We examined 2 mechanisms of change under paradoxical interventions: reactance and increased sense of self-efficacy. Procrastinating college students were randomly assigned to either paradoxical or self-control interventions. Effective study time and perceived self-efficacy were measured before and after treatment. In Study 1 nonverbal measure of initial reactance was employed. In Study 2 reactance was experimentally manipulated. Under paradoxical interventions, Ss higher on initial reactance benefited more from therapy than did Ss with low reactance; nonreactant Ss did not increase their effective study time, but they did improve in perceived efficacy to control their procrastination; increased study time was negatively correlated with increased self-efficacy. In self-control treatment, increased study time was accompanied by increased self-efficacy. Paradoxical interventions seem to reduce procrastination through the mechanism of reactance in some clients, whereas in others they lead to a cognitive change, possibly mediating a subsequent behavior change. The term paradoxical interventions refers to a wide variety of therapeutic injunctions and directives, the common denominator of which is essentially their attempt to induce change by discouraging it (Seltzer, 1986). Although there are enthusiastic clinical reports attesting to the efficacy of all types of paradoxical interventions, empirical evidence from controlled studies indicates that there are vast differences in the efficacy of different types of paradoxical interventions (see Shoham-Salomon & Rosenthal. 1987, for a review). Moreover, the differential psychological factors that mediate the effects of paradoxical interventions were rarely investigated in a controlled way. In one of the first attempts to understand how paradoxical interventions operate, Watzlawick, Beavin, and Jackson (1967) explicated the logic of what they called the therapeutic double bind. It was seen as a mirror image of the pathogenic double bind: The directive to deliberately engage in the symptomatic behavior, the very same behavior the client wishes to change, is similar to a "be spontaneous" paradox. The only way to obey such a directive is by disobeying it. According to Watzlawick et al. (1967), two possible consequences follow a paradoxical directive (e.g., to bring about one's anxiety). In many cases the client is not able to produce such a spontaneous occurrence on demand. In each attempt to bring the symptom about, he or she will experience reduced symptomatology. Such a success is expected to lead to a sense of controllability, a feeling of mastery We thank Hal Arkowitz and the three anonymous reviewers for their insightful comments on an earlier version of this article. We also thank Anat Goren-Falach and Eyal Dahari for their tremendous amount of help at all stages of this research.

This study evaluated the use of pretherapy patient variables as correlates of 3 categorical types of outcome: negative response (negative change of more than 1 normative SEest on depression measure); nonresponse (change within +/- 1 SEest on depression measure); and positive response (positive change of more than 1 SEest on depression measure) to psychotherapy among 62 patients with major depressive disorder. By using 4 scales from the Brief Symptom Inventory, the Inventory of Interpersonal Problems, age, and sex, 75.8% of the Ss were correctly classified into the 3 groups. Negative responders were characterized by high levels of interpersonal difficulty and low levels of subjective distress. Nonresponders displayed moderate levels of both interpersonal difficulties and subjective distress. Positive responders displayed high levels of both interpersonal difficulties and subjective distress.

The effectiveness of paradoxical interventions in psychotherapy was evaluated in a meta-analysis of 12 data sets. Overall, paradoxical interventions were as effective as (but not more effective than) the typical treatment mode. However, paradoxical interventions showed relatively greater effectiveness than other interventions (a) 1 month after treatment termination and (b) with more severe cases.Finally, the analyses point to the advantage of positive connotations but raise serious doubts concerning the effectiveness of symptom prescription that does not follow a positive connotation. We offer specific research questions pertaining to the factors that mediate the differential effectiveness of paradoxical interventions.

Given the contextual boundedness of process variables, 3 assumptions implicit in present process research are questioned: Process variables (a) have fixed meanings, (b) discretely contribute to outcomes, and (c) have a decontextualized net worth, evaluated by their correlations with outcomes. From a systemic perspective, a conceptual framework is suggested that integrates both theory-driven and discovery-oriented strategies into a cyclical, ecologically oriented model. Experiments, observations of change events, and detailed descriptions partake in this cyclical model, informing and enriching each other. The core of this integration is a transactional conception that sees change processes as mutually defining rather than interacting with each other, thus leading to holistic examinations of changes of pattern in addition to patterns of change. The transactional core serves as the point of departure and home base for both theory-driven and discovery-oriented research strategies.

No abstract