Based on social identity theory (Tajfel & Turner, 1986) and distributive justice norms (Walster, Berscheid & Walster, 1976), hypotheses were set forth outlining expected changes in social evaluations of resource allocators with changes from interpersonal to intergroup situations. In interpersonal situations, fair allocators were expected to be more favourably evaluated than unfair allocators. This difference was expected to decrease, however, in weak intergroup situations (intergroup attenuation hypothesis) and reverse in strong intergroup situations (intergroup reversal hypothesis). In Expt 1 (N = 126), support for the former hypothesis was predicted and found in the minimal group paradigm. In Expt 2 (N = 82), support for the latter hypothesis was predicted and found following actual intergroup confrontation. In Expt 3 (N = 128), neither hypothesis was supported when the target and the recipients were out‐group members. The discussion extends these hypotheses to other group processes such as leadership endorsement and social influence.

Within social identity theory, in-group-favouring allocations in the minimal group paradigm are interpreted as strivings to differentiate positively one's in-group from an outgroup (Tajfel & Turner, 1986). An alternative interpretation, derived from equity theory (Adams, 1965), suggests that in-group-favouring allocations represent the perceived worth of one's in-group relative to an out-group (Bruins, Platow & Ng, 1995). We evaluated predictions based on these two interpretations in two studies. In Study 1 ( N = 73), relative in-group and out-group worth were directly manipulated. A main effect for out-group worth obtained, with participants making in-group-favouring allocations when the out-group had high but not low worth. In Study 2 ( N = 42), participants were categorized into conditions based on median splits of their apriori reported levels of perceived in-group worth and personal self-esteem. In-group-favouring allocations were made by participants with high personal self-esteem and low perceived in-group worth, and low personal self-esteem and high perceived in-group worth. Overall, the data from these two studies are more consistent with a social identity analysis than the relative ingroup worth hypothesis. Implications for social identity and equity theoretical interpretations of minimal intergroup allocations are discussed.

No abstract

The Turbo Pascal 6.0 unit, Scales, provides single-line commands for the presentation of a generic rating scale, and appropriate anchors for Likert-type scales, semantic differentials, and true-false scales. This software affords researchers a versatile and relatively simple means of incorporating rating and attitude scales into their own Pascal programs. The Scales unit (written to be used in text mode) contains 17 different features, each of which can be used independently or jointly with each other.The primary feature of the unit is the function rating_ scale. By providing appropriate parameters, this function will present a generic rating scale on the computer screen and return an integer value corresponding to the respondent's choice. Researchers have the freedom to specify scale coordinates, text and background colors, and the number of scale points (up to 12). Responses are made by moving a box on the screen via right-and leftarrow keypresses. A boolean parameter allows experimenters to permit respondents to skip questions, or to require respondents to answer each question without skipping.Three procedures in the Scales unit easily allow anchors to be placed at the endpoints of the scale. Two of these-s-likertscale and semantic_differential-anchor the scale with "strongly agree" and "strongly disagree," and specified word pairs, respectively. By providing the same coordinates and scale size that were used in the ratingscale function, the anchors are automatically adjusted to the scale. The procedure true_false assumes that the scale size is 2, and anchors the 2 points with the words true and false. Each ofthese three procedures contains parameters to allow counterbalancing of the anchors.Three of the remaining procedures display text in specified colors, either at specified x and y coordinates (LineWrite) or centered horizontally at specifiedy coordinates (center_write, HeadFoot). Three other procedures assist in the description of the rating scale usage (right arrow_key, left_arrow_key, scale_instructions). The final procedure, border_window, employs Turbo Pascal's window procedure in specifying an active window and places a specified character border around this window.Of the six remaining functions, four return specific values from the keyboard (geCchar, getjnteger, getreal, get string), The boolean function tum_the_page presents the instruction "Press C to continue" at specified coordinates and returns a value of true and clears the screen when C is pressed. The boolean function "yes" returns a value of true if a Y is pressed, and a value of false if an N is pressed.A sample of 60 subjects completed an hour-long computer-administered questionnaire employing the Scales unit features. In the final question, the subjects were asked how easy the computer rating scales were to use. On a lO-point scale (with 10 being very easy), the mean response was 8.03 (SD = 2.26). In general, they found these scales easy to use.AvaiIability. The Scales unit, its source code, full documentation, and an example program ca...