Cocaine: Self-Efficacy

Listed below are a number of situations that lead some people to drink. Enter the numbers in the boxes that best corresponds to your present feelings of temptation and confidence in each of the situations below.

	When I am feeling angry inside
	When I am feeling depressed
	When I see others drinking alcohol
	When I am craving a drink
	When I am feeling really good
	When I have the urge to try just one drink
	When I am bored
	When I am worried about something
	When I think I have overcome my problems with alcohol
	When I want to test my willpower
	When I am celebrating a special occasion
	When I am lonely
	When I feel a physical need for alcohol
	When things are going really well for me
	When other people encourage me to have a drink
	When I see an ad about alcohol
	When I become overconfident about my sobriety
	When I am offered a drink by someone
	When I am passing a bar
	When I am with friends I used to drink with
	When I am feeling really positive about the way things are going for me
	When I am nervous
	When I feel like having a good time
	When I go by a liquor store
	When I have a strong urge for a drink
	When I think I can drink again
	When I am in a situation I used to drink
	When I am really happy
	When I want to see how far I can push myself
	When it's late and an urge comes over me

Self-Efficacy Scale
The self-efficacy questionnaire consisted of 30 items generated to measure specific situational confidence (to abstain from alcohol use) and temptation (to engage in alcohol use) factors. Items were written to represent five constructs, including negative affective situations, positive/social situations, craving or habit items, situational cues (environmental), and testing personal control. The format for such a factor structure was taken from work previously done in the smoking field (Velicer et al., 1990), and from interviews of replapse situations proposed generally for addictive behaviors (e.g., Marlatt & Gordon, 1985).

The resultant 30x30 item matrices (Confidence and Temptation) were subjected to principal components analysis using the MAP procedure. As previously stated, this program offers a range of choices in the number of components to retain. Scree plots, MAP (Velicer, 1976) and Horn's rule (Horn, 1965) were utilized to determine the number of components to retain.

Due to the nature of the sample (mean length of sobriety at assessment = 5.93 years), both of these scales exhibited a considerable restriction of range in scores, as the mean of the temptation items was quite low (X = 2.06 ; SD = 0.93), while confidence was uniformly high across theoretically problematic situations (X = 4.35 ; SD = 0.75). Once again, the similarity pattern across various addictive behaviors is noted, as this closely represents what has been found in large groups of smokers, despite the fact that the item content has markedly differed between the samples (e.g., Prochaska & DiClemente, 1984; Rossi et al., 1989).

Recent investigations on self-efficacy in the smoking area have suggested that a restriction in range may not represent a response bias, but may reflect an accurate assessment of self-efficacy as a function of stage of change This type of response pattern (with high inter-item correlations) has been identified as one in which an obfuscation of distinct factors occurs, which might lead to an interpretation of a single, general, underlying factor to these scales (Velicer & Steiger, 1990). Like the processes of change, subjects were included in the final analysis only if complete responses were present for all items. This resulted in an effective sample of 150 for the temptation items and 163 for the confidence scales (listwise N = 143). These are considerably smaller than the process scales (listwise N = 175) and possibly represents a pitfall in conducting research on long-term maintenance. While process use represents what individuals are currently doing, subjective ratings of efficacy with this group of maintainers may be a less salient domain to measure for these individuals, although this is speculative.

Confidence
Within the Confidence scale, the range of factors was from 1 (scree plot) to 2 components suggested by both MAP and parallel analysis. Both 1 and 2 component PCA's were performed on the matrix, and the resultant solutions examined for maximum clarity and interpretability. From these analysis, it was clear that a single, general factor emerged as the most interpretable of the solutions. All 30 of the confidence items leaded at .53 or higher on the single general factor, with a mean loading of .70. An estimate of internal consistency (coefficient alpha) was .97 for the 30 item scale. Table 9 provides the means, standard deviations, reliability estimates, and intercorrelations between the temptation and confidence scales.

Martin, R.A., Rossi, J.S., Rosenbloom, D., Monti, P.M., & Rohsenow, D.J. (1992, November). Stages and processes of change for quitting cocaine. Paper presented at the 26th annual meeting of the Association for the Advancement of Behavior Therapy, Boston, MA.

Rosenbloom, D. (1991). A Transtheoretical Analysis of Change Among Cocaine Users. Unpublished dissertation, University of Rhode Island, Kingston, RI.

Rossi, J.S. (1992, August). Common processes of change across nine problem behaviors. Paper presented at the 100th annual convention of the American Psychological Association, Washington, DC.