Sticker and Rock (1990) administered a behavioral situational judgment test, the Interpersonal Competence Instrument (ICI), to 131 undergraduates, along with other measures of social cognition (e.g., peer and self-ratings, accuracy in decoding nonverbal communication) and measures of general skill (e.g., verbal comprehension, general reasoning). Using multidimensional scaling analysis, they found little evidence of convergent or discriminant validity among the measures of social cognition and general skill. Some of the social-cognition measures appeared to tap verbal skill, whereas others seemed to measure general reasoning skill. In contrast to the findings of Ford and Tisak (1983), these findings failed to support the hypothesis that behavioral measures of social cognition would be more distinguishable from measures of general academic cognition than would be verbal measures of social cognition.

Brown and Anthony's (1990) findings suggested that the constructs of social and academic cognition are distinct, but potentially interrelated. They evaluated the relationship of GPA and American College Test (ACT) English and Mathematics scores of college freshman to self and peer ratings of personality and behavioral dimensions of social cognition. Using a principal-components analysis, they identified three distinct components in their data, represented by an academic component, a peer-ratings component, and a self-ratings component. They concluded that social cognition could be separated from academic cognition. Social cognition as perceived by others was also distinct from one's own assessment of social skills. However, they also found that GPA was the best predictor of self and peer ratings on behavioral aspects of social cognition, suggesting a relationship between social skills and school performance.

2.1.3 Nonverbal approaches to measuring social cognition

In addition to behavioral approaches to measuring social cognition, researchers also sought to distinguish social from academic cognition by pursing nonverbal measures. Nonverbal approaches to measuring social cognition assess primarily nonverbal decoding skills (Archer, 1980; Archer and Akert, 1980; Barnes and Sternberg, 1989; Rosenthal, 1979; Rosenthal, Hall, DiMatteo, Rogers, and Archer, 1979; Sternberg and Smith, 1985). Rosenthal et al. developed the Profile of Nonverbal Sensitivity (PONS) test, which presents a single woman in a variety of poses. Participants are asked to decode the implicit signals being emitted, and to figure out which of two alternative descriptions better characterizes what the test taker has seen and/or heard. The PONS has been found to have weak to moderate correlations with other measures of social and cognitive competence (Halberstadt and Hall, 1980; Rosenthal et al., 1979).

Archer (1980; Archer and Akert, 1980) developed an alternative to the PONS test, called the Social Interpretation Test (SIT). The SIT presents participants with visual and auditory information regarding a social situation. For example, the participant might see a picture of a woman talking on the phone and hear a fragment of the woman's conversation. The participants are asked to judge whether the woman is talking to another woman or to a man. In another situation, participants are asked to judge whether a man and woman shown in a picture are strangers, acquaintances, or friends. Research using the SIT has focused primarily on the accuracy of participants' judgments based on verbal versus nonverbal information.

Using a task similar to the SIT, Sternberg and Smith (1985) developed a measure of decoding skills and assessed their relationship to other measures of social and cognitive cognition. They presented participants with two types of photographs. In one type, a man and woman were shown posing as if they were in a close relationship. Participants were asked to judge if the photograph depicted a real or a fake couple. In the second type, the picture showed a supervisor and his or her supervisee. Participants were asked to judge who of the two individuals was the supervisor. Accuracy was assessed as the percentage of pictures the participant judged correctly. Participants were also given several measures of social and cognitive cognition including the PONS (Rosenthal et al., 1979); the Social Insight Test (Chapin, 1967); the George Washington Social Cognition Test (Moss et al., 1949); the Group Embedded Figures Test (Oltman, Raskin, and Witkin, 1971); and the Cattell Culture Fair Test of g (Cattell and Cattell, 1963). Nonverbal decoding accuracy only correlated significantly with performance on the Embedded Figures Test. Sternberg and Smith concluded that there was insufficient evidence to suggest that nonverbal decoding skills provided a valid measure of the construct of social cognition.