I (j "i: THE EFFECTS OF SELF AND SOCIAL MONITORING ABILITIES ON DECEPTION AND DETECTION By Carra Sleight A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Communication 1985 ABSTRACT THE EFFECTS OF SELF AND SOCIAL MONITORING ON DECEPTION AND DETECTION by Carra Sleight Like communication, deception is a negotiated process, but past deception research has concentrated on the behaviorial correlates of deception or accuracy of detection, an action-centered approach. Few studies have integrated the two activities to determine if the cues that are noticed are the cues that lead to accurate detection, an interactive approach. Because deception detection is a process of formulating perceptions, it is suggested that one's habitual perceptual style might help to determine why certain cues are noticed, why others are disregarded, and how accurately the cues are processed. Self and social monitoring are considered such perceptual styles. The former should be most closely related to deception, the latter to detection. A simplified card game was used to test these relationships. Results showed that the manipulation, while interactive and naturalistic in its approach, was not strong enough to produce significant results. ACKNOWLEDGEMENTS I would like to thank everybody I have ever known, with the possible exception of my ex—husband, for their contribution to my work. Actually, my ex probably deserves special thanks, for he made the largest single contribution to my interest in deception. Particular thanks for their help in actually producing this paper must go to: G. R. Miller whose work inspired (and inspires) me and whose guidance I could not do without; Jack Bain who is always on my side; Vince Farace whose elegant criticisms helped me to think more carefully; and Paul Mongeau who introduced me to Lennox & Wolfe's work. Thanks also to Barry Caplan and Joe Speaks for their volunteer work (in Barry's case over 20 hours) as card dealers. Without their cooperation, no experiment would have taken place. Finally, special mention must be made of the late L. B. Jones, my grandmother, whose belief in the benefits of education makes my career as a graduate student possible: Thanks, Gran, I know you know. ii Chapter 1, Introduction Chapter 2, Method . . Chapter 3, Results . Chapter 4, Discussion References . . . . . Appendix . . . . . . TABLE OF CONTENTS iii 18 22 25 LIST OF TABLES Self-Monitoring and Social Monitoring Means with Original Scales . . Self—Monitoring and Social Monitoring Cell Means with Original Scales ANOVA of Dependent Variable Deception using Original Self and Social—monitoring subscales . . . . . . . . . . . . . . . . . . ANOVA of Dependent Variable Deception using Original Self and Social-monitoring subscales . . . . . . . . . . . . . . . . Test-retest reliabilities were .65 (.88) . . . . . . . . . . . . . . Self Monitoring and Social Monitoring Cell Means with modified scales Test-retest correlation matrix from PACKAGE . . . . . . . . . . . . Self-monitoring and Social—monitoring Subscale Intercorrelations . iv . IO 11 12 . 12 14 15 16 . 17 CHAPTER 1 INTRODUCTION Persistent use of a static, action-centered model (Stiff & Miller, 1984a, b) in deception studies has resulted in research focused on either deception or detection with emphasis on the attendant behavioral cues generated by either activity. Concrete generalizable results regarding these external cues have been scant (see Kraut, 1980 and Zuckerman, DePaulo & Rosenthal, 1981 for reviews of the literature) and contradictory. While findings have been equivocal, some cues have been more commonly noted across studies than others. Nonverbal cues in this category include reduced postural shifts (O'Hair, Cody & McLaughlin, 1981), reduced eye contact, (Exline, Thibaut, Hickey & Gumpers, 1970), and decreased use of illustrators (Ekman, Friesen & Scherer, 1976). Catalogued verbal cues include shortened responses due to nervousness and uncertainty (Knapp, Hart & Dennis, 1974), increased disfluencies (DePault, Rosenthal, Rosenkrantz & Green, 1981) and discrepant message content (Mehrabian & Winer, 1967; Miller, Mongeau & Sleight, 1984). When asked, subjects report a similar list of expected cues (Miller et al., 1984), although a more recent compilation (Sleight, 1985) indicates a much more discrepant list of expectations. However, this group of common cues does not guarantee accurate deception detection. There seems to exist, then, a gap between what we study, how we study it, and what is actually happening. Given the fact that detection accuracy scores rarely exceed 65 percent, (Kraut, 1980) one might conclude that the wrong question is being addressed or that some important variable is being overlooked. 2 With regard to the first possibility, little attention has been given to investigating the correspondence between deception and detection, as Stiff and Miller (1984b) point out in their proposal for an interactive approach. This would permit investigators to identify behaviors that result in perceptions of deceptiveness, whereas the action-centered or source—oriented approached cannot. Instead of asking, "What are the cues of deception?" or "How accurate are subjects at decoding cues?"; the question of interest becomes "On what cues do individuals rely when formulating perceptions of deceptiveness and how are these cues related to actual deceptive behavior?" This perspective suggests an additional line of inquiry, "Why do individuals rely on the cues that they do?" which may point to the second possibility mentioned above. There may be an individual difference variable (or a set of them) that determines whether or not cues are comprehended and how accurately they are put to use. Attention should be drawn to the idea that detection involves formulating perceptions. Since no one cue is an idea of deception and thus a guarantor of detection, then the judgment must be made based on some internal criterion. The locus of the process is the individual detector; the process itself is cognitive, and‘the index of success may be social ability of some kind. In fact, Maier and Janzen (1967) found that subjects made judgments about veridical behavior and then defended them regardless of the evidence or their correctness. Like communication, deception is a negotiated process. It takes two people, one of whom intends to deceive, the other with the proclivity to detect. Effective deception can be seen as either a willingness to be deceived (for whatever reason) or an inability to correctly perceive and decode possible cues that would lead to detection. Given that the 3 former may occur to simplify the problems of daily life ("Poached eggplant! My favorite!") as a matter of course, the latter may vary due to personal abilities and interest in the behaviors of others. Heider (1958) notes, The fact that there is a lack of correspondence between the raw material of perception and the intended object of perception allows idiosyncratic approaches to the world on the part of the observer a much freer reign in the organization and interpretation of incoming proximal stimuli. The issue here...concern(s)...perceptual styles --what the person extracts from his world because of his manner of perceiving. (PP. 56-7) Self monitoring can be considered a perceptual style that might affect the deception/detection process, for it involves the ability of self to apprehend and manage external cues. It is of particular interest for two reasons: 1) The original conception of self monitoring (Snyder, 1974) correlates highly with empathy (Stiff, 1984), which suggests predictive accuracy. Mead's (1934) early work defined empathy as the ability to take the role of another. Such ability was conceptualized as predictive accuracy. High self monitors, in taking the role of another, should thus be better able to predict other's behavior. 2) Recent work of Lennox and Wolfe (1984) has refined Snyder's construct and improved the operationalization. The original conception stated: The goals of self monitoring may be (a) to communicate accurately one's true emotional state by means of an intensified expressive presentation; (b) to communicate accurately an arbitrary emotional state which need not be congruent with actual emotional experience; (c) to conceal adaptively an inappropriate emotional state and appear unresponsive and unexpressive; (d) to conceal adaptively an inappropriate emotional state and appear to be experiencing an appropriate one; (e) to appear to be experiencing some emotion when one experiences nothing and a nonresponse is inappropriate. (Snyder, 1974, p. 527) 4 These stated goals seem applicable to both successful deception and successful detection. Prior work can be cited to both support and contradict this supposition. Self monitoring was found not to have a significant effect on deception (Comadena, 1982; Kraut & Poe, 1980; Ekman et al., 1976), a counterintuitive result given the above criteria. High self monitors were found to use different detection strategies than low monitors (Elliott, 1979), a finding that agrees with the Stiff (1984) correlation. These results may be done to the original measure itself; ...an instrument specifically designed to discriminate individual differences in concern for social appropriate— ness, sensitivity to the expression and self-presentation of others in social situations as cues to social appropri— ateness of self-expression, and use of these cues as guide- lines for monitoring and managing self-presentation and expressive behavior. (Snyder, 1974, p. 529) Lennox and Wolfe determined that "The scale's multidimensionality (Snyder specified five factors) extends beyond the limits of the construct, creating a situation in which its factors compete with one another" (1984, p. 1350). The scale was found to dependably yield only three factors; two of which, acting ability and extraversion, were irrelevant to it and better measured by other instruments. The remaining factor of self monitoring, other—directedness, loaded on two distinct dimension, which they felt correctly represented "Snyder's (1979) description of the high self monitor as one who ‘is particularly sensitive to the expression and self-presentation of relevant others' (p. 89) and who uses these cues as a guide to regulating self— presentation" (p. 1359). The two dimensions were defined as "the ability to modify self presentation" and "sensitivity to the expressive behavior of others" (p. 1359). They were found to be correlated ( .22, p (.01, two—tailed). For the purposes of symmetry the former dimension will be called self 5 monitoring, the latter social monitoring. These two more sharply defined variables should affect both the deception and detection process more clearly. Self monitoring should be most closely related to successful deception and social monitoring to successful detection. This restores their intuitive relation to the process. Therefore, regardless of his/her social—monitoring ability, the high self monitor who is able to adapt his or her behavior cross— situationally should be a better liar. On the other hand, the high social monitor who attends more assiduously to the cues of others should be a better detector, regardless of his/her self—monitoring score. Consequently, those individuals who are both high social and self monitors should excel at both deception and detection, while those low in both skills should not. Stated formally, the hypotheses are: H1: Individuals who are both high social and high self monitors should be relatively more successful at both deception and detection. H2: Individuals who are high social monitors and low self monitors should be relatively successful at detection but relatively unsuccessful at deception. H3: Individuals who are low social monitors and high self monitors would be relatively unsuccessful at detection but relatively successful at deception. H4: Individuals who are both low social and low self monitors should be relatively unsuccessful at both deception and detection. This study then, attempts to take an interactive look at the deception/detection process by looking at why certain subjects may notice the cues that they do, and what they then do with those cues. CHAPTER TWO METHOD To establish scores for test—retest reliability, students at a large Midwestern University enrolled in a basic communication course were given the Lennox and Wolfe Revised Self—Monitoring Scale (1984). This scale is a 13—item instrument. Seven items pertain to self monitoring, while six items measure social monitoring. A five item, Likert—type scale, ranging from strongly disagree (1) to strongly agree (5), was used to tally the responses. The ordering of the items was the same random ordering as that used by Lennox and Wolfe. At this first scale administration students were also asked to enroll for the experiment itself. Those who did participate were given extra credit for their work. One hundred students actually participated; 66 women, 34 men, all of whom were to both deceive and detect. Ten days later, subjects were called into the lab in dyads. The lab setup consisted of two rooms. In one, measurement scales were administered, in the other, the experimental manipulation took place. Once in the lab, subjects were either ushered directly into the manipulation room or asked to wait in the adjacent room and fill out the measurement scales. This alternating schedule expedited the work and helped to keep all subjects busy most of the time. Respondents were asked to fill out a reordered Lennox and Wolfe Scale. Empathy, dogmatism, and Machiavellianism measures were also administered. This was done for two reasons: 1) to both relax subjects and focus their attention on the experimental setting if they had to wait, 2) to provide additional information for future work. Manipulation The manipulation consisted of a card game similar to both blackjack and poker. The game was directed by an assistant trained to be the dealer. Dyads were seated opposite each other at a table. A written set of rules was provided for each player. After reading the rules, players were asked if they had any questions. If they did, the dealer was instructed to reread the rules with the players until they understood and to caution them that two practice hands would be played. The practice hands served as additional explanation. The rules stated that each player would be dealt a pair of cards. Player A's cards would be dealt face down, Player B's cards face up. Player A, who could look at his/her own cards and also see B's hand, was the deceiver in the game. Player B, who could see only his/her own cards, was the detector. An ante of 25 cents or one dollar was preset for each hand. The dealer informed the players of the ante. Both then anted the required amount. After the deal of the cards a waiting period of 30 seconds was observed. During this time the players could talk to one another or remain silent. Player B could then either call or fold his/her hand. If B folded, and A's cards were high, then B would win the hand. If B called, and A's cards were low, then B would also win the hand. The opposite was also true. If B folded when A had low cards, A would win; or if B called when A had high cards, A would win again. The order of the cards was predetermined, so all dyads played the same series of 14 games. To give each player a chance to both deceive and detect, players switched roles at the end of six hands. The ante was arranged so that the total amount of money at risk in each series of six hands was the 8 same. Players thus were not monetarily penalized for being better at one task than the other. Real money was used as antes rather than poker chips, since a series of pretest games revealed that subjects preferred money and felt more motivated to win when using it. Unfortunately, subjects could not be allowed to keep their winnings. They were told that the top four money winners would win gift certificates for record albums. Given the age range of the respondents, this seemed to stimulate their desire to win. A number of players inquired about the amount they had to beat in order to win the albums. At the end of play, two free response questions were asked. One was a manipulation check: "What strategies did you use in playing the game?" The other was not: "Did you notice any behaviors in the other player that affected your strategy? If you did, what were they?" A list of the behavioral correlates of deception was the object of the second question. Having completed the game and filled out all the scales, players were thanked for their participation in this "study of game playing strategies" and told they would be contacted if they were the overall winners. Gift certificates were sent out two weeks later. CHAPTER 3 RESULTS Using SPSS (Nie, Hull, Jenkins, Steinbrenner & Bent, 1975), two 2 x 2 ANOVAs with independent groups were used to test the hypotheses. In the first case, deception was the dependent variable, in the second, detection was the criterion variable. Monitoring scores in the test condition were used as the independent variable as all 100 cases were available to be used. Only 95 subjects had answered pretest items. Means, which represent number of games won, are shown in Tables 1a and 1b. These homogeneous cell means indicated no significant treatment effects. ANOVA results, found in Tables 23 and 2b, confirmed this. While the alphas for the subscales were relatively high, (.76 for self monitoring and .75 for social monitoring), the possibility that the Lennox and Wolfe measure itself might have contributed to these results was entertained. Being a new and relatively untested scale, a confirmatory factor analysis using PACKAGE (Hunter, Cohen & Nicol, 1982) was run. Internal consistency and parallelism are the criteria by which a factor is judged to be uniformly measuring a trait (Hunter, 1980). Using these criteria, each subscale contained a weak item. Item 6 "I can usually tell when others consider a joke to be in bad taste, even though they may laugh convincingly," a social—monitoring item from the original Lennox and scale, was one; and Item 12, "Even when it might be to my advantage, I have difficulty putting up a good front", a self-monitoring question, was the other. lO Lennox & Wolfe found that Item 12 correlated .30 with other questions on the self subscale. All other interitem correlations ranged from .42 to.60. When Item 12 is removed from the cluster, the alpha is Table la Self-Monitoring and Social Monitoring Means with Original Scales Condition Level Mean Deception High Self Monitoring 1.62 (N=47) Deception Low Self Monitoring 1.82 (N=51) Deception High Social Monitoring 1.73 (N=53) Deception Low Social Monitoring 1.72 (N=53) Deception Grand Mean 1.72 Detection High Self Monitoring 4.23 (N=47) Detection Low Self Monitoring 4.31 (N=51) Detection High Social Monitoring 4.40 (N=45) Detection Low Social Monitoring 4.28 (N=53) Detection Grand Mean 4.28 Table 1b Self—Monitoring and Social Monitoring Cell Means with Original Scales Self Condition Mean monitor High social monitor High Deception 1.68 (N=25) Low Deception 1.80 (N=20) High Detection 4.16 (N=25) Low Detection 4.40 (N=20) Low social monitor High Deception 1.55 (N=22) Low Deception 1.84 (N=31) High Detection 4.32 (N=22) Low Detection 4.26 (N=31) 11 Table 2a ANOVA of Dependent Variable Deception using Original Self and Social-monitoring subscales SOURCE §§_ g;_ MS_ .3 p_ eta _5 self monitor 1.088 1 1.088 .503 >.05 .011 .104 social monitor .051 1 .051 .047 >.05 .0005 .02 self X social .179 1 .179 .164 >.05 .002 .038 S/self monitor and 102.288 94 1.088 — — .986 .992 social monitor TOTAL 103.561 97 1.068 1.00 Table 2b ANOVA of Dependent Variable Deception using Original Self and Social—monitoring subscales SOURCE §§_ g§_ M§_ .F .p eta _5 self monitor .149 1 .149 .139 >.05 .001 .031 social monitor .001 1 .001 .001 >.05 .OOO self X social .537 1 .537 .500 >.05 .005 .07 S/self monitor and 100.868 94 1.073 — - .994 .996 social monitor TOTAL 101.561 97 1.047 1.000 12 13 increased to .80. Item 6 correlated .35 with its companion social— monitoring items, whose correlations ranged from .36 to .53. While removal of this item did not change the alpha, it can be argued that its deletion conceptually tightened the cluster. The remaining five items all pertained to subjects' ability to utilize the leakage of facial cues (Ekman & Friesen, 1969). These "improved" subscales are based on an N of 100, far less than the N of 1538 recommended by Hunter, Schmidt and Jackson (1982) for statistically valid judgments. Nevertheless, the ANOVAs were rerun using these instruments. Means are found in Tables 3a and 3b. The pattern of means has shifted slightly here, when compared to the first analysis, but their uniformity again suggests little treatment effect. The ANOVA (see Table 4) again confirms this. While it is tempting to compare the changing patterns of these two analyses to see the change in distribution of subjects/cell and to note that the pattern of means in the detection condition with unimproved scales is as predicted, such observations are meaningless given the lack of significance. As noted earlier, test—retest reliabilities were calculated by correlating pretest with test scores on the monitoring scales. This was done to provide two reliability scores for each measure. The greater of these was to have been used to correct for attenuation due to error of measurement had the ANOVA results been significant. Given the outcome and the consequent development of "improved" scales, reliabilities for both original and modified measures were figured. Similar results were found using both SPSS (Nie et a1., 1975) and PACKAGE (Hunter et al., 1980; see Table 5). 14 Test-retest reliabilities (with correlations corrected for attenuation due to error of measurement in parantheses) were .65 (.88) Table 3a Self-Monitoring and Social-Monitoring Means with Modified Scales Condition Level Mean Deception High Self 1.63 (N=56) Deception Low Self 1.83 (N=41) Deception High Social 1.66 (N=50) Deception Low Social 1.77 (N=47) Deception Grand Mean 1.71 Detection High Self 4.20 (N=4.20) Detection Low Self 4.39 (N=4.39) Detection High Social 4.39 (N=4.39) Detection Low Social 4.28 (N=47) Detection Grand Mean 4 28 Table 3b Self Monitoring and Social Monitoring Cell means with modified scales Self Condition Mean Monitor High Social Monitor High Deception 1.69 (N=32) Low Deception 1.61 (N=19) High Detection 4.22 (N=32) Low Detection 4.39 (N=19) Low Social Monitor High Deception 1.54 (N=24) Low Deception 2.00 (N=23) High Detection 4.17 (N=24) Low Detection 4.39 (N=23) 15 Table 5 Test—retest correlation matrix from PACKAGE 501 502 503 504 505 506 507 508 509 501 502 503 504 505 506 507 508 509 100 97 28 28 65 65 26 25 15 97 100 27 27 65 65 23 22 11 28 27 100 95 27 27 59 58 23 28 27 95 100 31 31 58 62 26 65 65 27 31 100 97 28 28 16 65 65 27 31 97 100 28 29 15 26 23 59 58 28 28 100 97 28 25 22 58 62 28 29 97 100 29 15 11 23 26 16 15 28 29 100 501 502 503 504 505 506 507 508 509 pretest self subscale (modified) pretest self subscale (original) pretest social subscale (modified) pretest social subscale (original) test self subscale (modified) test self subscale (original) test social subscale (modified) test social subscale (original) items from all other administered scales 16 17 for the modified self scale and .65 (.91) for the original scale. For the social scale these reliabilities were .59 (.84) and .62 (.93) respectively. While these reliabilities are not as robust as one might like, they are certainly within sampling error of each other. The inflated corrected values for the original scales reflect the smaller sample size (95) thus including more sampling error. Use of the subscales as independent dimensions in this study was based upon Lennox and Wolfe's finding that they were correlated .22. Although not eXpressly stated in their paper (1984), one can assume that they administered the original scales in a laboratory setting. The correlation of .29 (.38) that was obtained using the same scale in a similar (lab) setting corroborates their finding, considering the smaller sample used here. The correlation of the modified scale, .28 (.36), in a lab setting is more supportive of their claim. Correlations of the subscales-—i.e. self X social (pretest- classroom) and self X social (test-lab) when compared to each other—— suggest that these scales are stable over time (see Table 6). That is, regardless of the conditions under which they are administered, the relationship of the two subscales to each other remains constant. Table 6 Self-monitoring and Social-monitoring Subscale Intercorrelations Modified scales Original Scales Pretest (c1assroom)--- Self X Social = .28 (.41) Self X Social = .27 (.42) Test (lab)—-- Self X Social = .28 (.42) Self X Social = .29 (.38) CHAPTER 4 DISCUSSION The conclusion most readily drawn from these results is that the manipulation was not strong enough to provide adequate tests of the hypotheses. The game did not sufficiently engage the respondents; with no deception then, there could be no detection. Responses to the manipulation check item reinforced this idea. Of the 91 subjects who responded to this question, 24% said they used no strategy in playing, 60% said they played the odds, and only 11% said they watched the other player. The fact that most subjects won when being asked to ostensibly detect (when they could see only their own cards, but were asked to call or fold) indicates that the fall of the cards alone gave away who would win. Too many high pairs were repetitively dealt to detectors in combination with low pairs to deceivers. In the detection condition, 96% of the respondents won $3.00 or more; in the deception condition, this was true of only 21% of the players. Additionally, having both players ante each time rather than having the deceiver only do it, further exacerbated the non-deception condition. The ante, which might have been used as a clue (either positive or negative) by the detector, was denuded of its possible significance. Nevertheless, the operationalization is appealing for a number of reasons: 1) it was my chairperson's idea, and seemed like a good one; 2) it is an interactive way to operationalize the construct, one which involves all subjects in both the deception and detection process; 3) it is a relatively naturalistic way of getting at the process; and 4) it is a fairly straightforward manipulation involving little deception of the respondents by the experimenter and therefore minimal debriefing. 18 19 Improvements would not be difficult to make: 1) Pairs of cards should be drawn from several decks thus obviating the player's ability to simply count pairs played. During pretest games, several combinations of pairs were changed because it was found, for instance, that a pair of kings dealt face up always elicited the same response—-P1ayer B folded because two pairs of aces had already been dealt. 2) A more careful monitoring of pairs, that is controlling whether face cards or low number pairs were dealt face up and in what combination, would encourage deception and detection. This would be particularly true if several repetitions of the same pair were dealt early in the game. 3) A method to allow play to continue so that Player A (the deceiver) would have a chance to raise the ante if he/she desired before B (the detector) called or folded might enhance interactive deception. Play might progress this way: 1) Dealer sets the ante; 2) Cards dealt; 3) Player B (who can see only his/her own pair) antes or doesn't; 4) Player A antes (raises) or doesn't; 5) Player A is dealt a second pair; 6) Player A antes or doesn't; 7) Player B antes (calls) or doesn't (folds); or 7) Player B is dealt a pair which instructs her to call or fold and A being aware that this will happen at random must control her behavior accordingly. The benefit of an effective manipulation is obvious; the effects of individual difference variables such as self and social monitoring can then be tested. The present results suggest alterations that will permit a more powerful test of the hypotheses. During the course of analyses, two additional issues emerged. A valid confirmatory factor analysis for both the subscales used is needed. Results from the minor adjustment made in the self—monitoring scale indicate that a study with a reasonably sized N is warranted. In 20 addition, while the removal of an item from the social scale made no difference in the alpha level with a small sample, it might well do so with a large one. Conceptually, this subscale might benefit from the addition of items that tap issues other than facial cues and leakage. There are certainly other kinds of cues that social monitors heed: eye contact, or lack of it, is the only one included in the current version of the measure. The stability of the scales over time was also brought to this researcher's attention during the course of this work. While the scales seem to be stable, as the intercorrelations show, no studies could be found that specifically addressed this issue. Earlier researchers (Briggs, Cheek & Buss, 1980; Gabrenya & Arkin, 1980; Lennox & Wolfe, 1984; Snyder, 1974, 1979) have confined themselves to confirmatory factor analysis of the self—monitoring scale alone. Indeed, no test—retest reliability scores could be found to be used as a basis for comparison. It appears that a high alpha level is inferred to imply stability. The logical connection seems tenuous. Future work might take a systematic look at the situational variables that may or may not affect scale administration. As administered, this study demonstrates only the effects of sampling error and error of measurement. However, these results and this discussion indicate: 1) that with the suggested alterations the manipulation can be strengthened sufficiently to properly test the hypotheses, for conceptually, a replication seems warranted; 2) that a confirmatory factor analysis with a meaningful N should be performed on the subscales; and 3) that a measurement study to confirm the test—retest reliabilities is in order. Therefore, despite the apparently 21 discouraging results reported here, some bit of progress has been made toward improving future investigations. LIST OF REFERENCES LIST OF REFERENCES Briggs, S.R., Cheek, J.M., & Buss, A.H. (1980). An analysis of the self—monitoring scale. Journal pf_Personality and Social Psycholgy, 38, Comadena, M. (1982). Accuracy in detecting deception: Intimate and friendship relationships. In M. Burgoon (Ed.), Communication yearbook 6_(pp. 446-472). Beverly Hills: Sage Publications. DePaulo, B.M., Rosenthal, R., Rosenkrantz, J., & Green, C.R. (1981). Acutal & perceived cues to deception. Unpublished manuscript. Ekman, P. & Friesen, W.V. (1969). Nonverbal leakage and clues to deception. Psychiatry, 32, 88—106. Ekman, P., Friesen, W.V., & Scherer, K.R. (1976). Body movement and voice pitch in deceptive interactions. Semiotica, 16, 23-27. Elliott, G.C. (1979). Some effects of deception and level of self-monitoring on planning & reacting to a self-presentation. Journal 9f_Personality and Social Psychology, 32, 1282-1292. Exline, R.V., Thibaut, J., Hickey, C.B., & Gumpert, P. (1970). Visual interaction in relation to Machiavellianism and an unethical act. In R. Christie & F.L. Geis (Eds.), Studies ip_Machiavellianism (pp. 53- 75). New York: Academic Press. Gabrenya, W.K. Jr., & Arkin, R.M. (1980). Self-monitoring scale: Factor structure & correlates. Personality and Social Psychology Bulletin, 6, 13-22. Heider, F. (1958). The psychology pf interpersonal relations. New York: John Wiley & Sons. Hunter, J.E. (1980). Factor analysis. In P. Monge & J. Cappella (Eds.) Multivariate technigpes ip_human communication research (pp. 229- 257). New York: Academic Press. Hunter, J.E., Cohen, S.H., & Nicol, T.S. (1982). PACKAGE: a system of routines to do correlational analysis. Unpublished manuscript, Michigan State University. Hunter, J.E., Schmidt, F.L., & Jackson, G.B. (1982) Meta—analysis: Cumulating research findings across studies. Beverly Hills, CA: Sage Publications. 22 23 Knapp, M.L., Hart, R.P., & Dennis, H.S. (1974). An exploration of deception as a communication construct. Human Communication Research, 3, 15-29. Kraut, R. (1980). Humans as lie detectors. Journal pf Communication, 39, 209-216. Kraut, R., & Poe, D. (1980). Behavioral roots of person perception: The deception judgments of customs inspectors and laymen. Journal p£_Personality and Social Psychology, 32, 784-798. Lennox, R.D., & Wolfe, R.N. (1984). Revision of the self— monitoring scale. Journal pf_Personality and Social Psychology, 46, 1349-1364. Maier, N.R.F., & Janzen, J.C. (1967). Reliability of reasons used in making judgments of honesty and dishonesty. Perceptual and Motor Skills, 33, 141-151. Mead, G.H. (1934). Mind, self and society. Chicago: University of Chicago Press. Mehrabian, A., & Weiner, M. (1967). Decoding of inconsistent communications. Journal 9f Personality and Social Psychology, 3, 109- 114. Miller, C.R., Mongeau, P.A., & Sleight, C. (1984). Fudging with friends and lying to lovers: Deceptive communication in personal relationships. Paper presented at the Second International conference on personal relationships, Madison, WI. Nie, N.H., Hull, C.H., Jenking, J.C.. Steinbrenner, K., & Bent, D.H. (1975). Statistical package for the social sciences. New York: O'Hair, H.D., Cody, M.J., & McLaughlin, M.L. (1981). Prepared lies, spontaneous lies, Machiavellianism, and nonverbal communication. Human Communication Research, Z, 325-339. Sleight, C. (1985). [Subject reports: deception cues]. Unpublished raw data. Snyder, M. (1974). Self—monitoring of expressive behavior. Journal pf_Personality and Social Psychology, 39, 526—537. Snyder, M. (1979). Self-monitoring processes. In L. Berkowitz (Ed.), Advances 32 Experimental Psychology (Vol. 16, pp. 85-128). New York: Academic Press. Stiff, J.B. (1984). Construct validity of two measures of empathy. Unpublished manuscript, Michigan State University. Stiff, J.B., & Miller, C.R. (1984a). Interrogation, level of message exposure and judgments of honesty and deceit: toward a more interactive model of deceptive communication. Paper presented at the 24 annual meeting of the Western Speech Association, Seattle, WA. Stiff, J.B., & Miller, C.R. (1984b). Deceptive behavior & behaviors which are interpreted as deceptive: an interactive approach to the study of deception. Paper presented at the annual meeting of the International Communication Association, San Francisco, CA. Zuckerman, M., DePaulo, B.M., & Rosenthal, R. (1981). Verbal & nonverbal communication of deception. In L. Berkowitz (Ed.), Advances lg experimental social psychology (Vol. 14, pp. 1—59),, New York: Academic Press. APPENDIX APPENDIX A Lennox and Wolfe Self—Monitoring Scale Items 1-7 are the self—monitoring subscale Items 8—13 are the social-monitoring subscale 1. 10. 11. 12. 13. In my social situations, I have the ability to alter my behavior if I feel that something else is called for. I have the ability to control the way I come across to people, depending on the impression I wish to give them. When I feel that the image I am portraying isn't working, I can readily change it to something that does. I have trouble changing my behavior to suit different people and different situations. I have found that I can adjust my behavior to meet the requirements of any situation I find myself in. Even when it might be to my advantage, I have difficulty putting up a good front. Once I know what the situation calls for, it's easy for me to regulate my actions accordingly. I am often able to read people's true emotions correctly through their eyes. In conversations, I am sensitive to even the slightest change in the facial expression of the person I'm conversing with. My powers of intuition are quite good when it comes to understanding others' emotions and motives. I can usually tell when I've said something inappropriate by reading it in the listener's eyes. I can usually tell when others consider a joke to be in bad taste, even though they may laugh convincingly. If someone is lying to me, I usually know it at once from that person's manner of expression. 25