Role of physical attractiveness in impression formation

Role of physical attractiveness in impression formation! Table I Means and Standard Deviations for Picture Sets ARTHUR G. MILLER, Miami University, Oxford, Ohio 45056 Physieal A ttraetiveness Level Male and female judges were given photographs, previously scaled as high, moderate, or low in physical attractiveness, and were asked to record their impressions of the stimulus persons on an adjective checklist. The results showed high attractiveness to be associated with positive traits, the reverse holding for low attractiveness. The data are consistent with the hypo thesis that, in a first-impression situation, a person's level of attractiveness may evoke in a perceiver a consistent set of expectancies by a process of trait inference. This kind ofprocess accords weil with previous research relating physical attractiveness to interpersonal processes. Physical attractiveness has been a central independent variable in a number of recent studies. Using a gain-loss model of interpersonal attraction, Sigall & Aronson (1969) predicted, and found, that highly attractive persons were strongly disIiked if critical of naive Ss because of a greater initial drive to please attractive as opposed to unattractive individuals. Within the framework of Rotter' s (1966) development of the intemal-extemal control construct, Miller (1970) found support for the prediction that unattractive persons would be perceived as extemal in their locus of control. Several investigators have found physical attractiveness to be a significant source of variance in interpersonal perception (e.g., Byme, London, & Reeves, 1968; Mills & Aronson, 1965; Walster, Aronson, Abrahams, & Rottman, 1966). In no case, however, has the underlying basis for nie resuIts been specified. One hypothesis is that attractiveness levels are perceptually related to certain psychological traits or dispositions. When attractiveness constitutes the initial stimulus input about another person, a set of expectancies regarding other aspects of that person may be activated by a process of trait inference (Bruner, Shapiro, & Tagiuri, 1958, p. 278). This question was tested in the present study. SUBJECTS The Ss in this study were 360 male and 360 female undergraduates from introductory psychology courses at Miarni University. PROCEDURE As part of another study (Miller, 1970), 200 male and 200 female photographs Psychon. Sei., 1970, Vol. 19 (4) Sex of Judge Male Sex of Stimulus Male FemaJe Male FemaJe Female High Moderate Low Mean SD Mean SO Mean SO 5.92 5.78 6.62 6.50 1.78 1.59 1.45 1.53 4.56 4.52 4.34 4.38 1.56 1.55 1.56 1.59 3.11 3.22 2.14 2.18 1.25 1.37 1.28 1.19 6.42 6.46 6.96 7.00 1.89 1.76 1.38 1.61 3.88 3.90 4.04 4.12 1.53 1.54 1.41 1.41 2.14 2.26 1.96 1.72 1.11 1.16 1.11 0.83 Table 2 Mean Adjeetive Scale Values for Eaeh Dimension at Eaeh Level of Attraetiveness and Analysis of Variance F Ratios for Male Judges Dimension Sex of Stimulus (S) Indifferent: Curious Physieal A ttraetiveness (PA) F~AxS High Med Low Male Female 7.02be 7.08be 7.53 e 7. 28 bc 6.25 b 4.30 a 17.68** Simple: Complex Male Female 5.98b 6.00b 5. 15b 3.43 a 22.75** Insensitive: Pereeptive Male Female 7.40 c 6.83 c Careless: Careful Male Female 5.63b 7.12 e 7.08 e 7.38 e 7.53b 8.08be Praetica1: Aeademic Male Female 3.87 a 5.22 e 4.38 ab 5.02be Calm: Restless Male Female 4.08 3.97 4.53 5.27b 3.67 a 5.72 a 5.15 a 4.9Übc 3.77 a 4.04 4.08 Unsure: Confident Male Female 7. 52be 8.00 e 7. 13be 6.6Üb 4.75 a 4.28 a Submissive: Assertive Male Female 5.4°e 4. 93 be Happy: Sad Male Female 5.1 Ob 1.85 a 4.22b 3.05 a 6.35 c 5. 18b Passive: Aetive Male Female 6.92b 8.07c 4.68 a 4.07 a 75.32** Competitive: Cooperative Male Female 6. 13b 7.85 c 5.20 a 6. 28b 17.91** 19.99** 1.82 Aloof: Amiable Male Female 5.05be 5.82 e 1.53 a 2.18 a 7.70be 7.85 e 7.25 e 7.80 e 8.00 e 8.38 e 5.32 a 5.92 ab 32.11 ** 13.71 ** 3.63* Candid: Guarded Male Female 4.00bc 3.68 ab 6.48b 8.52 e 4.50 ed 3.03 a 4.98d 3. 85 be 4.27* 23.93** 2.95 Serious: Humorous Male Female 4.42d 4.17 cd 2.10 a 4.57d 3.07b 3.45be 9.42** 15.98** 14.33** Self·eontrol: Pleasure·seek Male Female 4.77 e 5.05 c 2.67 a 4.77 c 3.62h 3.48 ab 12.33** 9.48** 7.92** Reserved: Outspoken Male Female 6. 28d 4.95 c 3.43 ab 5.05 c 3.60 ab 4.00bc 12.78** Rigid: Flexible Male Female 6.95b 7.42b 4.57 a 7.43b 5.22 a 5.37 a 17.35** 8.57 e 7.48b 3.97 5.14* 3.99* 11.87** 43.65** 5.18* 4.03* 34.24** 1.67 2.89 1.06 1.91 12.92** 52.95** 1.53 15.94** 1.06 4.05* 61.10** 18.73** 15.08** 4.15* 7.82** 19.20** 10.48** Note-Differences between means which do not have a subscript in common are significant at the .05 level or beyond. The second term of each bipolar dimension indicates the plus·keyed direction of each scale. * p < .05; ** 11. < .01 ldf = 2. 354; 2 dl = 1. 354; 3 df = 2. 354 241 x 3~ in.) weIe obtained from the yearbook office. These were professional ph otographs of senior students. One-hundred male and 100 female Ss rated each of the 400 pictures on a 'nine-point scale of physical attractiveness, one being extremely unattractive, nine being extremely attractive. Means and standard deviations were obtained for each pieture, separately for male and female judges. From the distribution of means for each sex (of stimuli and judge), the upper, rniddle, and lower 10 photographs were selected. From each of thesegroups, the two photographs with the lowest standard deviation were chosen to represent the three ranges of physical attractiveness. The means and standard deviations of the stimuli are shown in Table 1. In the present study, male Ss were supervised by a male experimenter (E), female Ss by a female E. The Ss were seen in groups of as many as four, although each S was given individual instructions and worked in aseparate cubic1e. Each S was assigned randomly one of the 12 photographs relevant to his or her sex shown in Table 1. The Ss were not informed of the physical attractiveness dimension. Ss were asked to record their impressions of the person in the photograph on the Adjective Preference Scale (Jackson & Minton, 1963). This scale consists of 17 dimensions (Tables 2 and 3), each of which contains 10 pairs of bipolar adjectives in forced-choice format. The score is the number in a particular direction checked on each scale (maximum 10). The psychometrie status of this instrument is impressive, with Kuder-Richardson 20 reliabilities ranging from .55 to .96 (median .68) and low scale . t ercorre1at'Ions ( generally we11 be1ow m .30). RESULTS AND DISCUSSION Scores on each dimension for the two photographs at each attractiveness level were combined. This procedure increases the representativeness of each a ttractiveness level, since any single photograph may evoke a variety of responses in the perceiver in addition to his impression of its attractiveness. For each of the 17 dimensions, a 2 by 3 analysis of variance was computed for the factors of stimulus sex and level of attractiveness. The cell means and F ratios for male judges are shown in Table 2 and for female judges in Table 3. For both groups of judges, there are sign i fi cant effects for physical attractiveness in 15 of the 17 dimensions. A consistent pattern emerges, that of the unattractive person being associated with the negative or undesirable pole of the (2~ 242 Table 3 Mean Adjeetive Seale Values for Each Dimension at Each Level of Attractiveness and Analysis of Variance F Ratios for F emale J udges Physical A ttractiveness (PA) Dimension Sex of Stimulus (S) Indifferent: Curious Simple: Complex Insensitive: Pereeptive Med Male FemaIe Male Female Maie Female 8.08d 6.88be 7.33d 6.55 ed 6.70be 6.47b 5.97be 5. 37 b 6. 32 b 7.08be Careless: Careful PraetieaI: Aeademic Calm: Restless Unsure: Confident Submissive: Assertive Happy: Sad Passive: Aetive Competitive: Cooperative Maie FemaIe Maie Female Maie FemaIe Male Female Male Female Male FemaIe Male Female Male FemaIe 7.38a 6.85 a 8. 28 b 8. 92 b 4. 75 b 4.08 ab 4.73b 4.22ab 4.88d 4.33 ed 4.27bed 3.55 ab 7.97de 7.13ed 8.42e 6.85 e 6.02e 5. 15 be 5. 47 be 3.57 a 1.97b 2.82b 0.85 a 1. 97 b 8.22de 7.20be 8.37e 7.57ed 6.97 ab 6.60 a 7.48ab 8.92e 7.78 ed 4.87a 5. 30b 3.72a 6. 58 b 5.07 a 8. 30 b 6.58a 4.45b 3.68a 3.83abe 3.50a 5. 13 b 4.20a 4. 93 b 3.02a 4.15 e 5. 12d 6.77b 4.62a 6.73 ab 7. 65 b Aloof: Amiable Male FemaIe 8.52be 7.90b 7.03 a 9.17 e Candid: Guarded Serious: Humorous Self-eontrol: Pleasure-seek Reserved: Outspoken Rigid: Flexible Male Female Maie FemaIe Male Female Maie Female Male Female 3.63 ab 5.07e 4.48 e 4.33e 5. 92 d 4.77 e 3.92b 3.02a 3.72be 3. 37 b 4.62 e ~.93b 5.55b 2.70a 6.13b 7.87e 7.75 e 7.65 e 6.32b 5. 28 b 7.0%e 6. 67 b Low Fl F2 _ __.:.xPA.'L_ _~;LS High -=:.'~_" 5.27** 22.48** 6.58** 24.14*' 11.95 .* 1.11 16.22** 1.11 6.12'* 1.21 2.69 19.37** 4.61 • 1.20 1.98 6.86** 7.92** 61.25** 2.28 14.93** 23.63** 42.52** 1.42 9.52*· 47.01** 6.00* 13.13*· 2.18 25.25** 4.81·' 6.70a 7.02a 11.70" 5.83* 10.17*' 4.02b 3. 80 b 2.33 a 2.02a 2.18 ab 1.97a 2.53 a 1,;.68a 4.72 a 6.1% 5.84'* 27.95** 2.19 10.73·* 1.23 59.49" 17.28*' 3.07' 53.83*' 29.41** 4.81** 13.33*' 11.25** 5.33** Note-Differences between means which da not have a subscript in common are significantat the .05 level or beyond. The second term of each bipolar dimension indicates the plus·keyed direction of each scale. ..ldfp = < 2,.05;354;.... p.2 < .01 df = 1, 354; 3 df = 2, 354 adjective scales and the highly attractive person being judged significantly more positively. The status of modera tely attractive persons is variable, generally falling between the high- and low-attractive stimuli, but not significantly different from both extremes. Physical attractiveness, thus, is a potentially strong determinant of first impressions. The effect is pervasive, occurring in a large array of impression responses and with respect to male and female stimulus persons. That male and female judges responded to different sets of photographs serves as a kind of replication and adds to the generality of the fmdings. There are a number of significant effects for the sex of the stimulus persons. Some reflect sex -role stereotypes, e .g., female judges perceiving females as more simple, subrnissive, passive, and reserved, whereas other sex effects seem to lack psychological meaning or generality, e.g., male judges seeing males as significantly more sad than females. Regarding the Sex by Attractiveness interactions, exarnination of the pairs of cells at each attractiveness level reveals seven significant differences between male and female photographs at the high-attractive level, 20 at the moderate-attractive level, and 18 at the low-attractive level. It appears that as one departs from high-physical attractiveness, a stimulus person's sex becomes a more influential impression determinant. However, as previously stated, the precise meaning or significance of different impressions of male and female stimulus persons is not always c1ear. In troducing a Psychon. Sei., 1970, Vol. 19 (4) behavioral consequence of the first BRUNER, J. S., SHAPIRO, D., & TAGIURI, R. The meaning of traits in isolation and in impression response might clarify the combination. In R. Tagiuri and L. Petrullo implications of such interaction effects. A (Eds.), Person perception and interpersonal plausible hypothesis in this context might behavior, Stanford: Stanford University Press, be that unattractive males are perceived 1958. Pp. 276-288. more adept at compensating for their BYRNE, D., CLORE, G. L., JR., & GRIFFITT, W., Response discrepancy versus attitude unattractiveness than are females, i.e., if similarity-dissimilarity as determinants of one must be unattractive, it is a better fate attraction. Psychonomic Science, 1967, 7, to be male than female. 397-398. Most significant, in view of the present BYRNE, D., LONDON, 0., & REEVES, K. The effects of physical attractiveness, sex, and data, is the relative paucity of experimental attitude similarity on interpersonal attraction. interest in the attractiveness variable, as Journal ofPersonality, 1968, 36,259-271. Sigall & Aronson {l969, p.93) have JACKSON, D. N., & MINTON, H. L. A indicated. For it seems quite probable that forced-choice adjective preference scale for personality assessment. Psychological Reports, this dimension has real significance as an 1963,12,515-520. antecedent of "liking" or interpersonal A. G. Social perception of attraction, in addition to such well-studied MILLER, internal-external control. Pereeptual & Motor factors as propinquity (Newcomb, 1961), 'ikiJIs, 1970, 30, 103-109. cognitive balance (Aronson & Cope, 1968), MILLS, J., & ARONSON, E. Opinion change as a funetion of the communicator's attractiveness and attitude similarity-dissimilarity (Byrne, and desire to influenee. Journal of Personality Clore, & Griffitt, 1967). Questions & Social Psychology, 1965, I, 173-177. suggestive of needed research are: When do NEWCOMB, T. M. The acquaintance process. the trait implications of physical New York: Holt, Rinehart, & Winston, 1961. attractiveness appear, developmentally, and ROTTER, J. B. Generalized expectancies for internal versus external control of what are the sources of these relationships? reinforcement. Psychologieal Monographs, To what extent are correlates of physical 1966,80, l(Whole No. 609). attractiveness empirically based, in SIGALL, H., & ARONSON, E. Liking for an addition to being assumed, as shown in the evaluator as a function of her physical attractiveness and nature of the evaluations. present data? Does attractiveness influence Journal of Experimental Social Psychology, social interaction in the manner of a 1969,5, 93-100. self-fulfllling prophecy, i.e., if unattractive WALSTER, E., ARONSON, V., ABRAHAMS, persons are perceived as inadequate on D., & ROTTMANN, L. Importance ofphysical attractiveness in dating behavior. Journal of personality factors, are such assumed PersonaJity & Social Psychology, 1966, 4, correlates subsequently "validated" in 508-516. interaction? What social context and individual difference factors increase or decrease the importance of-or interact with-the attractiven'ess variable? REFERENCES ARONSON, E., & COPE, V. My enemy's enemy is my friend. Journal of Personality & Social Psychology, 1968,8,8-12. NOTE 1. The writer would like to thank Wesley Penn and Betty Jerger for their assistance and David Probert of the Miami University Computing Center. The effects of sentence length and grammatical structure in aserial learning task 1 HO WARD B. ORENSTEIN2 and - Although the facilitative effects of DONALD A. SCHUMSKY, University of structure (i.e., syntax and/or morphology) Gncinnati, Gncinnati, Ohio 45221 on recall of nonsense strings have been demonstrated by Epstein (l961, 1962) and Nonsense strings eontaining four, five, F orster (1966), a number of recent studies six, and seven nonsense words were have failed to fmd such positive effects. eonstrueted for three grammatieal Bogartz & Arlinsky (1966) did not obtain strueture eonditions: morphology and syntactical facilitation when immediate syntax, morphology alone, and syntax ordered recall of strings that possessed alone. Analysis of the pereentage ofwords both function words and bound reealled revealed a signifieant leaming morphemes in syntactic order was effeet that was inversely related to compared to recall of strings possessing sentenee length. More important, however, either function words or bound lWlS a lai/ure to find any facilitation effeets morphemes alone. However, the facilitative on reeall due to morphology and/or effects of function words and bound syntax. morphemes were found to be significant Psychon. Sei., 1970, V01. 19 (4) when compared to appropriate control strings that lacked either of them. Bogartz and Arlinsky, however, did not include a control condition in which both function words and bound morphemes were present but not in syntactic order. O'Connell, Turner, & Onuska (1968), using orally presented strings in which such a control was present, found that syntax did not facilitate recall. Further negative findings were reported by Rosenberg (1964), who found that the addition of bound morphemes in syntactic order to short (fIVe-unit) strings did not facilitate their immediate recall. Bryk & O'Connell (1967), using strings (10 units) adopted from Epstein's (1961) original strings, tested for immediate recall under three levels of constraint: no morphology and no syntax (NS), morphology alone (LS), and morphology and syntax (HS). They found that the high structure condition (morphology and syntax) was significantly different from the condition of no structure. The difference between the high and low levels of structure was nonsignificant. Consideration of the available evidence suggests that the influence of syntactic and/or morphological cues may, among other things, be dependent upon the length of the string itself. Bryk (1968) suggests that syntactic facilitation may occur primarily because of "S's facility in organizing the individual items in astring to form more easily storable and recallable chunks of information [p. 3]." Extending Miller's (1956) concept of chunking, Bryk states that the advantages of gramrnatical structure should increase as the amount of information to be stored increases. Thus, for relatively longer strings "prechunked" syntactic strings should be more apparent than for the relatively shorter strings. In a 3 by 3 by 3 factorial design, Bryk used immediate written recall of nonsense strings to study the effects of: structure (NS, LS, HS), meaningfulness of nonsense sterns, and string length (5,9, 12). He states that bis results revealed that recall of HS strings relative to NS and LS strings increased as a function of increases in the meaningfulness and length of the nonsense string. His failure to find syntactic facilitation for the five-unit strings suggests that previous failures (e.g., Rosenberg) to find such facilita tion are explainable in light of the relatively short string length used. However, careful examination of the procedure and results leads to a different interpretation. Since the lengths of strings contained different numbers of items, analysis of the mean number of items recalled, rather than the percentage of Hems recalled, indicates a possible confounding. It is impossible to determine 243