An analysis of the effects of the forced-choice format of rotter's internal-external scale

Abstract

The forced choice Rotter I-E scale and a scale that involved rating percentage agreement with the separate item alternatives both were completed by 294 undergraduates. Disagreement between the two scale formats was greater when pairs of alternatives were seen as more equal. Disagreement also was related to self-rated level of conflict when Ss were responding to the forced choice item. Correlations between responses to pairs of separate item alternatives were low, which replicates previous studies. However, correlation between the forced choice I-E score and an I-E score derived from ratings of the separate alternatives w&s high, a replication of other previous work. These results were integrated by showing support for a model of locus of control as comprised of independent external and internal subscales, in which forced choice I-E represents a balance score.The forced choice format of the Rotter I-E Scale has been identified as a source of phenomenological conflict (Sanger & Alker, 1972), as a safeguard against response bias (Rotter, 1966), as an approximation of the type of decisions required by the real world (Tyler, Gatz & Keenan, Note l), and as a potential source of inconsistent factor analytic results (Kleiber, Veldman, & Nenaker, 1973).Studies that assessed the forced-choice format (Collins, 1974;Kleiber et al., 1973;Klockars & Varnum, 1975) have yielded some comparable and some dissimilar results. Iileiber et al. (1973) and Klockars and Varnum (1975) questioned whether pairs of alternatives are bipolar. They reasoned that if a pair of internal and external statements from a given item are opposites, then (a) they should be significantly negatively correlated; and (b) both should load on the same factor, with opposite signs. Neither conclusion was supported. Rather, their results suggested that on the majority of items the respondent is choosing between different subscales as well as between internal and external attributions of control. The average item pair correlation reported by , and the range was +.14 to -.47. Item-by-item correlations reported by Klockers and Varnum were very similar. Their average was -.15; their range was +.06 to -.46, and the same two items represented the ends of the range.Collins (1974) also factored the 46 separate statements and found that the alternatives tended to load on different factors. He took the further step of testing whether answers to the 46-item Likert format approximated scores obtained from the 23-item forced choice format. If so, then factor analysis results could be generalized to the forced-choice I-E scale. He scored both formats for externality and found a correlation of .82 between them, which suggests that the two were measuring the same construct.The present study was designed to replicate, resolve, and extend these previous studies, as well as to explore the implications of the results for use of the I-E scale. METHOD Subjects and ProcedureThe Rotter I-E scale was presented in two different formats to 294 students (75 male, 219 female; 260 white, 34 ...

A constructivist analysis of the Rotter I‐E scale1

“…In fact there have been some studies which help us to verify that men and women do organize their senses of locus of control differently. Gatz and Good (1978) have found that men and women are equally external but that men are more internal. Sanger and AUcer (1972), in studying internality and feminism, suggest that women endorse internality to the extent that they are socialized by the Protestant ethic ideology to believe that hard work pays off, even though they have not found it to hold for them.…”

“…Several researchers (Kleiber, Veldman, & Menaker, 1973;GoUins, 1974;Klockars & Vamum, 1975;Gatz & Good, 1978) have pointed out that one reason for the uninterpretabUity of factors based on the forced choice I-E items is that the item alternatives are not symmetrical. Thus, a response involves more than a bipolar decision.…”

Individual differences matter : personality traits and preferences in smallholders’ farm management

“…We elicited LoC following the standard practice of Rotter (1966) using the 10-item inventory with a 5-point Likert-type scale. We followed the suggestions by Collins (1974) and Gatz and Good (1978) that external and internal LoC should be treated as two independent traits, and constructed two separate LoC traits (i.e., internal and external)…”

“…Some studies, such asCollins (1974),Gatz and Good (1978) have suggested that external and internal locus of control should be treated as two independent traits. We follow the same practice in our study.…”