Individuals were randomly assigned to DPA or IPS. Afterwards they were followed for two years. Objective data related to paid employment outcomes was collected through quarterly participant interviews. Data pertaining to the predictor variable, working alliance were collected by participant interviews every 6 months for individuals who were working at the time.
Demographic variables were measured by the Uniform Client Data Inventory and work history by the Employment and Income Review at baseline. Psychiatric symptoms were assessed at baseline and semiannual periods using the Positive and Negative Symptom Scale. Employment outcomes included total duration of paid work and mean paid job tenure. A scale was developed to measure working alliance. It included items related to emotional support, instrumental/informational support, frequency of performance feedback, stressfulness of the relationship, how critical the vocational worker was to the client and the person's overall satisfaction with the relationship.
Fidelity to the IPS or DPA model tenants were assessed every 6 months for both sites where individuals received services.
Stepwise hierarchical multiple regression analysis were run to determine relationships between the working alliance and employment outcomes. Baseline variables were controlled for and were entered into the regression model. Related to the number of days of paid work and job tenure applicable covariates were entered.
The working alliance variable was added in the second step of the regression analysis. Zero order correlates were also run. Independent t test was used to determine differences between IPS and DPA on working alliance.