research in the so com lab

People routinely make social judgments about others (e.g., sex, race, and age), and such judgments have been described as obligatory.  It is unsurprising, therefore, that the study of person perception enjoys a long history in both social and cognitive psychology.  Yet social and cognitive psychologists have approached the study of person perception using drastically different methods and in virtual isolation from one another.  Social psychologists have focused on the fact that social categorizations are pervasive and that they carry serious consequences (e.g., social costs of being categorized to be a woman or minority).  The insights from this research provided breadth, but could be regarded as lacking precision.  Cognitive psychologists, in contrast have focused on precisely modeling the parameters that engender specific categorizations (e.g., the biomechanical parameters that characterize body motion). Consequently, the insights from cognitive research were precise but lacked breadth. 

My research borrows methods from both cognitive and social psychology to better understand how people reach obligatory perceptions of others.  I use a variety of stimuli and methods.  My stimuli range from carefully controlled computer animations that afford precision to more naturalistic videos of real people that afford breadth.  My methods include psychophysiological and behavioral measures (e.g., facial EMG, corneal reflection eye-tracking, motion analysis, and skin conductance), social cognitive measures (e.g., response latency, judgment biases, etc.), and subjective evaluative judgments. 

My approach – to regard person construal to be a fundamentally social process, and to exploit methods and techniques that are more common among cognitive psychologists  – has proven useful to understand the complexities of basic person construal.  I have found, for example, that observers are prone to look toward particular regions of the body when judging the sex of others, and that this tendency changes after providing observers information about the target’s social category membership.  In other research, I have described how some body cues that are generally reliable indicators of one social category (e.g. sex-typical body motion) are interpreted differently in the context of other social information.  Knowledge about a target’s sex, for example, changes how body motions are interpreted – affecting both evaluative social judgments such as perceived attractiveness and social categorizations such as perceived sexual orientation.  Similarly, when body motions that differentiate men from women are enacted in stereotyped emotional states such as anger versus sadness, observers’ sex category judgments are more closely aligned with the irrelevant stereotyped information, rather than on the embodied differences between men and women.  Thus, my approach to questions of person construal has revealed the combined importance of both bottom-up (such as physical parameters) and top-down (such as prior social knowledge or stereotyped) processes.

Click here to see sample stimuli.

Mechanisms of Person Perception

•Perceiving Sex:  Much of my research has explored how observers use physical cues to make categorical judgments of a target’s sex.  Two cues in particular, the body’s shape and motion, have received considerable attention, but prior studies investigating the perception of these cues have isolated either body shape or body motion.  In my work, I have brought these two cues together in computer animations.  These animations vary in body motion including gaits that vary from a masculine shoulder swagger to a feminine hip sway – motions embody the actual gait of men and women and extremetizations of those gaits.  These animations also vary in body morphology including waist-to-hip ratios that vary between a curvaceous 0.5 to a more tubular 0.9 – morphologies that embody the majority of the range of human bodies and extremetizations of those morphologies.

      1.  In one set of studies (Johnson & Tassinary, 2005), we explored how body cues differentially affect perceptions of sex and masculinity/femininity.  Using corneal reflection eye-tracking, we found that participants concentrated their visual scanning of a sexually dimorphic region of the body (i.e., the waist and hips) only when the sex of the target was unspecified to them.  When participants already knew the sex of the target, their visual scanning was evenly distributed across the targets’ bodies.  Additionally, using mediational analyses we demonstrated that perceiving sex from body motion is the product of inference.  First, body motion is perceived to be masculine or feminine, and then the target’s sex is inferred.  

2.  In another set of studies (Johnson & Tassinary, 2007), participants provided categorical judgments of each target’s sex.  Both body shape and motion proved to be important determinants of perceived sex, yet the effect size was substantially larger for body shape.  Body motion, in contrast, was more important for judgments of perceived masculinity and femininity.  This finding contradicts other research that found motion alone to be sufficient for accurate sex perception. 

•Perceiving Sexual Orientation:  Using both animations and real targets, I have explored how the gender typicality of body motions affects observers’ perceptions of sexual orientation. 

1.  In one set of studies (Johnson, Gill, Reichman, & Tassinary, 2007), participants categorized targets according to perceived sexual orientation.  Overall, the gender typicality of a target’s motion affected perceived sexual orientation.  Targets that were judged to be women were more likely to be judged heterosexual when they walked in a characteristically feminine way than when they walked in a characteristically masculine way.  The opposite pattern obtained for targets that were judged to be men, and the magnitude of the effect was substantially stronger among perceived male targets, a finding that corroborates developmental findings that gender atypicality among men and boys is more heavily penalized than is gender atypicality among women and girls.  Moreover, when judging the sexual orientation of male targets, participants rely solely on body motion to inform their judgments of sexual orientation.  When judging the sexual orientation of female targets, in contrast, participants rely on both body shape and body motion to inform judgments.  This finding helps explain why other researchers have found sexual orientation categorizations from photographs to be accurate for female, but not male, targets. 

2.  In another set of studies (Reichman & Johnson, in prep) we have explored how these perceptions relate to participants’ judgments of targets’ intent to convey their sexuality to observers and participants’ acceptance of various targets.  We found that feminine body motion, in general, is perceived to be a deliberate attempt to communicate to observers, regardless of whether the feminine motion is exhibited by a woman or a man.  When a motion is perceived to be intentional and it is also gender atypical (i.e., when men walk with feminine body motion), it tends to meet disapproval.  These findings help explain how people can simultaneously profess liberal attitudes toward homosexuality yet still stigmatize homosexual individuals. 

3.  Finally, we have recently completed studies in which we used a three dimensional motion capture system, Optotrak, to record the body motions of men and women who self identified to be gay or straight.  These data are currently being used in several projects.   First, we have related body motion parameters to observers’ judgments of sexual orientation as well as the accuracy or error associated with those judgments (Johnson, Gill, Reichman, & Tassinary, 2007).  Second, we are specifying similarities and differences between the movements of each group at a biomechanical level.  Finally, we are examining how stable individual differences (i.e., one’s sex role identity) are revealed through nonverbal body motions. 

•Evaluative Social Judgments: Other research has explored the mechanisms by which specific body shapes are perceived to be attractive.

1.  I have explored how both body shape and body motion affect perceived attractiveness (Johnson & Tassinary, 2007).  Participants in these studies judged the sex, masculinity/femininity, and attractiveness of animations and real walkers.  Gender typical combinations of body shape and motion were judged to be more attractive than were gender atypical combinations, and this finding obtained in several conceptual replications including studies that:  a) present stimuli from different visual perspectives, b) relate perceived attractiveness to motion and morphology parameters, c) relate perceived attractiveness to the perceptions (i.e., perceived sex and masculinity/femininity) that those parameters engender, d) manipulate the purported sex of ambiguous targets, and e) measure the perceptions of real human walkers and animations.  Collectively, these data shed light on how body cues are evaluated by observers, and they help explain a range of contradictory findings for how (and why) the waist-to-hip ratio affect judgments of women’s attractiveness.

2.  I have described perceived attractiveness as interpersonal metaperception (Tassinary & Johnson, in press). This theoretical approach characterizes perceived attractiveness (and other more global evaluations) as a social Gestalt that emerges from the compatibility of more automatic and obligatory social perceptions (e.g., sex, gender, race, and age).

•Person (Mis)Perception:  In several studies, I have explored whether people’s cognitive representation of men’s and women’s bodies are accurate, or error prone.

1.  In one set of studies, I have compared people’s perceptions of the distribution of women to men varies across different body shapes to the actual distribution of women to men who embody those body shapes.  Perceptions and reality differ considerably.  In some studies, participants provided categorical sex judgments for bodies that appeared sequentially on a computer screen, and both judgments and their latency are evaluated. Observers perceive variability in body shapes that, in reality, are exclusive to women, and they readily report extreme body shapes (specifically extremely curvaceous hourglass silhouettes) to be women, even though such extreme bodies do not appear in nature (Johnson & Tassinary, in prep).  In other studies, I have examined the extremity of the perceived prototypical man and woman by asking participants to identify from an array the bodies that best represent the average man and woman.  Participants willingly identify extreme bodies to represent these averages – so extreme, in fact, that they are not represented in nature.  The perceived female prototype has a more extreme waist-to-hip ratio (i.e., is more curvaceous) than the anthropometric mean among women; the perceived male prototype has a more extreme chest to waist ratio (i.e., a more pronounced V shape) than the anthropometric mean among men.  Importantly, these effects obtain even when participants are advised to “beware, the experimenter included impossible bodies.” 

2. Another set of studies continues to explore the mechanisms that undergird person misperception.   Specifically, these studies: a) ruled out the possibility that these extreme prototypes emerged merely due to scaling effects, b) explored the possibility that sex perception fits the narrow definition of true categorical perception, and c) are exploring how these categorical extremes emerge.

3.  Finally, another set of studies explored (Johnson, Lurye, & Tassinary, under review) and continues to examine the development of sex perception among children between the ages of 4 and 6 (an age that is coincident with rapid changes in gender identity). In one task, children categorize static stimuli to be either male or female using a modified card-sort task. In another task, children identify the “best” man and woman from a line-up (analogous to the prototype studies described above). In a final task, children judge the sex and masculinity/femininity (using an age appropriate scale) of computer-generated targets.

Stereotypes Resolve Social Ambiguity

•Race is Gendered:  My collaborators and I are exploring how social categories of race and sex are intertwined.  Put simply, we are exploring how race is gendered.  Stimuli include faces that are categorically distinct (i.e., male or female for sex; white, black, Asian for race) and categrorically ambiguous (morphs between sex and race categories).  In separate trials, participants in these studies categorize stimuli by sex or by race.  We are looking for stereotype consistent facilitation effects – how one unambiguous (and stereotyped) category may facilitate categorization along the other dimension (e.g., unambiguously black faces facilitating “male” judgments; unambiguously female faces facilitating “female” judgments).  Additionally, we are exploring the time course of these judgments by examining reaction times and the trajectory of mouse movement.  Finally, using fMRI, we are exploring the cortical mechanisms that are recruited to resolve social ambiguity in race/sex.

•“Throws Like a Girl”:  In another set of studies (Johnson, Pollick, & McKay, under review), we are exploring how sex stereotypes that involve emotion (e.g., women respond with sadness and men respond with anger) can disambiguate or mislead judgments of a target’s sex.  Stimuli are point-light displays of men and women throwing a ball in four emotional states (sadness, happiness, anger, and neutral).  Participants categorized the “point-light-throwers” by sex, rated their confidence in their categorization, and judged the depicted emotion.  Judgments of stereotyped emotions (i.e., anger and sadness) are more error-prone (i.e., angry throws are presumed to be men, sad throws are presumed to be women), yet participants are most confident in these judgments.  Thus both top-down (stereotype consistency) and bottom-up (motion parameter perception) affect person perception.

Self-Relevance of Person (Mis)Perception

•Extreme Comparisons:  One consequence of Person (Mis)Perception has self-relevance.  In several studies (Johnson, in prep), I have explored how misperceiving social prototypes sets up unflattering social comparisons (especially among women) that compromise body satisfaction.  In these studies, participants are confronted with the actual or the perceived male and female prototypes.  Then, various aspects of body image and satisfaction are measured.  Implicitly comparing oneself to extreme prototypes (whether manipulated by the experimenter or by exposing participants to their own prior judgments) leads women to perceive their own bodies to be significantly larger, relative to perceptions of women who were exposed to more realistic prototypes.  Currently, we are using facial EMG to explore the immediate affective consequences of these social comparisons.

•Implicit Self-Judgments:  Much of my current research employs exciting new technology that permits self-judgments that are genuinely implicit.  New software, developed by my collaborator Lou Tassinary and his visualization students at Texas A&M University, allows us to input up to 10 different body measurements and generate a computerized body that depicts those precise proportions.  In current research, we collect physical measurements of participants upon their arrival to the lab, and we generate a stimulus using these data.  Then, this self-image is imbedded in a series of normed stimuli that represent the full range of anthropometric variability found in nature.  Participants judge each of these stimuli, including their own body – unbeknownst to them, along several dimensions.  This set of judgments provides two things:  First, these data provide a foundation upon which we can weigh the relative importance of body cues for social judgments (using judgments of the normed stimuli).  Second, it provides an implicit self-judgment.  Both of these can be compared to judgments that are later made explicitly about one’s own body.  We’re finding that participants’ implicit and explicit self-judgments tend to differ, but in a somewhat surprising way.  For women who are high in eating disorder symptomatology, implicit and explicit self-judgments tend not to differ.  For women who are low in symptomatology, in contrast, explicit self-judgments are more favorable (i.e., more attractive and less fat).  Thus, it appears that women who are high in symptomatology are also robbed of their tendency to self-enhance.  Currently, we are exploring the relation of these findings to depressive realism and more general self-enhancement biases.  Additionally, we are using facial EMG to determine whether the favorable explicit self-judgments observed among low-symptomatic women are immediate (correlating with immediate zygomatic activation) or effortful (implying that corrugator activity should precede zygomatic activity).

•Appearing Extreme: A final set of experiments, currently underway, relates this line of research with the program of research described previously.  In these studies, we are exploring the circumstances that lead individuals to convey masculinity or femininity to onlookers via their gait (and their success, or lack thereof, in doing so).  In one version of these studies, participants are asked to walk on a treadmill while being recorded, and we manipulate participants’ expectations of precisely how the recordings will be used.  Some participants are told that their walk motions will be analyzed using a computer algorithm to quantify the parameters of human gait; other participants are told that their walks will be evaluated by opposite sex/same sex peers to examine the determinants of perceived attractiveness.  We expect that both men and women will produce more gender-typical motions when they assume their motions are for social consumption.