Directorates
·	Director General's Office
·	Data Development
·	Strategic Evaluations
·	Partnership and Continuous Evaluations
Publications
·	Evaluation Reports
·	Evaluation Briefs
·	Labour Market Development Agreement Series
·	Lessons Learned Series
·	Evaluation Tool Kit Series
·	Evaluation Forum Newsletter
·	From Theory To Practice

Client impacts of jobLink were examined from two perspectives: clients' subjective ratings of the benefits of the resource centre; and clients' objective outcomes such as their current employment status and non-labour market outcomes (e.g., attitudes and satisfaction with other aspects of their life). Both subjective and objective outcomes are compared with the results of non-participants. These findings are presented below.

5.1 Rated Benefits

Exhibit 5.1 presents respondents' responses in terms of the perceived benefits of the jobLink resource centre services and the corresponding ratings of the benefits of employment services received by the comparison group. In general, neither the participants nor the comparison group provided high ratings of the benefits of the employment services they received, though the ratings of the participants tend to be marginally higher. The highest rated benefit for both groups was helping individuals to get information about the job market. Participants provided a significantly higher rating of this benefit with 58% indicating they had benefited "to a great extent" compared to 47% of the comparison group.

Just under one-third of participants believed that the jobLink resource centre had helped them to find a job (30%) and to decide to upgrade their education (31%). In terms of the latter, the comparison group provided a significantly higher rating of the benefits of their employment services than jobLink participants.

Both jobLink participants and the comparison group provided lower ratings of services in terms of helping them to get off social assistance or employment insurance and helping them to enroll in another employment program. Only a smaller proportion of respondents — 16% in both groups — believed that the employment services they received had helped them to increase their wages.

jobLink clients who had received the more intensive staff-assisted services rated the benefits more highly on most of the aspects listed. Not surprisingly, jobLink participants who were employed at the time of the survey also provided higher ratings of the benefits of the resource centre in terms of helping them to find a job and helping them to get off social assistance or employment insurance.

5.2 Employment Status

The following sections present the results of the bivariate survey data analysis of program outcomes (including labour market outcomes such as employment and use of income transfers and non-labour market outcomes such as satisfaction). Note that these data do not control for pre-existing differences between the participant and comparison groups. The results of the multivariate analysis, which controls for these variables, are presented in the final section of this report.

jobLink participants were significantly more likely to have looked for work than the comparison group (91% and 75%) (Exhibit 5.2). The most frequently used job search methods for both groups were sending out résumés or applications and answering ads. The comparison group was somewhat more likely than jobLink participants to make use of informal job search methods such as through family and friends.

Since starting to use the jobLink resource centre, 42% of clients had found a full-time or part-time job or worked at their own business. This is significantly higher than the 30% of the comparison group who had found employment during the same period.

At the time of the interview, 22% of jobLink participants were employed full-time, 14% were employed part-time and 2% were self-employed. Based on the bivariate results, the participant group was significantly more likely to be employed full-time than the comparison group. Similar proportions, 44% of clients and 41% of the comparison group, were unemployed and looking for work at the time of the interview. About one in five clients were out of the labour market, either in school, a homemaker or not looking for work. The "out of the labour market" individuals are over-represented among the comparison group. Of those who were not currently working, the majority of respondents in both groups were interested in entering the labour market in the next 12 months.

5.3 Current Job

Of respondents who reported having a job, the jobs that were found by the jobLink participant group were more likely to be full-time and year-round (though not statistically significant). The earnings for the participant group were also higher (though this is at least a partial reflection of the greater number of hours worked). Both the participants and the comparison group were concentrated in sales, service, and clerical occupations (Exhibit 5.3). jobLink participants spent significantly fewer weeks unemployed during the post-program period than the comparison group.

Within the jobLink group, the job characteristics were predictably different along gender lines with women more likely to be in sales and service and clerical occupations compared with men who were over-represented in trades, transport and equipment operator positions. Women were also more likely than men to be working part-time and also had lower average weekly earnings as a consequence. Weekly employment earnings were highest for clients in the middle age category (31 to 45) and for clients who had a full-time job prior to using the resource centre.

5.4 Income and Utilization of Public Support

jobLink participants were significantly more likely to have collected employment insurance in the post-program period than the comparison group in the bivariate analysis (Exhibit 5.4). Both the participant and comparison groups were likely to have accessed social assistance in the post-program period, though there was no statistically significant difference between the two groups. jobLink participants, however, did collect social assistance for a shorter period of time compared with non-participants.

There were few differences between the participant and comparison groups in terms of household and personal income. jobLink clients' current household income was $16,870 on average compared with $16,980 for the comparison group. Personal income was $11,080 for jobLink clients on average, compared with $11,353 for the comparison group. Almost two-thirds of clients (65%) reported that it was difficult for them to live on their total household income. One in five said living on their current household income was somewhat difficult and 12% responded not difficult.

Income levels were higher for clients who were currently employed and who had stronger prior work experience. Older clients (over 45 years of age) reported lower household incomes and also expressed greater difficulty in living on their current income.

5.5 Non-Labour Market Outcomes

A limited number of measures were included on the survey instrument to capture other possible outcomes not related to employability. Exhibit 5.5 shows satisfaction levels at the time of the interview on dimensions such as skills and security. Overall, the jobLink resource centre clients have a moderate degree of confidence in their skills — 65% are satisfied with their current business or job skills. Just under half of clients are satisfied with the level of control they feel they have over their life. Not surprisingly, satisfaction with overall financial security is rated lowest on this battery with only one in five feeling satisfied with this aspect of their life. There were no significant differences between the responses of the participant and comparison groups.

Clients who are currently employed express greater satisfaction on each of these indicators. Satisfaction with level of financial security decreases with the age of the respondent but increases with respondent's personal income. Satisfaction with skills is associated with higher levels of education.

Exhibit 5.6 presents the results of a second series of attitudinal indicators. Most clients and those in the comparison group are confident in their abilities to search for work (74% and 71% respectively agree they know how to find a job). There is, however, pessimism around actually finding a job. About one in three clients agreed that they would have no problem in picking up and finding a good new job. The comparison group expressed greater confidence on this item. The responses indicate that most respondents attribute this lack of flexibility to local labour market conditions. Only one in four jobLink clients and a similar proportion in the comparison group agreed that there are a lot of jobs available in their community for a person with skills like theirs. Those with lower personal incomes are more likely to feel that they would have trouble finding a good new job. Clients in the older age category have the greatest degree of pessimism around demand for the skills in their community.

5.6 Multivariate Analysis

An essential question in the analysis is the incremental impact of jobLink on participants. Simple comparisons between program participants and non-participants on key outcome indicators (e.g., employment status) may yield a biased estimate of program impact because of pre-existing differences between the comparison groups and the jobLink participants. To ensure that differences in measured outcomes were not the effect of pre-existing differences between the jobLink and the comparison groups such as differences in terms of their labour market history and background characteristics, multivariate analyses were conducted.

Six dependent variables representing key outcome variables were tested:

• whether respondents were employed or not at the time of the interview (discrete);

• whether respondents were full-time employed or not at the time of the interview (discrete);

• number of weeks unemployed during the post-program period (continuous);

• number of weeks on EI during the post-program period (continuous);

• personal satisfaction (continuous, represented by an index of respondents' rated satisfaction with their financial security, business/work skills and level of control over their life);

• confidence in the labour market (continuous, presented by respondents' rating to the item "I would have no problem picking up and finding a good new job").

A broad number of independent variables were included in the analysis to assess antecedent conditions such as sociodemographic characteristics, prior labour market experience and type of intervention. Specifically, variables that were included in the analysis are listed below. Unless otherwise stated, the source of the data is the survey conducted as part of the evaluation.

• age;

• sex;

• education (no post-secondary education/post-secondary education);

• marital status (married or not);

• number of dependants;

• site;

• ever full-time employed pre-program;

• number of weeks employed pre-program;

• gross earnings three years prior (source: HRDC administrative data);

• EI benefits three years prior (source: HRDC administrative data);

• social assistance benefits three years prior (source: MCSS administrative data);

• whether actively looked for work or not;

• whether used self-assisted services or not;

• whether used staff-assisted services or not;

• employed or not prior to using resource centre/reference date.

A flag representing participant versus comparison groups was included among the independent variables to measure the impact of the jobLink intervention compared with the outcomes for the non-participant comparison group. The coefficients for the group variable measures the difference in the dependent variables between the participant and comparison groups, controlling for all measured pre-existing differences. Dummy coding was used for all categorical variables.

Multiple regression was utilized for the continuous dependent variables while logistic regression was used to test the categorical variables. The modelling process was conducted in two stages. First, a base model was developed to control for the basic sociodemographic and labour market characteristics of respondents, such as age, sex and education. The jobLink program variable was also included. The coefficient and significance level of the program variable in the base model would indicate the impact of the program on the outcome variables after controlling the background characteristics of the respondents. The full model includes a more comprehensive set of independent variables, specifically, the use of self-directed and staff-assisted services. The purpose of this model is to better understand the effect of the use of different types of services.

The coefficient and significance level for the program variable for each of the independent variables tested in the base and full models are presented in Exhibit 5.7. The overall tolerance level for each of the models is also included as a measure of collinearity. The detailed regression results are reproduced in a separate technical appendix.

View Exhibit 5.7

Briefly, the multivariate analysis provides some indication of the potential impacts of jobLink, though these impacts appear to be relatively short term. There is a statistically significant reduction in the time not working following the program intervention for the jobLink participants — eight weeks. This also translates into a significant difference between the participant and comparison groups in terms of amount of time collecting social assistance during the post-program period. jobLink participants collected social assistance 2.8 months less during the post-program period compared with non-participants.

The remaining indicators such as employment status and weeks on EI were not significant in the multivariate analysis. The latter finding — weeks on EI — is perhaps not surprising given that respondents would have had to find and lose a job (in significant numbers) during the post-program period in order to collect EI. The analysis, however, also may not have been sensitive enough to capture other impacts given the relatively small number of cases available, as well as other challenges around constructing the participant and comparison group sampling frames. Employment during the post-program period was influenced by respondents' sex (men had more positive outcomes than women), age (younger respondents were more likely to have found work), marital status (married respondents were more likely to be working), whether the individual had looked for work and level of prior earnings (those with higher earnings were more likely to be employed).

The multivariate analysis of the non-labour market outcomes suggests that there is no significant difference between the participant and comparison groups with respect to personal satisfaction. The comparison group, however, reported higher levels of confidence in the labour market.

5.7 Summary

Perceived ratings of the benefits of adjustment and employment services were modest for both the jobLink participant and comparison groups. While most respondents looked for work (higher among the jobLink group), re-employment rates were generally low; less than half of both groups were employed at the time of the survey. A significant positive advantage for the participant group in terms of employment status faded in the multivariate analysis, which indicated that pre-existing differences between the two groups accounted for the difference. Sex, age, marital status and prior labour market experiences were more important in predicting outcomes. A positive advantage for the jobLink group in terms of weeks not working in the post-program period was maintained in the multivariate analysis (though this should be interpreted cautiously given the difference in the timing of the surveys for the two groups). The jobLink group spent significantly less time not working (8 weeks) during the post-program period than the comparison group and this translated into less reliance on social assistance (2.8 months). The jobs that were found by the jobLink participants were more likely to be year-round and full-time with higher earnings than the comparison group. There were few differences between jobLink participants and the comparison group in terms of non-labour market outcomes.