A reply to Norman H. Anderson's critique of the Subject-Verb-Object approach to social cognition.

Abstract

The Subject-Verb-Object (S-V-O) approach to social cognition provides a method of conceptualizing several social cognition problems within the context of item sets that contain eight subject-verb-object sentences, such as "The kind man avoided the intelligent swindler" and "Conservative voters enjoy golf." Anderson recently criticized the S-V-O model on numerous grounds that were primarily technical in nature. The present reply refutes all of Anderson's major criticisms. In discussing an alternative to the S-V-O model, Anderson recommends that problems raised by the pervasive tendency of subjects to respond to information configurally be dealt with by collecting data that can be described by simple algebraic models. It is argued that Anderson's alternative avoids, rather than deals with, the important issues raised by the configural responding of subjects. Although the specific content of the present article is primarily concerned with the S-V-O model, much of the discussion is relevant to issues of general importance to investigators who use analysis of variance models and/or simple algebraic models in their work. A postscript replies to the article by Anderson that follows this one. Anderson (1977) recently wrote a detailed critique of Gollob's (1974aGollob's ( , 1974b Subject-Verb-Object (S-V-O) approach to social cognition as well as of and Insko, Songer, and McGarvey's (1974) applications of that approach. Wyer and his co-workers (Wyer, 1974, 197S; Wyer, Henninger, & Wolfson, 197S) have extended Gollob's model and have applied it in a broad variety of areas concerned with social inference and attribution theory. Insko et al. applied the S-V-O model to the analysis of pleasantness judg-

Social Inference

“…In a second experiment social desirability (virtuousness) and competence were shown to underlie the admirableness judgment. Gollob (1979) and N. H. Anderson (1979) authored two parts of a heated and instructive exchange concerning the merits of the subject-verb-object approach. The level of discourse in this exchange is remarkably sophisticated, but accessible, and it exemplifi es the discussion of methodological and theoretical issues in the context of a well-specified theoretical approach.…”

Indeterminate theory: Reply to Gollob.

“…If Gollob (1974) considered that two different sets of weights were involved, he should have said so then. Since Gollob (1974) does not even allude to two different sets of weights, the complaints in Gollob (1979) seem ill-taken.…”

“…The central criticism in my discussion of the balance triad model is conceded by Gollob to be correct (Anderson, 1977;Gollob, 1979). Since this central issue is becoming lost under a mass of technical detail, it deserves reemphasis in simple terms: The balance triad model yields a theoretical representation of the response to a single sentence that is arbitrary and indeterminate.…”

“…In his reply, Gollob (1979) agrees with this criticism: The model "assumes that the judgment of a given sentence is a linear combination of the various psychological biases," but the model analysis "does not yield direct estimates of the weights for that linear combination" (p. 940).…”

The effects of behavioral intentions and consequences on judgments of the actor and other: An S-V-O analysis.

“…The approach taken to investigate the issues outlined above was stimulated by the subject-verb-object (S-V-O) formulation of social cognition developed by Gollob (1974b). This formulation, which has been applied successfully to a variety of problems in impression formation and attribution (Gollob, 1974a(Gollob, , 1974bWyer, 1974, 197S;Wyer, Henninger, & Wolf son, 1975;Wyer & Hinkle, 1976) is described in detail elsewhere (Gollob, 1974b(Gollob, , 1979Wyer & Carlston, 1979) and so will not be elaborated here. According to this approach, the information in a social interaction situation is analyzed by the judge into a series of informational cues, each cue defined in terms of values along one or more dimensions.…”