measuring expectancies that might predict health behavior so a more

idiosyncratic scale was necessary. In this study I used the Wallston

Scale to control for locus of control.

 The scale consists of 11 items chosen from a 34-item pool following

appropriate statistical protocols. Alpha reliability of the 11-item

scale is .72. In developing the scale, the authors controlled social

desirability using the Marlow-Crowne Social Desirability Scale. The

correlation was -.01. Test-retest reliability on the Wallston Health

Locus of Control Scale was .91.

 "Concurrent validity of the Health Locus of Control Scale (HLC) was

evidenced by a .33 correlation (p < .01) with the Rotter I-E Scale for

the original sample" (Wallston, Wallston, Kaplan, & Maides, 1976, p.

581). The authors reported two studies that suggest evidence of con-

struct validity. They predicted and found that internalss" with high

health values would seek more information about a given health situation

and found it to be true. The second study was based on the belief that

internalss would be more likely to take steps to better their

environmental condition than externals" (p. 583). This was a weight-

reduction program, and the results were as predicted. Internals lost

more weight on a "self-directed program and externals lost more on a

group program" (Wallston et al., 1976, p. 583).

 In general, although the whole concept of locus of control is

currently under study, I believe that the area-specific assessment

instrument has met the criteria for inclusion in this study.

Controlling for this variable in any study on health behavior seems

important.