Table 2.

Social judgment theory

Theory	Design
Social Judgment Theory (SJT) studies the relationship between the case factors, or cues, and the assessments and decisions which are actually made (research questions 2 and 3). It is based on the notion that people must make assessments (judgments) based on the information available to them and that this is often incomplete or ambiguous. From available information they make inferences about the “true” situation. Exploring the way in which particular information factors are used provides evidence about judgment accuracy and variability between the judgments people make.	In SJT vignettes are used in which the same case information (factors) is presented to each participant. Using SJT vignettes may be narrative or graphical in form although where SJT and SDT are combined graphical vignettes are used. Factors included in vignettes are typically elicited from people who are experienced in making the assessment to be studied through interviews from which relevant factors are abstracted. The same factors are included in each vignette however, the level or weight of each factor is randomly varied across the vignettes, to provide a range of risk levels (0-100) for each (0 represents no concern and 100 the highest possible concern). Selection of relevant factors and realistic factor weights is achieved by developing and piloting vignettes with appropriate experts, this ensures that although the vignette format may be abstract the factors and weights are recognisable (ecological validity) as those which could occur in real life.
Administration	Analysis
Study participants are asked to rate the overall level of risk (0-100) for each vignette and decide whether they would act or not act. Vignette tasks using SJT can become very large. Between five and 10 vignettes are required for each factor included to allow for the regression analysis; therefore no more than 10 factors are usually included in each vignette so that the overall task does not become onerous for participants.	Judgment analysis identifies the relative contribution of each factor and weight of each factor to the overall assessment of risk in the case, and the decision to act (i.e. what factors are used and how they are used). Linear regression is used to model the continuous judgment about level of risk in each vignette (0-100) and logistic regression to model the dichotomous choice (act or no action). Varying the factor information presented to study participants across vignettes, allows the responsiveness of clinicians to differing factor information to be established, this is fit and is measured by the multiple correlation coefficient (values above 0.6 are expected). Comparison of scores between participants identifies variability within and between clinician’s judgments. Mean scores between individuals and groups are compared using t-test for independent samples. Repeat cases are used to identify judgment consistency.