Chapter 2: Participation in Employment-Related Activities After Random Assignment

This chapter analyzes the use of program services by sample members in the Jobs-First GAIN Evaluation during the two years after random assignment. First, the chapter explores the extent to which AFDC-FG experimental group members were exposed to the mix of services and messages offered by Jobs-First GAIN. As discussed in Chapter 1, the Jobs-First GAIN program ran high-quality job clubs supported by job development activities, as well as short-term education and training for people who completed job club without finding employment. These activities made up a key component of the Jobs-First GAIN approach to promoting rapid entry into the labor market. This chapter also examines the extent to which program staff used formal enforcement procedures, especially financial sanctions (reductions in welfare grants), to enforce mandatory participation requirements. Further, it compares levels of participation for subgroups of the experimental group.

The chapter then examines the extent to which AFDC-FG experimental group members participated in employment-related activities outside Jobs-First GAIN. Such participation could have occurred when experimental group members were no longer required to participate in the program or had exited from welfare, or when they were still in Jobs-First GAIN as an addition to or instead of participation in their regularly assigned activities. Next, the chapter compares participation patterns for AFDC-FG experimental group members with those for members of the control group to estimate the effects of Jobs-First GAIN on use of employment-related services. The chapter concludes with a brief summary of participation patterns for AFDC-U experimental group members.

1. Key Findings

• All experimental group members attended Jobs-First GAIN’s six-hour informational and motivational orientation session and were exposed to the program’s strong pro-work message.

• Relatively few experimental group members — 42 percent of AFDC-FGs and 34 percent of AFDC-Us — participated in a subsequent Jobs-First GAIN activity during the two years after orientation. Of those who participated, nearly all attended job club, reflecting the Work First character of the program. Nearly all participants began attending activities soon after orientation and finished participating by the end of year 1.

• About 30 percent of AFDC-FG experimental group members experienced a reduction in their welfare grant (a sanction) for noncompliance with Jobs-First GAIN’s mandatory participation requirements at least once within two years of orientation. This sanction rate was high compared to the sanction rates recorded for other employment-focused programs in the 1990s and well above those recorded for welfare-to-work programs evaluated in the 1980s. Jobs-First GAIN case managers cited most other AFDC-FG and AFDC-U experimental group members for noncompliance and initiated formal enforcement procedures on them at some point during the follow-up period. The process, however, did not lead to imposition of financial sanctions for these people.

• A relatively large proportion of AFDC-FG experimental group members attended education and training activities outside Jobs-First GAIN, boosting their rate of participation in any type of employment-related activity other than Jobs-First GAIN orientation to 62 percent.

• About 44 percent of AFDC-FG control group members also participated in employment-related activities, mostly education and training, on their own initiative. As a result of this unusually high participation rate for control group members, the experimental-control difference in participation was modest (about 18 percentage points).

• Jobs-First GAIN produced a much larger impact (31 percentage points) on participation in job search. In effect, Jobs-First GAIN induced many experimental group members to participate in job search who would otherwise have participated in education and training.

2. Framework for Interpreting Participation Findings

1. What Can Be Learned from Studying Participation

The discussion of participation patterns for sample members presented in this chapter will provide important context for interpreting the findings on program costs, impacts, and benefits in the chapters that follow. The first task in the participation analysis is to measure the extent of experimental group members’ involvement in Jobs-First GAIN. When the evaluation was designed, it was expected that restricting access to Jobs-First GAIN’s services, messages, and mandates to the experimental group would enable its members to attain a higher rate of employment, higher earnings, and lower levels of welfare receipt than their counterparts in the control group. Whether these effects actually occur depends on many factors. A key question addressed by this chapter is whether experimental group members’ exposure to the program was sufficient to change their labor market behavior. Put differently, for this to be a fair test of the program’s Work First strategy, a relatively high percentage of experimental group members must have participated in Jobs-First GAIN activities, encountered its mandates, or received its messages.

Participation in Jobs-First GAIN was not the only way in which experimental group members could prepare for work. Some may have participated in other types of activities on their own initiative, probably after they were no longer required to participate in Jobs-First GAIN. These experiences outside the program could also have affected the timing of experimental group members’ search for employment, success in finding and keeping jobs, and levels of income and self-sufficiency. It is therefore important to learn how often experimental group members participated in employment-related activities outside Jobs-First GAIN.

Further, previous experimental evaluations of welfare-to-work programs have shown that many control group members enroll in employment-related activities – typically, basic education, vocational training, or post-secondary education – on their own initiative.1 Jobs-First GAIN's effects on employment and welfare will likely fall short of expectations if control group members received similar types of pre-employment services as experimental group members or took part in other activities that enhanced their ability to find and keep jobs. For this reason, it is important to measure control group members’ levels of participation and to estimate Jobs-First GAIN’s incremental effects on participation – that is, the extent to which the program raised experimental group members’ level of participation (overall and in particular activities) above that of control group members.

The same reasoning applies to the study of Jobs-First GAIN’s cost-effectiveness. Federal, state, and county government budgets not only funded Jobs-First GAIN services, but also many of the activities in which experimental and control group members enrolled on their own initiative – through the federal Job Training Partnership Act (JTPA), community colleges, Pell grants, and state and county education programs. Thus, the total cost to government budgets of helping Jobs-First GAIN experimental group members prepare for work depends on experimental group members’ level of participation in pre-employment activities both within and outside the program. Moreover, estimating the experimental-control difference in pre-employment costs requires knowledge of participation levels for control group members.

2. Alternative Definitions of Participation

This chapter follows the analytical framework used in previous MDRC studies of participation patterns in welfare-to-work programs. It defines participation as attendance at an employment-related activity for at least one day, whether within or outside Jobs-First GAIN, but does not count program orientations, appraisals, or other meetings with Jobs-First GAIN staff. This definition of participation assumes that program enrollees who take part in activities such as job clubs or education and training courses receive the most exposure to the program “treatment.” For Jobs-First GAIN, however, the distinction between attendance at a program activity and a meeting with program staff is not clear-cut. As discussed in the previous chapter, all experimental group members attended a long informational and motivational meeting at orientation during which program staff strongly communicated the program’s Work First message. In addition, experimental group members could receive job leads from program staff during orientation, appraisal, or at any time afterward. Thus, by a more inclusive definition of a program activity, one could conclude that 100 percent of experimental group members participated.

3. Employment-Related Services Are Only Part of the “Treatment”

Participation in employment-related activities was only one of several ways in which program enrollees could experience the Jobs-First GAIN “treatment.” Equally important, program administrators and staff communicated a strong Work First message to all people entering the program. Enrollees first heard this message during a long motivational session at program orientation. Program staff repeated these ideas during appraisals and other one-on-one meetings, during program activities such as job club, and in informational handouts.

Further, as discussed below, DPSS implemented a tough enforcement-oriented response to experimental group members who did not participate in program activities without showing good cause. Jobs-First GAIN staff frequently issued warnings to experimental group members that they were not complying with the program’s mandatory participation requirements and imposed financial sanctions (welfare grant reductions) on a relatively large proportion of the caseload at least once within two years of orientation. Program administrators intended this “high-enforcement” case management approach and the strong pro-employment message to complement the program’s high-quality, motivational job clubs. Together, these components of the Jobs-First GAIN approach encouraged enrollees to find work quickly and discouraged them from spending a long time in the program.

4. Measurement Issues

Sample members’ participation in Jobs-First GAIN and non-Jobs-First GAIN activities was measured over a two-year follow-up period, starting with their date of random assignment. (For experimental group members this date is the date of orientation.) Automated program tracking records from DPSS’s GAIN Employment Activity Reporting System (GEARS), combined with automated welfare and earnings records, were used to estimate rate and length of participation in Jobs-First GAIN activities. The same measures of participation in non-Jobs-First GAIN activities were calculated from the responses of a sample of 746 AFDC-FG experimental and control group members to the Two-Year Client Survey. Survey respondents reported the start and end months and average number of hours per week of participation. Like any self-reported data, these participation data are subject to recall error. That is, some respondents may not have reported participation in all activities in which they in fact participated, especially short-term activities, whereas others may have reported participation in activities that occurred prior to random assignment.2 Some respondents may also have underestimated or overestimated the duration of their attendance or the average number of hours per week of participation.

Some measures presented in this chapter apply only to sample members who participated in an employment-related activity (that is, job club, education, or training). These people will be referred to as participants. The findings for AFDC-FGs represent weighted averages of the participation levels estimated for regular and early enrollees.3

3. Results for AFDC-FGs

1. Participation by Experimental Group Members in Jobs-First GAIN Activities

As shown in Table 2.1, in the two-year follow-up period, 58 percent of AFDC-FG experimental group members were assigned to a Jobs-First GAIN job search, education, or training activity, and 42 percent participated in a Jobs-First GAIN activity for at least one day. The level of participation in Jobs-First GAIN activities for experimental group members is low compared to the participation levels recorded for other employment-focused welfare-to-work programs evaluated by MDRC, including Riverside GAIN (60 percent). The Riverside Labor Force Attachment (LFA) program, however, had a similar proportion of experimental group members (44 percent) who participated in work-related activities.4 Jobs-First GAIN also attained a lower level of participation than the earlier, education-focused Los Angeles GAIN program (51 percent).5

On average, AFDC-FG participants remained in Jobs-First GAIN activities for a relatively short time — a little more than four out of the 24 months of the follow-up period (see Table 2.2). About 60 percent attended for two months or less. Nearly all participants began attending activities soon after orientation and finished participating by the end of year 1. At the high end, 8 percent participated for at least a year; a slightly higher percentage were still attending a Jobs-First GAIN activity at the end of the follow-up period. The participation patterns displayed in Table 2.1 reflect the Work First orientation of Jobs-First GAIN. Out of all activities, job search (usually job club) drew the largest percentage of experimental group members (37 percent). In contrast, only 11 percent of experimental group members participated in basic education or vocational training.6 DPSS did not assign any Jobs-First GAIN enrol-lees to post-secondary education courses.

The data in Table 2.2 provide additional information on use of program services for the 42 percent of experimental group members who attended a Jobs-First GAIN activity. As expected of a Work First program, nearly 90 percent of AFDC-FG participants attended job club in the two-year follow-up period, whereas only about a quarter of them participated in education or training. A little more than three-fourths of job club participants attended only one “spell” of job club (that is, one block of consecutive weeks), and nearly as many took part in job club as their only activity in the program.

Los Angeles Jobs-First GAIN Evaluation

Table 2.1

Rates of Participation and Status Within Two Years of Orientation for AFDC-FG Experimental Group Members

Participation Status (%)			All	Regular Enrollees	Early Enrollees
Assigned to any activity a			58.3	54.7	72.4
Ever participated in b	Any activity c		41.9	37.9	57.0
	Job search		37.1	33.8	49.6
	Any education or training		10.6	8.8	17.5
	Basic education		4.3	3.2	8.3
		ESL	1.2	1.0	2.2
		ABE	1.1	0.9	1.8
		GED	1.6	1.2	3.3
		High school	0.5	0.3	1.2
	Vocational training		7.6	6.6	11.6
	Work experience		2.8	2.3	4.6
	OJT		0.0	0.0	0.1
	Assessment		9.6	8.4	14.0
Deregistered for any reason d			93.9	93.3	95.9
	For employment		48.3	47.4	51.9
	For sanction		39.1	39.0	39.5
	For other reason		46.3	46.5	45.4
	In conciliation		83.0	83.1	82.6
	Sanctioned		30.4	30.9	28.5
	Deferred for any reason		30.5	31.6	26.4
	For unapproved SIT e		8.5	9.1	6.3
Sample size			11,521	8,620	2,901

SOURCE: MDRC calculations using data from the GAIN Employment Activity and Reporting System (GEARS). NOTES: a The assignment rate includes assignment to all activities listed, except assessment.(back) b Activity participation rates include participation for at least one day in either a program-referred or approved self-initiated activity.(back) c "Any activity" includes all activities listed, except assessment.(back) d Subgroup percentages sum to more than the total deregistration percentage because some recipients were deregistered more than once during the follow-up period.(back) e An SIT is a self-initiated activity ("self-initiated training"). (back)




Los Angeles Jobs-First GAIN Evaluation

Table 2.2

Participation Patterns Within Two Years of Orientation for AFDC-FG Experimental Group Members Who Participated in Jobs-First GAIN Activities

Outcome
Participated in job search (%)		88.5
	One spell	68.0
	Two or more spells	20.5
Participated in (%) a	Job search only	74.2
	Education and training only	10.8
	Job search and education and training	14.4
Average number of months in which individuals participated in a Jobs-First GAIN activity b		4.2
Number of months in which there was participation (%) c	1	32.5
	2	27.0
	3	8.7
	4 to 6	11.3
	7 to 12	11.9
	13 or more	8.3
Still participating at end of year 2 (%)		9.0
Sample size		4,509

SOURCE: MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS) NOTES: Full sample mean and percentages are weighted averages of results for regular and early enrollees. Measure = (regular enrollee result x percent of experimental and control group regular enrollees in AFDC-FG sample) + (early enrollee result x percent of experimental and control group early enrollees in AFDC-FG sample). a The percentage of participants who participated only in work experience or OJT is not shown in the table.(back) b Participants for whom data are missing were excluded from the calculation of the mean. (back)




2. Granting Temporary Deferrals and Longer-Term Deregistrations from Mandatory Participation Requirements

Almost every experimental group member experienced the transition to nonmandatory status, that is, was deregistered from the program, at least once before the end of year 2. As shown in Table 2.1, 94 percent of experimental group members were deregistered within two years of random assignment, about half of them because they were working full time, that is, for at least 30 per hours per week.7

Under California regulations effective prior to April 1, 1998, welfare recipients with certain barriers to participation were temporarily excused, or deferred, from Jobs-First GAIN’s participation requirements. Common reasons for granting deferrals included medically verified illness and “severe family crisis.” As shown in Table 2.1, about 31 percent of AFDC-FGs were deferred from participation in Jobs-First GAIN for good cause at some point during follow-up.

Notably, about 9 percent of experimental group members received a deferral for an unapproved self-initiated activity, usually an education or training activity that experimental group members began attending on their own initiative prior to orientation that did not meet program requirements for promoting quick entry into jobs. Jobs-First GAIN staff granted enrollees a temporary deferral to complete their current semester of coursework, after which enrollees were supposed to be assigned to job club. DPSS staff did not monitor enrollees’ attendance at these unapproved activities closely and did not provide support service payments to participants in them. In this analysis, as in DPSS’s published reports, these unapproved activities were not counted in calculations of participation rates in Jobs-First GAIN activities.

3. Use of Formal Enforcement Procedures and Sanctioning

A Jobs-First GAIN program enrollee who failed to attend her assigned activity received a notice outlining the sanctions that would be applied if the problem continued. If she did not comply at that point, a conciliation process was initiated in which she was given another notice and another opportunity to resolve the problem. If the enrollee continued to fail to comply, she incurred a sanction, that is, a reduction in her welfare grant amount equal to the value of benefits for one person on the case.8 The first sanction remained in effect until the enrollee met with program staff and agreed to attend a program activity or convinced program staff that she had good cause for not participating. A second sanction lasted a minimum of three months, and subsequent sanctions at least six months, even if the enrollee resumed participation sooner.

The data in Table 2.1 demonstrate that Jobs-First GAIN case managers used formal enforcement procedures very often, although the process only sometimes resulted in imposition of a financial sanction. As shown, 83 percent of experimental group members entered the conciliation process at least once during follow-up, either because they did not show up for an assigned activity or a scheduled meeting with Jobs-First GAIN staff or because they stopped attending a program activity without good cause. The proportion ever in conciliation exceeds the proportion ever assigned to an activity because the conciliation rate includes conciliation for failure to show up at deferral reviews or at scheduled appraisal meetings following a deregistration. Some experimental group members who refused to accept an assignment to job club also entered conciliation status during their initial appraisal meeting following random assignment.

About 30 percent of AFDC-FGs incurred a sanction in the two-year follow-up period. The sanction rate for Jobs-First GAIN exceeds those for the earlier Los Angeles GAIN, Riverside GAIN, and Riverside LFA programs (all with sanction rates of less than 10 percent) by a wide margin. Some employment-focused welfare-to-work programs operating in the 1990s, however, sanctioned comparably large proportions of enrollees.9 Most of the experimental group members sanctioned in Jobs-First GAIN (23 percent) entered that status during the first year of the follow-up period. In year 2, Jobs-First GAIN staff continued to enforce the program’s mandatory participation requirements for those still in the program and sanctioned an additional 8 percent of experimental group members for noncompliance. In addition, some experimental group members who were first sanctioned in year 1 remained in that status for one or more months in year 2.

4. Participation Patterns for Key Subgroups

1. Regular and early enrollees. As discussed in the previous chapter, early enrollees asked DPSS to enroll them in Jobs-First GAIN before they reached the top of the waiting list for services, whereas regular enrollees waited until DPSS required them to enter the program. It was expected, therefore, that a higher percentage of early enrollees would participate in employment-related activities. Participation patterns for these two subgroups confirm this hypothesis. As shown in Table 2.1, the rates of activity assignment and participation for early enrollees exceeded those for regular enrollees by a wide margin. In the two-year follow-up period, program staff referred 72 percent of early enrollees to a Jobs-First GAIN activity, compared to 55 percent of regular enrollees. The overall participation levels of early and regular enrollees, 57 percent and 38 percent, respectively, reflect this difference in assignment rate. A larger proportion of early enrollees than regular enrollees participated in every specific type of activity: 50 percent versus 34 percent went to job search, and 18 percent versus 9 percent attended an education or training activity. These differences most likely reflect early enrollees' greater motivation to participate. Surprisingly, the same percentage of early and regular enrollees entered the conciliation process (83 percent), and a similar proportion incurred a sanction (29 percent and 31 percent, respectively).10

2. Educational attainment, prior employment, and prior welfare receipt. Jobs-First GAIN produced relatively consistent levels of participation for sample members who differed in education, employment history, and previous welfare receipt (see Table 2.3). For instance, about 35 percent of high school graduates and GED recipients and 39 percent of nongraduates participated in job club during the two years after random assignment; around 10 percent of each group attended an education or training activity. Similar results were observed for sample members who worked for pay and for those who did not work for pay in the year before random assignment, and for recent applicants for welfare, short-term recipients, and long-term recipients. The consistency of these results once again reflects the Work First focus of Jobs-First GAIN. Experimental group members without a high school diploma or GED certificate in the earlier evaluation of Los Angeles GAIN, in contrast, were five times more likely to attend basic education classes than job clubs. 11

3. GAIN region, racial/ethnic group, and level of English proficiency. Participation levels varied by race/ethnicity and by GAIN region (see Table 2.3). For instance, about half of experimental group members in the Central and Southeastern regions took part in a Jobs-First GAIN activity, usually job club, compared with around 35 percent in the outlying northern regions of San Fernando Valley and San Gabriel Valley. Participation levels were highest among African-American and Hispanic experimental group members (46 percent for both groups) and markedly lower among whites and Asians (32 percent and 24 percent, respectively).

   Participation levels varied only slightly between Hispanics with and without proficiency in English (see Table 2.3), in large part because DPSS offered job clubs in Spanish. In keeping with the Work First philosophy of the program, Jobs-First GAIN staff assigned most people without English proficiency to job club rather than to English as a Second Language (ESL) classes — unlike in the previous, education-focused GAIN program. In contrast, a higher percentage of whites with English proficiency (34 percent) than those without it (13 percent) attended Jobs-First GAIN activities. A similar pattern was seen among Asian experimental group members. DPSS operated job clubs in Armenian and in Vietnamese and other Southeast Asian languages, so it is not immediately clear why these differences were observed.

   Interestingly, sanction rates were consistently higher for experimental group members who were proficient in English than for those who were not (see Table 2.3). Among AFDC-FGs who were proficient in English, more than 30 percent of whites, African-Americans, and Hispanics and close to 20 percent of Asians incurred a sanction during the follow-up period. In comparison, sanction rates for experimental group members who were not proficient in English ranged from 14 percent (Asians) to 19 percent (Hispanics). Again, it is not clear why these differences occurred.12

5. Participation in Activities Outside Jobs-First GAIN

As discussed above, enrollees in welfare-to-work programs often participate in pre-employment activities on their own initiative. Typically, they enroll in these activities after entering non-mandatory status or leaving welfare. Less commonly, people attend alternative employment-preparation activities when still required to participate in their program – either during deferrals for an unapproved self-initiated activity (discussed above) or without the knowledge of program staff. This pattern occurred in Jobs-First GAIN as well. Table 2.4 shows the rates and total hours of participation in Jobs-First GAIN and non-Jobs-First GAIN activities for AFDC-FGs in the two-year follow-up period.

Los Angeles Jobs-First GAIN Evaluation

Table 2.3

Rates of Participation and Status Among AFDC-FG Experimental Group Members Within Two Years of Orientation, by Region and Subgroup

Region or Subgroup		Sample Size	Participated (%)			Sanctioned (%)	Deregistered (%)
			Any Activity	Job Club	Any Education and Training
Region	San Fernando Valley (Region 2)	2,021	33.3	31.0	6.2	26.5	94.1
	San Gabriel Valley (Region 3)	2,847	36.2	32.8	8.1	26.2	94.8
	Central (Region 4)	1,962	50.6	44.1	14.8	36.7	92.4
	Southern (Region 5) a	2,538	42.5	37.3	10.3	32.9	94.8
	Southeastern (Region 6)	2,153	49.9	42.7	15.1	31.5	92.5
Gender	Male	834	38.5	35.7	6.3	29.4	91.8
	Female	10,687	42.1	37.2	10.9	30.5	94.0
Race/ Ethnicity	White	1,977	31.5	28.4	7.1	29.7	92.9
	African-American	3,606	46.0	39.8	12.2	36.6	94.0
	Hispanic	5,235	45.5	40.7	11.6	28.3	94.1
	Asian	671	23.5	21.0	4.9	15.5	94.1
Proficient in English b_i	Total	9,172	42.9	37.5	11.1	33.4	93.8
	White	1,773	33.6	30.4	7.6	31.1	92.9
	Hispanic	3,574	45.5	39.5	12.3	32.6	94.2
	Asian	274	31.7	28.9	5.8	18.2	91.7
Not proficient in English b_ii	Total	2,349	37.7	35.3	8.6	18.7	94.2
	White	204	13.4	11.0	3.3	18.0	93.2
	Hispanic	1,661	45.4	43.2	10.1	18.8	93.7
	Asian	397	17.9	15.6	4.2	13.6	95.6
Has a high school diploma or GED		5,232	40.4	34.9	10.0	29.5	93.3
Does not have a high school diploma or GED		6,289	43.1	39.9	11.1	31.2	94.4
Applicant		2,069	43.9	37.8	11.8	30.3	93.8
Short-term recipient		2,370	39.5	34.5	9.7	29.2	93.4
Long-term recipient		7,082	42.0	37.7	10.5	30.8	94.0
Employed in year prior to random assignment		4,145	40.5	35.2	9.3	28.8	94.9
Not employed in year prior to random assignment		7,376	42.6	38.2	11.3	31.3	93.3
Most disadvantaged c		2,910	43.0	39.8	11.0	31.5	93.9

SOURCES: MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS), California Employment Development Department Unemployment Insurance earnings records, and LA DPSS Integrated Benefit Payment System AFDC/TANF payment records. NOTES: Full sample means and percentages are weighted averages of results for regular and early enrollees. Measure = (regular enrollee result x percent of experimental and control group regular enrollees in AFDC-FG sample) + (early enrollee result x percent of experimental and control group early enrollees in AFDC-FG sample). Subgroup sample sizes may not sum to the total sample size because of missing data. a This region serves the low-income communities of Watts, Compton, and North Long Beach. (back) b Includes outcomes for African-Americans and Native Americans/Alaskan natives. (back: b_i, b_ii) c The "most disadvantaged" subgroup consists of long-term recipients who did not have a high school diploma or GED certificate at random assignment and who did not work for pay in the year prior to random assignment. (back)




Among experimental group survey respondents (the only group for which data on Jobs-First GAIN and outside activities exist), about 40 percent reported participation in an activity outside Jobs-First GAIN. Half of these respondents had not participated in a Jobs-First GAIN activity according to DPSS records, whereas the remainder attended at least one activity within the program and another activity on their own initiative. Combining the proportion of experimental group members who took part in a Jobs-First GAIN activity (42 percent) and the proportion who participated only outside the program (20 percent) yields an overall participation rate of 62 percent.13

Experimental group members who took part in activities outside Jobs-First GAIN most often participated in vocational training or post-secondary education (see Table 2.4). Notably, about 20 percent of experimental group respondents reported attending post-secondary education classes, an activity that Jobs-First GAIN did not offer. In comparison, less than 10 percent of respondents took part in job search, basic education, work experience, or on-the-job training (OJT), although the levels of participation for survey respondents in three of these activities (all but job search) slightly exceeded the levels recorded for the whole sample of Jobs-First GAIN enrollees.

6. Comparison of Participation Levels in the Experimental and Control Groups

Precluded from participating in Jobs-First GAIN activities, control group members often sought to enhance their job skills or earn an educational credential during the two years after random assignment (see Table 2.4). About 44 percent of control group members in the survey participated in an employment-related activity, usually vocational training or post-secondary education. Between 10 percent and 15 percent of control group respondents reported participation in job search or basic education, and fewer still in work experience or OJT.14

Experimental-control differences in participation represent the Jobs-First GAIN program’s impacts on participation. Welfare-to-work programs like Jobs-First GAIN may affect participation in several ways. They may

1. induce people who would not have attended on their own initiative to attend an employment-related activity;

2. change the type of activity that most people attend; or

3. alter the total number of months or number of hours that people attend an employment-related activity.

Los Angeles Jobs-First GAIN Evaluation

Table 2.4

Two-Year Impacts on Rates and Total Hours of Participation in Jobs-First GAIN and non-Jobs-First GAIN Activities for AFDC-FGs

Outcome			Participated (%)a			Hours of Participation			Hours of Participation Among Participants
			Experi-mental Group	Control Group	Difference (Impact)	Experi-mental Group	Control Group	Difference (Impact)	Experi-mental Group	Control Group	Difference (Impact)
Any Activityb			61.8	44.1	17.6	n/a	n/a	n/a	n/a	n/a	n/a
	Job Search		44.6	13.5	31.1	53.2	13.1	40.1	119.3	97.5	21.8
	Basic education		10.7	11.2	-0.4	62.1	106.9	-44.8	578.1	957.3	-379.1
		ESL	5.9	7.6	-1.7	29.5	65.6	-36.0	501.3	858.6	-357.3
		ABE-GED	4.8	6.6	-1.8	32.5	41.3	-8.7	671.8	621.4	50.4
	Vocational training/ post-secondary		30.8	32.8	-2.1	296.2	314.5	-18.2	962.9	958.2	4.6
		Vocational training	17.7	18.2	-0.5	143.4	155.6	-12.2	810.1	856.5	-46.4
		Post-secondary education	20.1	18.5	1.6	152.8	158.8	-6.1	760.2	858.2	-98.1
	Work experience/ OJT		7.6	2.3	5.3	n/a	n/a	n/a	n/a	n/a
		Work experience	6.0	1.4	4.6	n/a	n/a	n/a	n/a	n/a
		OJT	2.6	1.6	1.0	n/a	n/a	n/a	n/a	n/a
Sample Size			372	374		372	374		(varies)	(varies)

SOURCES: MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS) and from the Two-Year Client Survey. NOTES: Full sample means and percentages are weighted averages of results for regular and early enrollees. Measure = (regular enrollee result x percent of experimental and control group regular enrollees in AFDC-FG sample) + (early enrollee result x percent of experimental and control group early enrollees in AFDC-FG sample). Tests of statistical significance were not performed. Italicized estimates pertain only to sample members who participated in at least one activity. Therefore, the italicized differences between the experimental and control groups are not true experimental comparisons. Some subgroup percentages sum to more than the corresponding group percentages because some sample members participated in more than one activity. N/a = not available or applicable. aParticipation rates include participation for at least one day in Jobs-First GAIN activities or in activities outside of the program. (back) b"Any activity" includes all activities listed. (back)




In the case of Jobs-First GAIN, the first of these effects was modest relative to other Work First programs.15 The program raised overall the participation level about 18 percentage points above the control group level. The program produced a much larger (31 percentage point) increase in use of job search services, however, and a small increase (5 percentage points) in participation in unpaid work experience jobs. This pattern suggests that Jobs-First GAIN caused some experimental group members who would otherwise have attended only education and training classes to attend Work First-type activities (see Table 2.4).

About the same percentage of experimental and control group members attended education and training activities. In addition, the program did not increase the number of hours of participation among respondents who took part in these activities (see Table 2.4).

7. Additional Effects

1. Impacts on educational attainment. DPSS administrators did not intend Jobs-First GAIN to increase the proportion of welfare recipients who attained an education credential. As expected, about the same proportion of experimental and control group respondents (about 14 percent) reported earning either a GED or a training certificate within two years of random assignment. This rate of degree attainment for experimental group members is somewhat surprising, given Jobs-First GAIN’s Work First orientation, but likely reflects their relatively high rates of participation in education and training activities outside the program (results not shown).

2. Impacts on attitudes toward work and welfare. As discussed in Chapter 1, Jobs-First GAIN imparted a strong pro-work message that included both warnings about the impending time limits on welfare and instruction about the financial advantages of combining work and welfare in the short term. Responses to the Two-Year Client Survey suggest that the program achieved modest success in conveying these messages. When asked at the two-year point, about 45 percent of experimental group respondents indicated that they knew about welfare time limits compared with 42 percent of control group respondents, a difference that was not statistically significant (results not shown). On the other hand, Jobs-First GAIN decreased the proportion of respondents who believed that they could provide for their family better by staying on welfare than by working to 8 percentage points below the control group level of 24 percent (a statistically significant difference). In addition, a slightly smaller percentage of experimental group members indicated that they would rather stay home full time to take care of their family than work for pay, but the difference was not statistically significant (results not shown).

4. Results for AFDC-Us

1. Rates of Assignment and Participation

In general, members of two-parent families (AFDC-Us) showed similar patterns of participation and status in Jobs-First GAIN to those described above for single parents: Relatively few participated in Jobs-First GAIN activities, most attended one spell of job club, and very few were still participating in Jobs-First GAIN activities at the end of year 2 (see Table 2.5). Jobs-First GAIN case managers assigned 50 percent of AFDC-Us to a program activity — a lower assignment rate than for AFDC-FGs (58 percent). About a third of AFDC-Us (34 percent) participated in at least one activity for at least one day. This rate is slightly lower than that for two-parent families enrolled in the earlier, education-focused Los Angeles GAIN (36.0 percent) and about half that for two-parent families in the Riverside GAIN program (66.0 percent), both measured over one year of follow-up.16 A much larger percentage of AFDC-Us (32 percent) participated in job club than in every other activity. Only about 5 percent of AFDC-Us attended any of the education or training activities, including basic education.17

Nearly every AFDC-U experimental group member (92 percent) shifted to long-term nonman-datory status, that is, was deregistered, by the end of the two-year follow-up period. Most who were deregistered (57 percent) left mandatory status because of full-time employment. A much smaller percentage of AFDC-Us in the earlier, education-focused Los Angeles GAIN program (34 percent) were deregistered, whereas a similar proportion (80 percent) left Riverside GAIN.18 As mentioned above, the Two-Year Client Survey sample included AFDC-FGs only. It is therefore not known how many experimental and control group members in the AFDC-U sample attended education, training, or other employment-related activities outside Jobs-First GAIN on their own initiative.

2. Participation Patterns for Key Subgroups

As discussed in Chapter 1, the AFDC-U sample was relatively evenly divided between males and females. AFDC-U fathers had longer work histories than AFDC-U mothers, and would therefore be expected to have fewer difficulties finding employment after random assignment. Gender differences in work history and other background characteristics might have affected the experiences of male and female AFDC-Us in Jobs-First GAIN — for instance, the likelihood of their attending job club or the frequency with which they incurred a sanction or were deregistered. As shown in Table 2.6, male AFDC-Us had somewhat more contact with the program, but the differences between gender groups in patterns of participation and program status were not large. The participation level for AFDC-U men (37 percent) slightly exceeded that for AFDC-U women (32 percent), but a higher percentage of males than females incurred a sanction (26 percent versus 21 percent). In addition, deregistration rates were higher for males than females (95 percent versus 90 percent), especially for full-time employment (65 percent versus 48 percent; results not shown).

There was greater variation between subgroups defined by level of educational attainment, by race/ethnicity, and by GAIN region in level of participation and frequency of incurring a sanction (see Table 2.6). For instance, 37 percent of AFDC-Us who entered Jobs-First GAIN without a high school diploma or GED certificate attended job club, compared with just 26 percent of high school graduates or GED recipients. More dramatic differences in participation level were observed among the four largest racial/ethnic subgroups of the AFDC-U sample. Specifically, participation levels for Hispanics and African-Americans (47 percent and 43 percent, respectively) were more than twice as high as those for whites and exceeded the levels for first- and second-generation immigrants from Vietnam, Cambodia, and every other Asian country by nearly the same factor. African-Americans and Hispanics were also much more likely to incur a financial sanction during the first year of follow-up. Similarly, AFDC-Us from the San Fernando Valley, 60 percent of whom were white, were much less likely to participate in a Jobs-First GAIN activity than sample members residing elsewhere in the county. In contrast, AFDC-Us from the Southeastern region, nearly all of whom were Hispanic, had the highest level of participation of any racial/ethnic subgroup of AFDC-Us. As was the case for Hispanic AFDC-FGs, participation rates were about the same for Hispanic AFDC-Us with and without English proficiency. Among white and Asian AFDC-Us, a higher percentage of experimental group members who were proficient in English than of those who were not proficient in English participated in Jobs-First GAIN activities.

Los Angeles Jobs-First GAIN Evaluation

Table 2.5

Rates of Participation and Status Within Two Years of Orientation for AFDC-U Experimental Group Members

Outcome (%)
Assigned to any activity a			49.6
Ever participated in b	Any activity c		34.4
	Job search		32.1
	Any education or training		4.8
	Basic education		2.3
		ESL	1.4
		ABE	0.3
		GED	0.5
		High school	0.2
	Vocational training		2.9
	Work experience		2.6
	OJT		0.0
	Assessment		4.3
Deregistered for any reason d			92.4
	For employment		57.0
	For sanction		31.7
	For other reason		51.8
In conciliation			79.5
	Sanctioned		23.4
Deferred for any reason			47.0
	For unapproved SIT e		6.9
Sample size			4,039

SOURCE:MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS). NOTES: a The assignment rate includes assignment to all activities listed, except assessment. (back) b Participation rates include participation for at least one day in either a program-referred or approved self-initiated activity. (back) c "Any activity" includes all activities listed, except assessment. (back) d Subgroup percentages sum to more than the total deregistration percentage because some recipients were deregistered more than once during the follow-up period. (back) e An SIT is a self-initiated activity ("self-initiated training"). (back)




Los Angeles Jobs-First GAIN Evaluation

Table 2.6

Rates of Participation and Status Among AFDC-U Experimental Group Members Within Two Years of Orientation, by Region and Subgroup

Region or Subgroup		Sample Size	Participated (%)			Sanctioned (%)	Deregistered (%)
			Any Activity	Job Club	Any Education and Training
Region	San Fernando Valley (Region 2)	1,209	21.8	20.3	2.0	19.4	92.1
	San Gabriel Valley (Region 3)	1,095	32.9	31.4	2.9	20.7	93.2
	Central (Region 4)	472	34.7	31.4	6.1	27.3	92.2
	Southern (Region 5) a	481	40.3	37.6	7.9	25.8	92.9
	Southeastern (Region 6)	782	51.9	48.6	9.2	29.7	91.9
Gender	Male	2,118	36.5	34.8	4.1	26.1	95.0
	Female	1,921	32.0	29.2	5.6	20.6	89.6
Race/ Ethnicity	White	1,149	18.0	16.0	3.0	18.5	91.5
	African-American	212	43.4	42.0	7.1	35.4	87.7
	Hispanic	1,906	46.7	44.2	6.1	28.0	93.0
	Asian	766	25.8	23.6	3.7	16.4	93.9
Proficient in English bi	Total	1,947	40.0	37.2	5.8	29.4	91.2
	White	574	28.6	26.7	3.3	25.8	89.2
	Hispanic	963	47.0	43.2	7.3	31.8	92.8
	Asian	206	36.4	34.5	4.4	22.3	93.2
Not proficient in English bii	Total	2,092	29.1	27.4	4.0	17.9	93.6
	White	575	7.5	5.4	2.8	11.3	93.7
	Hispanic	943	46.3	45.3	4.9	24.1	93.2
	Asian	560	22.0	19.6	3.4	14.3	94.1
Has a high school diploma or GED		1,650	28.3	25.8	4.7	22.4	91.0
Does not have a high school diploma or GED		2,389	38.6	36.5	4.9	24.2	93.5
Applicant		446	41.0	38.8	7.2	24.0	91.3
Short-term recipient		1,226	39.6	36.8	6.3	25.6	90.7
Long-term recipient		2,367	30.4	28.5	3.6	22.2	93.6
Employed in year prior to random assignment		1,406	37.3	34.5	5.8	24.2	93.4
Not employed in year prior to random assignment		2,633	32.8	30.9	4.3	23.1	91.9
Most disadvantaged c		1,021	36.1	34.8	3.4	24.4	94.8

SOURCES: MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS), California Employment Development Department Unemployment Insurance earnings records, and LA DPSS Integrated Benefit Payment System AFDC/TANF payment records. NOTES: Subgroup sample sizes may not sum to the total sample size because of missing data. a This region serves the low-income communities of Watts, Compton, and North Long Beach. (back) b Includes outcomes for African-Americans and Native Americans/Alaskan natives. (back: b_i, b_ii) c The "most disadvantaged" subgroup consists of long-term recipients who did not have a high school diploma or GED certificate at random assignment and who did not work for pay in the year prior to random assignment. (back)