6.1 Method

6.1.1 Materials

Two main kinds of materials were used in this project.

The Everyday Situational Judgment Inventory. The Everyday Situational-Judgment Inventory (ESJI) consists of descriptions of various situations encountered by many people. After each situation, there are 8 options for handling the situation. For each option listed, participants were asked to rate the quality of the option on a 1 (low) to 9 (high) Likert scale, where the anchor points were 1=extremely bad, 3=somewhat bad, 5=neither bad nor good, 7=somewhat good, and 9=extremely good. Participants were asked to select the number corresponding to their judgment, and to write it in the blank preceding each option. Participants were told that there was no one "right" answer—that the options were simply things that people might do in the situations described.

An example of an item is as follows:

You've been assigned to work on a project for a day with a fellow employee whom you really dislike. He is rude, lazy, and rarely does a proper job. What would be the best thing for you to do?

• Tell the worker that you think he is worthless.
• Warn the worker that, if he is not "on his toes" today, you will complain to the supervisor.
• Avoid all conversation and eye contact with the other worker.
• Be polite to the other worker and try to maintain as business-like a manner as possible so that hopefully he will follow your example for the day.
• Tell your supervisor that you refuse to work with this man.
• The project is going to be impossible to accomplish with this worker, so you may as well not even try — you can always blame your bad work partner.
• See if you can convince one of your friends to take your place and work with this employee.
• Demand a raise from your supervisor; you should not have to tolerate these conditions.

Participants were given as much time as they needed to finish the inventory. Scoring for the ESJI was done in three ways:

1. Profile matching (d²). For each of the 30 problems, a given respondent's responses were compared to the averaged (prototypical) responses to that problem. The following specific scoring procedure was used. For a given option, the difference between the participant's response and the sample-mean response was computed and squared. Squared differences were summed across the 8 options and averaged. Then the square root of this average was computed. The same procedure was repeated for each of the 30 items. Total score was the sum of these values.
2. Rank-order correlation between individuals and mean profile (ρ). For this measure, the rank orders of the responses for the mean profile of responses to a given item were correlated with the rank orders of the responses for an individual's profile of responses for that item. Thus, 8 observations for an individual were correlated with 8 observations for the mean profile. This analysis yielded a rank-order correlation (rho, or ρ) for each item. These correlations were averaged across the 30 problems.
3. Dichotomized responses based on significance of ρ. Dichotomized scores were created for each of the 30 items by assigning the item a score of 1 if the ρ value for that item was statistically significant and a 0 otherwise. In other words, the respondent got credit (1) if the respondent'ys item response pattern rank-order correlated significantly with the averaged item response pattern.