Number of Visitors Who have viewed this page since 20-Feb-2006
SOCIAL NETWORKS, CAREER AND TRAINING PATHS
FOR PARTICIPANTS IN EDUCATION AND TRAINING PROGRAMS
Technical Report
The Social Network Study was conducted in 1996 as part of a larger research project looking at
the experiences of men and women with different racial, ethnic, educational backgrounds and
career paths in the Philadelphia labor market. This study was conducted with the cooperation of
the Private Industry Council of Philadelphia (PIC) and a number of training providers in the
Philadelphia area as a project of the Institute for the Study of Civic Values (ISCV). The social
network survey focused on the relationship between work experience, training, and the people
and organizations which welfare recipients use as resources in order to find jobs and educational
programs. Questionnaires were collected from 338 people currently enrolled in nine training
programs or community college.
This report includes a detailed summary of findings from the Social Network study. It is
designed as a supplement to a policy report using these data prepared for the 21st Century
League which focuses on ways to aid government, training providers and employers working
with public assistance recipients moving from welfare to work.
History of the Study and Methodology
Marital Status and Number of Children
Education and Training: General Patterns
Post-Secondary Education Experience
Major types of training programs taken
Comparing Free and Tuition Based Programs
Policy Eras and Types of Training:
Work Experience: General Patterns
Held a job which provided health insurance
Types of Employment Clusters and Career Ladders:
Work Characteristics and Reasons for Leaving Work
Demographic Composites and Work and Training Patterns
Training and Work Experience Patterns
Length of Time Between Training Programs
Relationship of training to previous work experience and demographic characteristics
The Training Track for African Americans and Latinos
Social Networks, Career and Training Paths
Demographic Patterns and Social Networks
The Training Track and Social Networks
Social Networks and Tuition Based vs Free Programs
Social Networks and the Role of Training Programs in Finding Work
Performance Based Contracts versus Tuition Payments in Program Outcomes
Appendix A: Differences Between Agencies Participating in the Study
History of the Study and Methodology
The Social Network Study grew out of analysis of anonymous surveys collected by Community Women's Education Project (CWEP) in the late 1980s and early 1990s. This statistical study included all 373 people enrolled in the CWEP workstart program over 5 years. It traced the previous work and training experience of program participants. Findings revealed that 52 percent of participants had completed high school, 7 percent had gone on to college, 59 percent had attended vocational training. Only 4 percent had never worked and 92 percent had started work before the age of 21. Most had held between two and five jobs. This project revealed a pattern where African American women would attend training programs linked to jobs which would end. Since the questionnaire lacked social network information and did not include time information which would allow researchers to understand the sequence of training and jobs, the reasons for this pattern were not clear. The Social Network Study was designed to explore several questions raised by this initial research:
The Social Networks, Career and Training Paths for Disadvantaged Populations study was
designed to address these questions. The questionnaire was revised and piloted at Community
Women's Education Project and the Institute for the Study of Civic Values. This work was
carried out by the author with the assistance of college students from Bryn Mawr College,
Haverford College, Hope College and Mansfield University enrolled in the Campus/Community
Collaborative program at the Institute. Redesign took place during 1993 through 1995.
The new study included several types of information. The entire questionnaire is in Appendix B.
The first section asked about demographics: age, gender, race/nationality, immigration year,
country of origin, number and age of children, if the respondent had ever married and current
marital status, highest level of education, reason for not completing high school for those who
were not formal high school graduates, and military history.
The second section asked for detailed information on up to four training programs which
participants enrolled in after leaving high school. This section asked for age and year the
participant entered the program, type of training, length of program, cost, whether or not the
person completed the program and if it led to employment. In order to understand the social
networks that led to training, participants were asked how they heard about the program. The
final set of questions in this section asked if the participants had taken out student loans, for how
much, and if those loans were in default.
The third section focused on employment. Participants were asked about all jobs after leaving
high school. They recorded the year and how old they were when they started this job.
Information on type of job, full or part-time, wages, whether or not the job provide health
insurance and how long they stayed at the job. This section also collected information on how
they found the job and why they left. Final questions in this section asked about problems
finding work.
The final section asked about current source of income. This included length of time on welfare
for this spell and, for those currently not on public assistance, total length of time ever on
welfare. The final question asked about why they entered their current program.
Data were collected from 338 people enrolled in nine training programs or community college in
Philadelphia in late 1995 and 1996. The Philadelphia PIC has contracts with the Commonwealth
of Pennsylvania to provide education and training for Philadelphia residents through several
Federal government programs, including JTPA and the Family Support Act of 1988 (JOBS).
Based on a combination of Commonwealth and Federal regulations, the PIC was responsible for
meeting training and placement goals for the people enrolled in their programs. For most of the
time documented in this study, in Pennsylvania, both JTPA and JOBS programs required that a
certain percentage of participants be placed in full-time jobs which paid $6.00. Half of the jobs
should provide health insurance. Placement rates were measured by counting the number of
people who remained at their job placement 30 and 90 days after they find work. The Family
Support Act required that participants be in a training related activity for 20 hours a week. The
PIC used both JTPA and JOBS funds to subcontract with non-profit and for profit organizations
to provide most of its training. Since both JTPA and JOBS standards had to be met for most
programs, they all had placement goals which fit both standards and must offer 20 hours of
activity.
Since many disadvantaged Philadelphia residents lack basic math and reading skills needed to
complete job specific skills programs, the Philadelphia PIC created a two tier training system.
People who did not have sufficient reading and math levels to succeed in a job specific skills
program were first sent to a "feeder" program which focused on teaching adult basic skills.
Feeder contracts required that they transition their graduates into either a job or a specific skills
program. The job specific skills programs were expected to place their graduates into paid
employment. Most study participants were enrolled in these two kinds of programs.
In addition to the skills training system, the PIC was involved in a number of other kinds of
Commonwealth and Federal programs designed to move public assistance recipients from
welfare to work. The Family Support Act required that two parent families on welfare (AFDC-U) participate in 16 hours a week of community service in order to qualify for benefits. The
Ridge administration had instituted six week mandatory job development programs for public
assistance recipients who met certain criteria. The study also included participants in both of
these kinds of programs.
With the support of the PIC, agencies were selected which include the range of adult basic skills,
job specific skills serving both men and women, and mandatory job preparation and workfare
programs available in Philadelphia. Care was taken to ensure that the study include participants
in all of the major types of programs funded through the PIC. These included a clerical program
(Arbor-Cite), a nursing assistant program (1199c), two job specific skills programs serving
primarily men (PHDC, OIC), a homeless shelter providing both basic and job specific skills
(People's Emergency Center), two "feeder" programs focused on populations who needed basic
skills remediation before skills training (CWEP, ISCV AWEP program preparation class), a
mandatory job development program (PIC-Upfront), and a mandatory community service
program for two parent families (ISCV AWEP-UP). Three classes at community college (one
general education, one upper level professional, and one upper level combined) and some STEP-UP enrollees also participated in the study. Questionnaires were given to all participants in
classes which included both disadvantaged and more general training program participants.
Students filled out and returned questionnaires on a voluntary basis.
Statistical analysis focused both on general relationships between demographic characteristics,
social networks and work/training experience and tracing time line experiences with work and
training. Most of the statistical analysis involved measures for qualitative data. These included
cluster analysis of demographic characteristics, types of employment and types of training. Much
of the analysis presented here involves examining cross tabulations of various information. Most
comparative statistics rely on chi-square, measures of linear relationships for nominal data,
correlations and analysis of variance. In quantitative research, a statistically significant
probability score measures how likely it is that a pattern did not occur by chance. Findings were
considered statistically significant at the p.01 level. Most of the patterns reported here were
significant at the p.001 or p.000 level.
Differences between participants in the various programs are discussed in Appendix A. The
participants in college fell into several groups: younger people who were more likely to be
working or supported by family and working age students which were very similar to people in
the various other programs. The mandatory community service program differed from many of
the other programs because it served a large number of older, white male dislocated workers and
refugees with high levels of education from their country of origin. Where differences in the
study population were due to significant differences among the people enrolled in a particular
program, that information is noted here.
Age: Most of the people in this study were working age:
18-21 9%
22-25 39%
36-45 28%
over 45 6%
Men were slightly older than women. Nearly 13 percent of men were over 45 compared to 5
percent of women. Most of these older men were older dislocated workers in the ISCV
mandatory community service program.
There was no significant difference in age for people from different races or nationalities.
Gender: 16% male, 84% female. More African American's (88%) and Latinos (82%) were
women then whites (68%) and Asians (40%).
Race:
African American 74%
White 16%
Latino 7%
Native American 1%
Asian 2%
Biracial 1%
Migrated to the U.S. Mainland: Eight percent of the people (26 ) in this study came to
Philadelphia from outside the U.S. mainland. This group included people who came to the U.S.
in several different ways. Thirty-six percent of this population were Puerto Rican citizens who
had migrated to the mainland. Thirty six percent were refugees: 20 percent from Eastern Europe
and the Soviet Union and 16 percent from Southeast Asia. Only 20 percent (7 people) were
actual immigrants to this country. They came from Asia, Central and South America, and the
Caribbean.
The majority of these people had been in the U.S. more than 10 years. Fifty percent arrived
before 1985, most after 1968. Seventy-two percent had arrived by 1992. Only 18 percent had
come to the U.S. mainland in the three years before the study.
Not everyone who was Latino or Asian should be thought of as a recent migrant with limited
grasp of English and little U.S. work experience. While all of the Asians had been born in
another country, many of these were young college students who may have been raised in the
U.S.
Only 48 percent of the Latinos were born outside of the U.S. mainland. While the majority of
the migrants had limited English and U.S. work experience, the education and work backgrounds
of U.S. born Latinos was similar to that of African Americans.
Very few whites (13%) and African Americans (2%) came from outside the U.S. mainland.
Most of these people were refugees or Haitian entrants.
Marital Status and Number of Children: Thirty-seven percent of the study population had
ever been married. Of those who had married, 62 percent were still married.
Marital status differed dramatically by gender and race. Sixty-two percent of men had married
versus 32 percent of the women. Seventy-eight percent of the men were still married compared to
56 percent of the women. This difference is largely due to the fact that many of the men in this
sample were in the mandatory two parent community service program.
Older age groups were more likely to have married. Eighty-nine percent of those over age 45, 56
percent of those ages 36 to 45, and 29 percent of those 22 to 35 had ever married. There was no
significant difference in age among those who were still married.
African Americans had been married much less frequently than any other group. Only 27
percent of African Americans had ever been married compared to 56 percent of whites, 72
percent of Latinos and 60 percent of Asians.
Most people in this study had children. Overall, 92 percent had children. African Americans
(94%) and Latinos (100%) were more likely to have children than whites (80%) and Asians
(60%). Differences in marital status and having children for Asians are primarily due to the fact
that many of the Asians were young college students.
Most of these people had small families. Forty-nine percent had two children or less. Twenty-one percent only had one child. Sixty-six percent had three children or less and another 10
percent had a fourth child. Only 15 people altogether had six children or more. There was no
difference in the number of children across race and gender.
Present Source of Income: Only 6 percent of this population had never been on welfare. Most were presently on welfare:
Welfare 83%
Job 10%
Spouse/family 5%
Other 3%
For those currently on welfare, statistics mirrored the national welfare figures:
Under 1 year 22%
1-3 yrs 28%
3-5 yrs 22%
5-10 yrs 15%
Over 10 yrs 12%
Of those who were not currently on welfare, 66 percent had been on it in the past. However, this group had spent less time on welfare overall. Figures for total time on welfare for this population broke down as follows:
Under 1 year 44%
1-3 yrs 31%
3-5 yrs 8%
5-10 yrs 14%
Over 10 yrs 3%
Men were less likely to be on welfare (66%) than women (86%). Men were more likely to be
currently working (28%) than women (7%). For those who were currently on welfare, 46 percent
of men had been on less than one year and 86 percent overall three years or less. Only 17 percent
of women had been on the system less than one year. Twenty-four percent had been on the
system three to five years and 32 percent five years or more. Overall, 56 percent of women had
been on public assistance three years or more. This difference between men and women was
primarily due to the fact that most men in this study were either in college or the mandatory
community service program which included many dislocated workers who had recently ran out
of the unemployment benefits and turned to welfare for support.
More African Americans (86%) and Latinos (90%) were on welfare compared to whites (66%)
and Asians (60%). More whites (19%) had jobs versus African Americans (11%) and Latinos
(10%). Thirteen percent of whites were supported by their families. There was no significant
difference across race and nationality for time on welfare.
Only 69 percent of the under 21 age group was currently on welfare. This trend was also due to
the fact that a larger part of the under 21 group were enrolled in college. Other age groups were
over 80 percent. Forty-six percent of the over 45 population had been on welfare less than one
year. Again, these figures reflect older dislocated workers. Thirty percent of the 36 to 45 age
group had been on public assistance for over 10 years.
Military experience: Roughly 4 percent of the people in this study had been in the military.
This included 36 percent of the men and 2 percent of the women. Fifty-five percent of those who
had military experience used the military as a significant part of their career, spending more than
four years in the service. There was no difference across race or age in military experience.
Education and Training: General Patterns:
The majority of the people in this study had completed high school or a GED: 56% finished high
school, 39% of those who had not completed high school finished a GED, 68% had a diploma or
GED. The "Diploma" variable included everyone who had actually received a diploma or GED,
as opposed to people who reported that they had finished high school, making the percentage
slightly lower than the percent for "finish high school."
Gender: Men (68%) were more likely to finish high school than women (53%). Nearly 80
percent of men had a diploma, compared to 66 percent of women.
Race: African Americans (53%) and Latinos (47%) were much less likely to finish high school
than whites (75%) and Asians (80%). However, more African Americans and Asians completed
GEDs, bringing statistics for actually having a diploma to 67 percent for African Americans, 70
percent for whites, 64 percent for Latinos and 100 percent for Asians.
Age: Younger populations were much less likely to have finished high school. Only 39 percent of
the under 21 population had completed high school, compared to 52 percent of those 22 to 35, 67
percent of those 36 to 45 and 75 percent of those over 45. The same patterns held once GEDs
were included. Fifty-five percent under 21, 64 percent 22 to 35, 81 percent 36 to 45, and 74
percent over 45 had a diploma.
Reasons for Leaving: Reasons for leaving school did not vary across race or age. Overall, 30
percent left due to pregnancy, 21 percent for money or work, 26 percent for family problems, and
10 percent did not want to go back. Women left more often because they were pregnant (35%)
while men left because they needed money or went to work (42%). An equal number of men
(25%) and women (26%) left for family problems.
Post-Secondary Education Experience:
Most people in this study had gone to a training program before entering the activity where they
were currently enrolled. 83 percent had attended a training program: 38 percent one, 33
percent two, 12 percent three or more. There was no difference in attending programs or the
number attended between men and women. The over 45 group(53%) and under 21 age group
(47%) were most likely to have only completed one program.
Program attendance did vary for whites and people of color. Eighty-five percent of African
Americans and Latinos had attended a training program compared to sixty-eight percent for
whites. African Americans took many more programs than other groups. Forty-eight percent of
African Americans compared to 28 percent for whites and 25 percent for Latinos took more than
one program. As discussed in the social network section, African Americans and established
Latinos have more access to training programs than whites due to low income and the familiarity
of these groups in using social service systems.
Completion rates also vary across race. Fifty-eight percent of African Americans and 60 percent
of Asians had completed at least one program, versus whites (52%) and Latinos (38%).
However, given that African Americans take more programs, and Latinos less, the completion
rates for these groups are actually much lower than for whites and Asians. Women (44%) are
less likely to complete any programs than men (33%). Overall, the number of programs
completed were: none 43 percent, one 40 percent, two 13 percent, three 4 percent.
In current programs, African Americans predominated in mandatory job development (83%), job
specific skills programs (88%), and the shelter (90%), compared to mandatory community
service (25%), feeders (62%) and college (55%). Forty-three percent of the people in this sample
in mandatory community service were white, 25 percent in feeders, 26 percent college, and under
11 percent for the shelter and JSS programs. Latinos also were found in the mandatory
community service programs (18%), feeders (11%), college (8%), but only 4 percent of the job
specific skills enrollees. This shows that Latinos were not being served by the job specific skills
training network. It also shows that whites were less likely to participate in the PIC funded
training program system. Whites and Asians were served by the community college and tuition
based training systems while Latinos were left out of training altogether.
Loans: 53 percent had taken out loans, 57 percent for more than $2,000, 26 percent for $500-2,000, 75 percent still owed and 64 percent of those were in default. There was not a significant
difference between men and women on the number that had loans or the amount. However, 52
percent of men versus 79 percent of women still owed on those loans. Sixty-eight percent of
women's loans were in default compared to 39 percent for men. There was no difference in loan
or default rates among various races and nationalities.
Major types of training programs taken: The study measured participation in training programs in two ways. First, information on the number of people who had ever taken a course in a particular subject were counted. Overall, the following percentages of the people in this study and ever taken courses in these topics:
Data/clerical 34%
Skilled trades 13%
College 16%
Nursing assistant or home health aid 20%
Business 17%
GED/ABE 14%
Job Preparation 13%
"Business" courses were often another term for secretarial training. In later cluster analysis of
these programs, they are combined with data/clerical, making this kind of training the most
popular kind of training overall. Over 50 percent of the people in this study had taken some form
or clerical or computer word processing training.
Gender and Race: The kinds of training programs varied dramatically by gender, race and basic
education. Only 7 percent of the men in the sample had taken data/clerical training. Thirty
percent of the women had taken one clerical program and another 10 percent two or more clerical
training programs. Twenty-three percent of women had trained to be nursing assistants,
compared to three percent of men. Fifteen percent of men had taken foodservice compared to 5
percent of women.
African Americans (37%) and Latinas (35%) were much more likely to take clerical training than
whites (13%). However, this was largely due to the fact that most whites in the sample were
men, not because whites did not take this kind of training. White women were most likely to
take clerical or GED training. Whites (34%) and Asians (80%) were most likely to go to college.
Only 12 percent of African Americans had gone to college. Nursing assistant programs were
exclusively taken by people of color: 25 percent of African Americans and 12 percent of Latinos
took these programs.
Basic Education: There was a significant difference in types of training taken based on high
school education. Sixty-seven percent of the people in clerical training had finished high school
and 79 percent had a diploma. People who had not completed high school went into nursing
assistant programs. Fifty-eight percent of the people in these programs had not finished high
school. Virtually everyone enrolled in college had a high school diploma or GED. There was no
significant difference in basic education for other programs.
Summary:Taken together, these statistics show some general patterns steering people toward different types of careers. People of color without high school educations, particularly women, went into nursing assistant programs. Foodservice also provided training for people with lower educational credentials, serving more men. Clerical programs were taken almost exclusively by women with solid basic education credentials. As explained later, these programs led to good career paths, particularly for women of color. Gender mattered little in going to college, but race made a large difference. Again, some of this trend is due to basic education credentials. All Asians and a larger percentage of whites had a high school diploma or GED.
Program Length: These figures show the percentage of people who have ever taken a program of this length. The majority of this population had taken short term training:
Under 2 months 17%
3-6 months 39%
6-12 months 38%
12-18 months 20%
Over 18 months 1%
Gender and Race: Program length did not vary much for men and women. However, there was
a significant different among people from different races. People of color were much more likely
to be in programs that ran three to six months. Forty-two percent of African Americans, 41
percent of Latinos and 40 percent of Asians had participated in this type of program compared to
22 percent for whites. This difference reflects participation in focused skills training programs.
African Americans (52%) were much more likely to go to six to twelve month programs,
compared to whites (25%) and Latinos (29%). Whites (41%) and Asians (80%) were much
more likely to be in 12-18 month programs than African Americans (21%) and Latinos (23%).
This reflects greater participation in college and union skilled trade programs.
Program led to a job: Overall, program participants stated that their training led to a job 30
percent of the time. The program did not lead to a job 40 percent of the time.
Comparing Free and Tuition Based Programs:
Using cluster analysis, a statistical procedure which groups together information with similar
characteristics, the various types of programs were combined into the following categories:
GED/Job Prep: GED and job preparation courses were under 6 months and free to participants
Nursing assistant/foodservice, 3-12 mo, free
Nurse assistant/foodservice, 3-12 mo cost
Skilled trade/security/ hair styling, Greater than 6 mo, free
Skilled trade/security/hair styling, Greater than 6 mo, cost
Clerical/business,Less than 6 mo, cost
Clerical/business, Less than 6 mo, free
Clerical/business, Greater than 6 mo, cost
Clerical/business, Greater than 6 mo free
College
Policy Eras and Types of Training: The kinds of training available changed over time. "Great
Society" programs include courses taken before 1980, most during the Johnson era CETA
program. Seventeen percent of the people went to their first training program during this time
period. Reagan era programs ran from 1981-1988. Twenty-two percent of the first training
programs were funded under JTPA and other Reagan initiatives. Sixty-one percent of the first
training included programs funded under the Family Support Act of 1988. Most second and later
programs occurred after the Family Support Act had gone into effect.
During the great society era, training focused on job specific skills. Forty-six percent of the free
skilled trades training happened in this time. This may refer to union apprenticeship programs.
Clerical programs included short, free programs (20%) and greater than six month free programs
(36%). Twenty-three percent of the cost nursing assistant programs also occurred during this era.
Sixty-six percent of college training also occurred during this time.
People participated in much more tuition based training during the Reagan era. Forty-six percent
of the tuition based nursing assistant programs, 58 percent of the cost skilled trades programs ,
64 percent of the under six month cost clerical training and 16 percent of the greater than six
months clerical programs happened during this time. GED and job development programs
started in this era. Twenty-percent of first programs were Reagan era GED or job development
programs.
Most of the GED and job development (73%) occurred during the Family Support Act. Both
free and tuition based job specific skills programs continued. Clerical programs became longer:
70 percent of both the free and cost greater than six months programs were Family Support Act
programs.
Drop-Out Rates:There is a dramatic difference in drop out rates between tuition based and cost programs. For the first program, drop out rates for tuition based programs range from 39 percent for nursing assistant programs to 30 percent for skilled trades. The only free job specific skills program that reported students who did not finish the program was skilled trades (22%). College had the highest retention problems, with a 44 percent drop out rate. Drop out rates for GED/Job Prep programs were 21 percent for the first program.
Work Experience: General Patterns
Number of Jobs: Like the CWEP study, most of the people in the Social Network study had
held multiple jobs. The Social Network study did not include jobs held while attending high
school, which underestimates the work experience of the population, but gives a better
understanding of adult work experience. Overall, 13 percent of this study population had never
worked, 42 percent had held one or two jobs and 45 percent had held three or more. For more
detailed analysis, the population that never worked was expanded to include people who had
only held one job in their adult life which lasted less than one year. This brought the total who
had never worked up to 23 percent.
Gender and Race: Women (15%) were much more likely to have never worked than men (2%).
Men (18%) were also much more likely to have held six jobs then women (5%). Number of jobs
was not otherwise statistically significant.
African Americans (25%) had the highest percentage of the study population which had held
only one job, compared to ten to thirteen percent for other groups. Latinos (18%), primarily
women, had not worked at all. Fewer African Americans (41%) had worked more than three
jobs, compared to whites (58%) and Latinos (54%).
People who had not finished high school (18%) were more likely to have never worked,
compared to those who had finished high school (9%). Twenty-nine percent of those who had
not finished high school had held only one job, compared to eighteen percent for those who had
finished high school.
Major types of jobs: The study measured overall work experience patterns and then looked at
experience over time. The most common jobs held by study participants were as follows:
Cashier 36%
Data entry or clerical 27%
Nursing assistant 17%
Sales 13%
Security guard 10%
Factory worker or driver 23%
Maintenance 10%
Restaurant work 19%
Professional or professional entry level 9%
Gender: As with training programs, jobs fell into traditional gender categories. Many more
women (41%) than men (12%) worked as cashiers. Women (31%) held most of the clerical jobs,
compared to men (10%). Forty-six percent of men worked in factories, compared to sixteen
percent for women. Men also worked in skilled trades and maintenance (36%), while only a few
women (6%) held these kinds of jobs.
Age: Types of jobs also changed by age. These differences reflect changes in the economy over
time from factory work to service occupations. Fifty-seven percent of those under 21 had
worked as cashiers, 44 percent of those 22 to 35, 21 percent of people 36 to 45 and 11 percent of
those over 45. Seventeen percent of those under age 21 had worked in sales, 21 percent of those
36 to 45 and less than 10 percent for other groups. Forty-four percent of those ages 36 to 45 had
worked in computer/clerical, 22 percent for both 22 to 35 and over 45 and nine percent for under
21. Fifty percent of those over the age of 45 had worked in factories, compared to 22 percent for
those 36 to 45, 19 percent of people 22 to 35, and 17 percent for under 21. There was no
difference across age groups for nursing assistants, restaurant work, security guards, childcare or
professional positions.
Race: Differences among people from different races followed similar patterns to training. There
was no racial difference in those who performed clerical work. Many more African Americans
(38%) and Latinos (44%) had worked as cashiers than whites (16%). Nursing assistants were
primarily African American (20%), compared to less than six percent for other groups. Only
African Americans worked as security guards. There was no difference for other occupations.
Basic Education: Basic education made a significant difference in some occupations. Ninety-two percent of the people in clerical positions had a high school diploma or a GED. Seventy-three percent of those in sales positions had finished high school. Ninety-one percent of those in
professional jobs had a diploma. There was no significant difference in basic credentials for
those in other occupations.
Wage Levels: Overall, people in the study population had held jobs paying the following wages:
Lt $5 an hour 50%
$5-7 62%
$7-9 34%
$9-11 10%
Over $11 10%
Gender: Wages for men and women did not differ until over $11 per hour. Twenty percent of
men earned more than $11 compared to 7 percent of women. Men saw more wage progression
between their first and last jobs than women. For example a progression of one category may
mean earning less than $5 an hour at a first job and $5-7 at the last one. A three category jump
might go from $5-7 to over $11. Fourteen percent of women saw no wage difference compared
to six percent for men, 39 percent of men went up one wage category between their first and last
job and 27 percent of men earned three wage categories or more difference from their first job
compared to 12 percent for women.
Race: African Americans started out at slightly higher wages than other groups. Fifty-one
percent of African Americans started at less than $5.00 an hour compared to 64 percent for
whites and 88 percent for Latinos. Thirty-nine percent of African Americans started at $5-7 an
hour. Whites had more wage progression than other groups. Fifteen percent of African
Americans and 18 percent of Latinos had no wage progression compared to none for whites.
Thirty-five percent of whites, 47 percent of African Americans and 54 percent of Latinos only
progressed one category between their first and last jobs.
Employed full time: A full 77 percent of the study population had held a full time job. Forty-five percent had been employed part time. There was no significant difference in holding a full time job between men and women, but more women (48%) had held part-time jobs than men (32%). Younger people were more likely to never have held a full time job. Only 48 percent of those under 21 had held a full time job, compared to 78 percent or over for older groups. There was no significant difference among age groups for part-time work. There was no significant difference among races for either full or part-time work. Those with a diploma (82%) were much more likely to hold full-time jobs than those without (66%). Education made no difference in holding a part-time job.
Held a job which provided health insurance Most of these people had not held jobs which
provided health insurance. Forty-two percent had held a job with health insurance. Thirty-eight
percent of the women had held jobs with health insurance compared to 56 percent of men.
Younger groups had not had jobs with health insurance. Only 8 percent of those under 21 had
held a job with health insurance, compared to 36 percent of those 22 to 35, 63 percent of those 36
to 45 and 50 percent of those over age 45. There was no difference among different races in
employment with health insurance. Fifty percent of those with a high school diploma held jobs
with health insurance compared to 23 percent of those who did not have a diploma or GED.
Length of time on the job The study looked at the length of time that people had held jobs.
These figures could represent part-time, summer jobs before entering a training program or
people who did not stay at a job for a long time. Overall people had held jobs for the following
length of time:
Under 6 months 41%
6-11 months 42%
1-2 years 48%
2-5 years 36%
More than 5 years 17%
Gender and Race: There was no significant difference regarding length of time on the job
between men and women or among different races. Younger people were more likely to be in
jobs a short amount of time while older people had held jobs for long periods of time. Eighty-two percent of those under age 21 had held a job for less than six months, 47 percent of those 22
to 35, 24 percent of those age 36 to 45 and only 17 percent of those over age 45. Thirty-four
percent of those between 36 to 45 and 44 percent of those over age 45 had held jobs for over 5
years. These differences reflect the fact that many of the older workers had been displaced from
stable, long-term employment.
Basic Education: Education made a difference for those who were in jobs under 6 months. Fifty percent of those who had not finished high school had worked at a job less than 6 months compared to 36 percent of those who had finished. Seventy-five percent of those who held jobs two to five years had finished high school compared to 25 percent for those who had not finished high school.
Types of Employment Clusters and Career Ladders: Information on job title, wages and
health insurance were analyzed using cluster analysis. Jobs were grouped into several categories.
Low end service refers to cashier, sales, restaurant, service and bus matron positions which
require little training and pay low wages. Clerical positions include both secretarial and data
entry jobs. Semi-skilled helping professions included jobs like nursing assistant, childcare, maid,
teacher's aid and laundry that involve little formal training, pay low wages and involve primary
care to others. Blue collar positions include both semi-skilled factory work and skilled trades.
Professional jobs require significant formal training, but may not pay well.
Ideally, people would progress from low end service positions as a first job to better jobs over the
life of a career. As discussed in detail in the section of worker types and employment paths, this
was true for some people. Others found themselves returning to low paid positions again and
again.
Table 1 shows the percentage of people in each job for three jobs. This table suggests two
patterns for employment over time. Examining the first three categories, labeled low end service,
show one pattern. If people were progressing from entry-level jobs to better paid positions, the
percentages in the least desirable jobs - low end service, part-time, no health insurance, would go
down between the first and third jobs. But these figures do not drop much over time. The same
is true for the other two categories. They do show some movement out of these kinds of jobs,
but not as much as a ladder toward better employment would suggest.
The same pattern is true for helping professions like nursing assistant. There is an increase in
part-time, low-wage jobs without health insurance, but no change for the better helping
profession jobs. People move into poorly paid, part-time nursing assistant positions, but do not
seem to move up from there.
Examining the categories for clerical positions, blue collar jobs and professional jobs shows a
different pattern. In these fields, more people have career ladders. People do seem to move from
entry-level jobs into work with better pay, hours and benefits. The percentage in entry level
clerical jobs drops in half, from four to two percent between the first and second job. This
category drops to one percent by the third job. Progress means several options. People seem to
move into full time work at various wage levels and benefits. Percentages for most of these
categories double by second and third jobs. The third clerical category: full time, low wages and
health insurance seems to serve as an entry level position for some people. Blue collar and
professional jobs follow similar patterns.
Table 1: Job Type (Percentage) * does not add to 100%, some small categories missing
Low wage: $7 an hour or less Low/mod wage: $9 or less, but usually 5-9
Mod wage: $7-9 High wage: over $9
| Job Characteristics | Job 1 | Job 2 | Job 3 |
| Low end service (sales, cashier, restaurant work) part time, no health insurance | 14% | 15% | 11% |
| Low end service, full time, no health insurance | 17% | 15% | 12% |
| Low end service, full time, with health insurance | 4% | 4% | 2% |
| Clerical, part time, low wages, no health insurance | 4% | 2% | 1% |
| Clerical, full time, low wages, no health insurance | 3% | 4% | 6% |
| Clerical, full time, low wages, health insurance | 3% | 3% | 1% |
| Clerical, full time, moderate wages, no health insurance | 1% | 1% | 3% |
| Clerical, full time, good wages, health insurance | 2% | 7% | 6% |
| Semi-skilled helping professions (nursing assistant, childcare, teacher's aid), low wages, part time, no health insurance | 3% | 7% | 6% |
| Semi-skilled helping professions, low wage, full time, no health insurance | 5% | 3% | 4% |
| Semi-skilled helping professions, low/moderate wages, full time, health insurance | 3% | 3% | 3% |
| Blue collar, low wages, part time, no health insurance | 3% | 1% | 2% |
| Blue collar, low wages, part time, health insurance | 2% | 3% | 4% |
| Blue collar, low wages. full time, no health insurance | 7% | 6% | 5% |
| Blue collar, high wages, full time, health insurance | 2% | 5% | 6% |
| Professional, low/moderate wages, part time, no health insurance | 1% | 1% | 1% |
| Professional, low/moderate wages, full time, no health insurance | 0 | 1% | 4% |
| Professional/high sales, moderate/high wages, full time, health insurance | 1% | 1% | 2% |
| Other/hairstyling, full time | 3% | 2% | 4% |
| Other | 18% | 12% | 18% |
Work Characteristics and Reasons for Leaving Work: Looking at length of time that people
stayed on jobs and reasons for leaving teaches much about the relationship between the quality of
work and employment patterns.
People leave part-time jobs with poor pay and no benefits quickly. For the first job, 50 percent
of the people in part-time, low end service jobs were gone in under six months. Forty-percent of
the people in part-time, low-wage, no benefit clerical positions, 68 percent of the lowest nursing
assistant jobs and 68 percent of the entry-level blue collar workers left in the same time period.
The same pattern persists for jobs two and three for these categories. Percentages are also high
for low-wage, full time jobs without health insurance. Twenty-five percent of those in full time
low end service, 38 percent in low-wage, no benefits clerical and 29 percent in similar blue collar
positions left before six months were up. Patterns for these jobs also persist over time.
People stay in low-wage jobs with health insurance only slightly longer. Thirty-six percent of
those in full-time low end service positions without health insurance and 33 percent of those in
comparable blue collar jobs stayed between six months and one year.
For jobs which may prove entry level to a field but which do not offer good wages and working
conditions, many people stay for one or two years. For example, someone may take a full time
job which does not pay enough or offer health insurance as an entry-level position in order to get
experience in the field but only stay a year or two before moving on to a better job. Fifty-four
percent of those in full time, low wage helping professions without benefits, 38 percent in low
wage, full time clerical jobs and the majority in the entry-level professional positions left in this
time period. The same patterns hold true for second jobs.
People stay in good paying jobs with benefits for a long time. For first jobs, 75 percent of people
in the best clerical jobs stayed more than five years. This was also true for 60 percent of those in
good blue collar jobs and 68 percent of those in professional jobs. People in the better nursing
assistant positions changed jobs more often, but tended to stay significant periods of time. One
third had been in their first job two to five years and another third more than five years. The
same patterns hold for second and third jobs.
Reasons for leaving jobs also reflect working conditions. Thirty-six percent left in under one year
because jobs were temporary or part-time. Thirty-eight percent of those who stayed in a job for
less than one year left due to childcare problems, in part due to hours on these jobs. Thirty-one
percent left due to maternity. However, it is important to note that the highest percentages who
left due to pregnancy were in good paying, women's jobs.
People also left low end jobs in short periods of time to go to school. For first jobs, the 12
percent who only stayed at a job for a few months before going to school represent students who
worked for the summer before entering adult training programs. After second and third jobs,
this trend reflects people who realize that they need to get additional training in order to succeed
in a field. One-third of the people in low-wage helping professions or professional jobs left
second and third jobs for this reason.
People also leave entry-level or stable positions for a better job. Twenty-five percent of those in
low wage, full time clerical positions with health insurance left for a better job. The same was
true for people in the best low end service (18%) and nursing assistant (22%) positions.
Changes in the economy send people in the best jobs looking for work. Seventy-five percent of
those in the best clerical positions and 50 percent of those in high paid, blue collar jobs, and one-third of the professionals left their first job because they were laid off or the business closed.
The same pattern continues for second and third jobs.
Demographic Composites and Work and Training Patterns
Race, gender, age and education all play a role in the career and training paths of people in this
study. However, these characteristics do not define a person by themselves. People with
different combinations of background characteristics may live very different lives than others of
the same race or sex. For example, an African American woman with a high school diploma
may fare differently from another African American woman who did not complete high school.
In order to understand the relationship between these various factors, the study population was
divided into groups using a statistical technique called cluster analysis. This technique groups
people together based on ways they are most similar. Using clusters shows patterns in a group of
people which may not show up if the researcher just guessed about what factors made people
most similar. For example, if I had assumed that all African Americans with a high school
diploma who were not married would be different from whites with a high school diploma who
had married, I would have missed both the differences among African American women and the
similarities along other key demographic characteristics among people from both groups.
Cluster analysis shows these similarities between people with several kinds of characteristics
which may effect their work and training experience like age, race, education, marital status,
working status or welfare or number of children.
For this study, I ran two types of cluster analysis and compared different groups from both types
of clusters on their work and training experience. These clusters show patterns within the
population, but no cluster includes only people with a particular set of characteristics. For
example, a cluster may include mostly people of color, but also have a few whites who have
similar characteristics. After looking at these two clusters, I discovered that race/nationality,
gender and education together seemed to account for significant differences. I then created a
further set of clusters which simply grouped all the people with the same three characteristics in
one group. This final set of clusters only includes people with a given set of characteristics. For
example, all white women with a diploma are in one cluster and all African American women
with a diploma are in another cluster.
The first cluster analysis included information on participant's age, gender, race, immigrant
status, whether or not they had a diploma, whether or not they had ever married, and currently
working or on welfare. The following clusters emerged from this analysis:
1 Married people of color, 36 to 45 (74 cases): 87% Black, 22.7% Latino, 7.2% Asian, 84.7%
Female, 51.4% ever married, 56.7% currently married, 100% 36 to 45, 82% welfare, 8.2% job
2 Under 21 year old women, no diploma (26 cases): 69% Black, 19% white, 7.7% Latino,
92.3% female, 92% never married, 100% under 21, 80% on welfare, 12% family support, 46.2%
diploma
3 Immigrant, married, people of color, 22 to 35, diploma (22 cases): 59% Black, 27.3%
Latino, 9.1% Asian, 63.6% female, 100% currently married, 100% 22 to 35, 85.7% welfare,
9.5% job, 36.4% immigrant, 90.9% diploma
4 19% white, female, working age, no diploma (77 cases): 72% Black, 18.7% white, 6.7%
Latino, 89.5 % female, 76.6% never married, 88.6% 22 to 35, 8.6% 36 to 45, 2.9% over 45, 91%
welfare, 6% job, 3.9% diploma
5 Black, never married, 14% working, diploma(125 cases): 79.8% Black, 14% white, 85.5%
female, 76.2% never married, 4.2% under 21 82.2% 22 to 35, 13.6% over 45, 78.9% on welfare,
14% job, 5.3% family support, 7% immigrant, 96% diploma
6 white, half male, half married, 36 and up, 17% working, diploma(14 cases): 100% white,
50% female, 50% currently married (all married at some point), 8.3% 22 to 35, 83.3% 36 to 45,
8.3% over 45, 75% welfare, 16.7% job, 8.3% family support, 7.7% immigrant, 100% diploma
The second cluster analysis included the same information except age. This analysis organized
the population differently. It brought together all of the African American women who were
working and the white women. It also rearranged the male population and some other groups.
1 Black women, never married, no diploma, welfare (85 cases): 76.5% Black,, 14.8% white
4.9% Latino, 95% Female, 23.8% ever married, 56% currently married, 95.2% welfare
2 Black/Latino, 75% female, married, diploma, on welfare (40 cases): 72.5% Black, 22.5%
Latino, 75% female, 84% married, 89% on welfare, 13.5% immigrant, 97.5% diploma
3 Immigrant, married, white/Asian, diploma (14 cases): 78.6% white, 21.4% Asian, all male,
100% curr married, 100% welfare, 53.8% immigrant, 85.7% diploma
4 Black female, never married, diploma, welfare (135 cases): 79.7% Black, 11.4% white,
7.3% Latino, 90 % female, 78% never married, 100% welfare, 7.8% job, 97% diploma
5 Black, never married, working,74% female, 66% diploma(50 cases): 95.9% Black, 4%
Asian, 73.5% female, 77.5% never married, none on welfare, 68% job, 12% family support, 20%
other means of financial support, 66% diploma
6 white women, half welfare, half family support, diploma(14 cases): 100% white, 100%
female, 63.6% currently married (all married at some point), 50% welfare, 8.3% job, 41.7%
family support, 8.3% immigrant, 100% diploma
There was significant overlap between the two cluster analysis, but not complete overlap for any
one cluster:
Comparison of second cluster analysis with clusters including age:
Second Cluster Number 1 (4 65.9%, 2 15%, 1 12.9%)
(1st cluster in brackets) 2 (1 30%, 3 45%)
3 (6 42.9%, 5 21.4%)
4 (5 60%, 1 25.9%)
5 (1 28%, 4 28%, 5 44%)
6 (5 57.1%, 6 42.9%)
Finally, I combined participants into groups which exclusively included people from different
races/nationalities, gender and diploma. Latinos and Asians were combined given small numbers
and similar patterns for earlier analysis. The number of people in each of these groups was as
follows:
1 Black women, no diploma (75 cases) 23.9%
2 Black women, diploma (133) 39.3%
3 Black men, no diploma (4) 1.3%
4 Black men, diploma (25) 8%
5 White women, no diploma (11) 3.5%
6 White women, diploma (23) 7.3%
7 White men, no diploma (4) 1.3%
8 White men, diploma (12) 3.8%
9 Latina women, no diploma (6) 1.9%
10 Latina/ Asian women, diploma (14) 4.5%
11 Latino men, no diploma (2) .6%
12 Latino/Asian men, diploma (5) 1.6%
None of these clusters by themselves explained the work and training experience of the people in
this study. However, by looking at the affects on training and work experience of all of the
cluster analysis together revealed a complex set of patterns. These patterns are described next.
Work experience varied dramatically across race, nationality, gender, education and marital
status. In general, people fell into the following categories:
Limited or no work experience: In addition to the 13 percent of the study population who had
never held a job, another 10 percent had only held one job for less than a year, adding up to 23
percent with limited work experience. This group tended to have been on welfare longer: 71
percent had been on welfare 3 years or more.
There was no significant difference across the individual variables of race, nationality or having
completed a high school diploma. However, the majority of the people in this category were
women who had never married. Latinas were the largest percentage who had never worked at 18
percent. Using the clusters for education, race and diploma, 25 percent of the African American
women without a diploma, 24 percent of the African American women with a diploma, 27
percent of white women without a diploma, 33 percent of the Latina women without a diploma,
21 percent of the Latina/Asian women with a diploma, 16 percent of African American men with
a diploma and 9 percent of white women with a diploma fell in this category. Looking at the
more complex cluster analysis revealed that age was not a factor, but the majority of people in
this groups fell into two of the demographic clusters which did not measure age. Forty two
percent of the never worked population were in cluster 4 (Black women, never married, diploma,
on welfare) and another 28 percent were in cluster 1 (Black women, never married, no diploma,
welfare).
These analysis reveal two things about the population with no real work history. First, while not
having a high school diploma did not help in finding employment, the majority of people who
did not finish high school in fact have worked. The group which has never worked seems to
have in common family, neighborhood or personal characteristics which lead them to be isolated
from employment networks and to have other issues which keep them from working.
Anthropological research on low income populations show that family ties often place women in
a dense web of obligations to family and friends. Since work is often unavailable and unreliable,
these kinship obligations become more important than work or school. On the other hand, family
networks could provide strong supports for these women if some family members are allowed
stay home to care for children and continue receiving support from public assistance. If they
live in neighborhoods where there are few jobs and their family and friends who do work are not
in good jobs, they also may not have someone to refer them to work. African American men
lacking similar social ties may be in the same position. Latinas with limited language ability may
have even more difficulty (Stack 1974, Oliker 1995, Wilson 1996).
Low skill workers: This group includes men and women with no high school diploma or limited skills who go from low skilled job to low skilled job. Job categories include mostly low end service and semi-skilled helping professions. This group primarily includes women of color.
They tend to move between similar kinds of jobs. Seventy-one percent of these people went
from one low end service job to a second on just like it. Those who began their careers in part-time, no benefit helping professions were not able to move up into better positions in this field.
Eighty-seven percent of those who were employed in the good nursing assistant positions had a
high school diploma. Some were able to move from low end service into helping professions
later in their careers. But they mostly ended up in the part-time, low wage positions without
health insurance.
This group works more jobs and leaves them more quickly: between 40 percent and 68 percent of
people in various kinds of low wage jobs left by 6 months and the majority changed jobs within
2 years. They left primarily because of the nature of the work (low wages, hours, part-time): 38
percent who left first jobs in six months or less left because the job was temporary or part-time,
child care or family problems. Forty-one percent left second jobs for the same reason.
Displaced workers: Between 60 percent and 75 percent held their first job for 5 years or more.
This group fell into two categories. The first included white and African American men who
worked in good paying factory or blue collar jobs. A few professionals also fell into this
category. These people stayed in this kind of employment until they were displaced. After
losing good, stable jobs, they either could not find work or ended up in service or clerical
positions which would end. This group also tends to be older than other participants and include
more people who were currently married.
The second group included African American, Latina and white women with high school
diplomas and high quality training, often clerical training. These women would start their
careers in entry level clerical or professional training and move up in their fields. The Latinas in
this population tended to be established on the mainland, not new immigrant populations. This
group also lost their jobs due to downsizing and were not able to find comparable work.
Migrants and refugees: Newcomers to the U.S. included mostly Puerto Rican citizens and
refugees. These groups fell into two subgroups: highly educated and skilled people who needed
to learn the language and gain U.S. mainland experience and people with low skills and limited
education.
The refugee population could be further divided into several groups. Eastern Europeans were
mostly recent emigres focused on regaining professional qualifications. Southeast Asians had
sometimes been in the U.S. a little longer and may have been placed into factory work in the U.S.
When those jobs ended, they were unable to find new employment.
The migrant population also falls into several categories. Many have worked a number of jobs.
Education level may be low, leading people to move from low skill job to low skill job. Some
Puerto Ricans may have not found work on the mainland due to language barriers and lack of
connections to employers. Others may work in the car repair shops or stores in the Latino
community. These jobs often pay poorly and are part-time with no benefits.
Training and Work Experience Patterns:
People with various demographic and work experience characteristics used training in very different ways. This section discusses those patterns. Analysis included comparing the demographic clusters to general work and training experience and looking at time lines for work and training experience for people with different backgrounds. Findings for this section can be summarized as follows:
Pre-Training Work Experience: The time line analysis counted the number of jobs that people
held before and after attending training programs. Sixty-two percent of the study population had
not worked before they started their first training program. This statistic is no surprise as many
people go directly from high school to training. Eighteen percent had held one job before
training, 11 percent two jobs and 10 percent three or more.
Whites entered training much later in their career cycle. Forty-five percent of whites had held
two jobs or more before going to a training program. These figures mostly reflect dislocated
workers who had gone into work directly out of high school. They enter training only after their
stable work ends or they are placed in mandatory community service or job development
programs by the Department of Public Welfare when they run through their unemployment and
end up on welfare.
Latinos show similar experience, with 23 percent having held two or more jobs before entering
training. However, for Latinos, jobs held before training are often less stable positions in the
service sector. In comparison, the majority of African Americans had only worked at one job
before entering training.
Forty-seven percent of the study population had not worked before going to their second training
program. Nineteen percent had held one job, 14 percent two and 20 percent three or more. Here,
a pattern begins to emerge regarding the way that race/nationality and education influence the
role of training in a career path. Percentages for African Americans are about even across the
number of pre-training jobs. For whites, however, 19 percent had worked at three or more jobs.
This trend suggests that some whites return to training late in their careers as retraining after a
layoff or to get skills after working for many years.
People who went to more than two training programs were labeled "repeaters". Their experience
is discussed in detail below. Looking at the number of jobs before training programs three and
four reveals that this population falls into two groups. Sixty-three percent of the study
population in a third training program and 72 percent for the fourth program had never held a job
before training. This shows a signficant population in this multiple training experience which are
simply drifting from program to program.
However, 20 percent for both the third and fourth program had worked three or more jobs before
returning to school. This second population includes both low skilled workers who keep
returning to training unsuccessfully in order to better their condition and displaced workers who
had used training successfully early in their careers but now found themselves displaced as the
economy changed.
People often chose training based on their previous work experience. People who had been in a
clerical job often went into clerical training. The same was true of people in nursing assistant
positions. This study showed that previous related work experience made a profound difference
in ability to get a good job in that field after completing training. This was particularly true of
clerical positions. Many women went from a part-time clerical position into a clerical training
program. These women were most likely to get the good paying, full time clerical jobs with
health insurance. While the pattern was not as strong for the semi-skilled helping professions,
the same trend appeared in those jobs as well.
Post Training Work Experience: Data on working after training also shows several patterns.
There is a significant group of people in this study who did not find work after training. Forty-three percent never worked after their first program. Figures go up after additional programs.
Fifty-six percent had not held a job after a second program, 83 percent after a third and 94
percent after a fourth. Since the populations in third and fourth programs largely include people
who have never worked, this trend shows that additional training was not able to overcome the
multiple barriers which kept these populations out of work. For the displaced workers in these
programs, the types of training appears largely unsuccessful at returning them to the workforce.
Training did make a significant difference for a large number of people. Thirty-five percent of
the population had held two or more jobs after a first training program and 14 percent held two or
more jobs after a second program. Thirty percent had held one job after a second program.
Eight percent had held one job after a third program and another 8 percent four or more. The
pattern for people who had gone through four training programs indicates that most of these
people were low skill workers who went from job to job after training. The six percent who had
worked after a fourth training program held three and four jobs.
Length of Time Between Training Programs: People did not go directly from training
program to training program. Overall, the average gap between starting training program one
and two was slightly over five years, three and a half years between programs two and three and
three years between programs three and four. For those who had worked, there are significant
gaps between starting a job and returning to school. For both first and second training programs,
people started work approximately five years before going back to training. These figures
represent both people who worked for several years and then upgraded skills and those who only
held a job for a short time and then waited awhile before going back to school. For working
populations, the gap between starting work and entering program three was an average of six and
a half years. Before program four, the average gap is over nine years. Looking only at repeaters
who had worked, the gap between starting their last job and entering their current training
program is over nine years, compared to six and a half years for those who had gone to two
programs or less. This shows dislocated workers returning to school late in their careers.
Patterns are very different for repeaters who had not worked. Gaps between programs two and
three average one year and less than a year between programs three and four. A large part of this
trend is due to the PIC training strategy of sending people first to basic skills training and then
immediately on to skills training. It also shows that this population needs basic skills
remediation before entering training, suggesting multiple barriers to employment which will
make it even more difficult for this population to ever find and keep work.
Many people in the training and advocacy communities promote training in construction trades
or blue collar skilled trades as a mechanism to get women and people of color into stable, good
paying jobs. This study suggests that this strategy does not work. People who graduated from
tuition based skills training programs went nowhere. Those who graduated from PIC funded
programs were able to get into low paid blue collar jobs, but were not able to move into the
stable, good paying positions.
The reasons for this probably reflect the different hiring and referral mechanisms for blue collar
employment. Scholars show that these kinds of "primary sector" jobs are found through friends
and family (Gordon, Edwards and Reich 1982). Before affirmative action, they were often
closed to people of color. In the last three decades, African Americans and established Latinos
have also moved into these good jobs. The fact that no-one went from a skills training program
into these kinds of jobs suggests that training does not substitute for access to appropriate
networks for primary sector employment. All blue collar work was also greatly affected by the
fact that stable, large manufacturing enterprises continue to leave Philadelphia in large numbers.
Given an even smaller pool of good, primary sector jobs to go around, it is even harder for
newcomers lacking social ties and previous work experience to get them.
While construction jobs involve work in many different kinds and sizes of companies, the best
paying union jobs often involve the same kinds of social networks as a basis for getting in to
good jobs as working for General Motors. Scholarly research also shows that sexism plays a key
role in hiring decisions in both construction and primary sector blue collar employment (Gordon,
Edwards and Reich 1982).
The difference between high paying blue collar employment and other kinds of work is also
illustrated by the training choices of people who leave these kinds of jobs due to downsizing.
None of the people in these positions went into another kind of blue collar skills training.
Instead, they went to a scattershot of nursing assistant, short-term clerical and other unrelated
programs. Very few of these displaced workers found jobs out of this training.
Relationship of training to previous work experience and demographic characteristics:
Repeaters: Eleven percent of the study population took more than two training programs. This group included only African Americans, whites and established Latinos. African Americans take the most training programs. Repeaters fell into two subpopulations:
Low skill workers: People with limited work experience or without a high school diploma
sought training in semi-skilled helping professions, clerical or blue collar trades. People without
high school diplomas primarily sought training in semi-skilled helping professions. Training
often led back to similar employment or into jobs which required some skills like nursing
assistant but which were often part-time and paid badly. People without high school diplomas
seldom were able to translate training into good jobs: 86 percent of the people who found
moderate wage, full time nursing assistant positions with health care had a high school diploma.
People who use training as a step up in the labor market. This group primarily includes
people of color, primarily African Americans and a few established Latino/as. Most people in
this group are women. These people have finished high school and may have some previous
related work experience before entering training. They take 6 to 12 month training programs,
often clerical, which eventually translate into good paying jobs.
This group often shows a progression in employment. Training programs, both free and tuition
based, place their clients without previous work experience into entry-level clerical jobs. These
people move up into the best paying clerical positions. Those who had been employed in either
good clerical jobs or entry-level clerical positions before training moved into the best paying
clerical positions.
Some return to college or another related skills training program to get additional training in their
field. Those who were dislocated workers were able to find clerical work again after longer term
training, but often in lower paying jobs or jobs without health insurance. These downgrades
reflect a combination of shifts in the economy and the fact that short placement goals pressure
people into lower paid jobs.
Displaced workers: In general, training for displaced workers did little to return them to stable
employment. Particularly for people previously employed in factory or skilled trades, training
had nothing to do with their previous work experience. The qualitative companion studies to the
social network research showed that displaced workers in large companies were either sent to the
PIC for retraining or guided toward similar programs through organizations hired to help laid off
workers adjust. People in companies too small to offer these services relied on friends and
family or information in the media about "growth" industries to look for training for new career
paths. Many people thought that learning "computers" or "getting in to health care" would
provide them with stable work again. Their training choices reflected the menu available
through the training provider network and/or programs that claimed to offer entre into the
growth industries. However, since training was short term, the health care programs were
primarily for semi-skilled helping professions and "computers" clerical or data entry training.
Retraining often meant work at much lower wage rates without benefits. Many found training
through a social service agency. Displaced workers fell into two subgroups.
The Training Track for African Americans and Latinos:
One of the questions behind this study came from observations that many more African
Americans took training programs than other groups and that this training often did not lead to
stable employment. The analysis of the social network study questionnaires shows that African
Americans do in fact go to training programs much more often than other groups. They are also
much more likely to go to vocationally focused programs than any other group. These
differences reflect two factors in this community. Many African Americans have observed
correctly that education is often a path to economic and social advancement. Like other
populations moving out of low skill or blue collar work, the first generation to go to post high
school training focuses primarily on skills programs which lead to specific kinds of jobs. Ability
to translate training into stable employment depended in part on basic skills.
The second factor leading to a training track involved the information available in the African
American community about jobs and training. Due to ongoing discrimination and the tendency
of the working and middle class to move away from poor, segregated neighborhoods, many inner
city African Americans lack ties to people who have graduated from good training programs
who could provide them with first hand information, as well as ties to good jobs through friends
and family with stable employment (Wilson 1996). This lack of social networks combines with
the fact that this population is much more often connected to the Department of Public Welfare
or social service agency programs which steer them toward certain kinds of training. Since
social networks play a profound role for everyone in their career and training paths, this study
focused on this factor.
The experience of Latinos born and raised on the mainland looked very similar to that of African
Americans. This fact reflects that the socio-economic backgrounds, education in public schools,
social networks and neighborhood conditions are very similar for these two groups.
Social Networks, Career and Training Paths
Participants were asked how they found training programs and jobs. This information looked for
several kinds of resources:
Social networks had a profound affect on getting people into training programs. Access to friends
and family with good resources also made a difference in finding work for people in blue collar
jobs and professional occupations. For people without ties to people in good jobs, background
characteristics of solid basic education mattered more in finding employment. Training
programs were best able to place people with good basic credentials.
Training: Overall, participant heard about training programs in the following way:
Family and Friends 31%
DPW or employment office 16%
Social service agency/PIC 22%
School 10%
TV or radio advertisement 20%
Jobs: Overall resources for jobs also stressed primary social networks:
Friends/family 65%
Training program or DPW 17%
Newspaper 27%
Walked in 38%
Unemployment office 4%
School 3%
Temp agency 1%
There was no difference between men and women in resources used to find either training or
employment. People with low education tended to rely more heavily than others on friends and
family. Race and nationality did show significant differences only for people referred to training
programs by agencies. Twenty-three percent of African Americans, 23 percent of Latinos and 20
percent of Asians found programs through agencies compared to 9 percent for whites. These
referral sources mean different things for Asians than for African Americans and Latinos. Asians
who get to training programs through agencies are usually refugees who are resettled through a
social service process which includes ESL training and skills retraining if appropriate. Many of
the whites who got to training through agencies are also refugees from Eastern Europe and the
Soviet Union. Latinos and African Americans find agency help because these populations
qualify for these services their incomes are low enough to qualify for these "means tested"
programs and because these populations have well established patterns of using social service
agencies and churches as resources.
Demographic Patterns and Social Networks: The various types of people used resources in
different ways. Patterns were as follows:
The Training Track and Social Networks: Social resources made a highly significant
difference in tracking people into training. People tended to rely on the same resource to find
training for first and second programs. Fifty-eight percent who found their first program through
friends and family used this resource to find a second program. Sixty-six percent who found
their first program through DPW found their second through either DPW or an agency. This
pattern reflects initial DPW referrals which lead people into a PIC training system where the
agency where they completed their first program will refer them to a second. People who went
to an agency to find a training program once used an agency again 73 percent of the time.
People quickly learned that media advertisements were not the best means to find training,
however. Only 26 percent used the media as a resource a second time. A full 48 percent of the
people who found their first program through the media went through either DPW or an agency
to find their second training experience. Schools also were used for referrals only once.
Repeaters: Use of official networks and the media made a significant difference for repeaters.
Eighty percent of repeaters who found their first training program through DPW or an agency
found the second in the same way. In comparison, only 50 percent of non-repeaters found their
second training program this way. The difference is even more dramatic for those who found
training through the media. A full 60 percent of repeaters used advertisements a second time
compared to 14 percent for non-repeaters. After two programs found in this way, these people
were usually referred to the free system by DPW or an agency. Many of those who used media
repeatedly had never worked.
Never Worked: Similar patterns appear for others who have never worked. Again, friends and
family are primary resources. The never worked population seems to be referred to training
programs twice by DPW, compared to populations which found jobs which progress from
finding training through DPW to an agency. This trend suggests that the never worked group are
less likely to progress through the training system and end up at DPW starting over again.
Agency referrals are used by both populations. Half of those who never worked used the media
twice compared to less than 23 percent who found work.
Social Networks and Tuition Based vs Free Programs: Various social networks led people
into very different kinds of training experience, which in turn, either allowed their training
experience to serve as a step up into the stable labor market or become a training track leading
nowhere. This was particularly true for the choice of free or trade school specific skills
programs. The first chart shows the kinds of resources used to enter training programs. Table 2: How Participant Found First Training Program
The pie charts show the kinds of resources used to enter training programs for job specific skills
programs. These charts show that people in free programs find them through referral from
official helping networks and friends and family. The same patterns hold for both first and
second programs. On the other hand, first trade school programs are found primarily through
recruiters or friends and family. For second programs, referral comes most often from a
advertisements. Social networks provide different kinds of information. Family and friends who
have connections to training programs or jobs which provide quality training or pay well are an
invaluable resources. On the other hand, family and friends who are themselves low skill
workers with limited success with training could provide bad advice. Employees of official
organizations or non-profit agencies can also provide good or bad advice. Individuals have no
independent way to evaluate the quality of a program found through an advertisement. This
pattern suggests that people in first trade school programs may either be getting good advice or
being lured to bad programs through ads. These are people with limited resources who often
drop out of programs or find poor programs. Second programs show even more social insolation
as most of these people find programs through the media.
While not pictured here, patterns for GED and job preparation programs are similar to that for
free job specific skills programs. Forty-five percent found the first training program through
friends and family, but only 13 percent for the second program. The Department of Public
Welfare referred 23 percent of the people in first programs and 44 percent for second programs.
Second programs may include more mandatory training. Social service agencies provided
resources for 13 percent for first programs and 26 percent for second programs. People found
colleges primarily through their schools (39%) and friends and family (17%).
This next chart looks more closely at the resources that participants in different kinds of
programs used to find their first training program. The same patterns hold for the second
program. Patterns for free programs are the same for both clerical and nursing assistant
programs: approximately even numbers of people are finding their programs through family and
friends or organizations. On the other hand, free skilled trade programs are found primarily
through agencies. An overwhelming number of tuition based nursing assistant programs are
found through advertisements. Tuition based skilled trades are recommended through friends
and family as well as found through the media. Tuition based clerical programs are found
through all three sources.
Training Quality and Social Networks: The quality of training can be measured first by whether
people finish the program. Recall that drop out rates are very high for the tuition based
programs, but very low for the free programs except skilled trades. These statistics suggest that
people entering most of the free programs are satisfied enough with their programs to complete
them and have the basic educational skills and motivation to complete the program. Low drop
out rates may also be due in part to the fact that many of the free programs offer participants case
management to help them handle life problems which may interfere with completing their
training. Many of the trade schools do not offer this service. In terms of social networks, these
differences imply that information from someone who has direct knowledge of a program but no
vested interest in that program is more likely to give advice on the best program for that
individual. On the other hand, the recruiters and media advertisements only show the school's
view of their program.
Given that the drop out rate for free skilled trades programs is 22 percent and that over fifty
percent of people in those programs are referred by an agency or the Unemployment office or
Department of Public Welfare, these data suggest that evaluation for this population is not as
good. Since the skilled trade programs serve mostly men, this suggests that the mechanisms
evaluating men for training are not as good as those for women.
GED and job preparation programs reflect another set of factors. Much of the scholarly literature
on adult basic education reports high drop out rates (for example Herr and Halpern 1991, 1994).
The twenty-one percent drop out rate for first GED programs in this study is actually
comparatively low. The ethnographic companion study suggests that people drop out of adult
basic education programs because they are frustrated with their pace through these programs and
their continued need for them. Many people in GED programs dropped out because they disliked
schooling and these programs too often remind them of these bad experiences. Many would
prefer to be doing something more directly related to getting a job and drop out because they do
not easily see the relationship between slow pace toward basic skills and eventual employment.
In this case, the agencies which are the second largest referral source for this type of program
may in fact be mandating programs or steering people who request skills training toward basic
skills.
Quality of the program can partly be measured by whether or not graduates get related jobs. As
reported in earlier sections, graduates of nursing assistant positions had variable success finding
related employment depending on the type of training and their background characteristics.
Clerical program graduates of both free and tuition based programs were able to find work.
Skilled trade programs were less successful at placing their graduates into good programs with a
career ladder. However, it is important to note that programs can not be held completely
accountable for results because the basic skills that a person brings into a program powerfully
effect the kind of training that they get and their ability to make use of it. Attitude or family,
health and substance abuse problems also affect who finds work after a program and who does
not. Even the best training programs can also not control the economy or the contracting
conditions which may impact on their service strategies. Those contract provisions will be
discussed in the next section.
Social Networks and the Role of Training Programs in Finding Work
This section looks at the social networks that participants used to find jobs out of training
programs. As with training programs, friends and family were the most frequent resources used
by most people in finding jobs. In general, the least educated populations and migrants and
refugees used friends and family more often than other methods to find work. People who found
a job through "walking in" tended to be me more isolated in terms of personal resources: the
same people who found training through TV or other advertisements were more likely to find
jobs by walking in. For example, 33 percent of the people who found work after a tuition based
nursing assistant program "walked in" to that employer. People who graduated from training
programs and college were more likely to use the newspaper than less educated people.
These charts show the way that people found work after their first and second training programs.
College graduates depend most often on friends and family. Part of this trend is due to the fact
that community colleges ordinarily do not have active placement offices for their graduates.
Friends and family for people with the basic education to finish college may also have better
resources than for other populations.
Placement through a direct referral from a training agency happened most often for free job
specific skills programs. For first programs, 25 percent of placements were through program job
developers and 30 percent were placed directly by the program out of a second training
experience. Tuition based job specific skills programs do a much poorer job of actively placing
their participants. Only 15 percent reported finding work through their training program for the
first program and none for the second. These figures suggest that the tuition based programs put
much less emphasis on placement than the government funded programs. Government funded
GED and job placement programs only succeeded in finding participants jobs for six percent of
their participants for the first program, but reached 30 percent for a second program. Part of the
difference here is accounted for by the fact that more second training experiences in this category
were job readiness programs. In first programs, 35 percent of participants in this category were in
job preparation and placement programs compared to 62 percent for the second program. Even
with this difference in focus, the job preparation programs were not the most frequent source for
employment for their participants. Only 25 percent of the people in job preparation in program
two found their jobs through the training program.
Performance Based Contracts versus Tuition Payments in Program
Outcomes
Most of the training documented in this study occurred during the time that the PIC funded
training for agencies without a commonwealth set tuition rate as performance based cost
reimbursement contracts with a hold back of 30 percent pending job placement. Agencies
working under contract with the PIC had to pay close attention to meeting their placement goals
in order to receive full funding for their contract. Contracting pressures clearly affect placement
strategies for PIC funded programs. Programs had 90 days to find their participants work which
paid $6.00 an hour, full time, preferably with health insurance. Given the labor market, agencies
often found themselves choosing between placing people into part-time, training related
employment or full time unrelated jobs with similarly poor wages and working conditions.
For many programs working with participants with limited previous work experience, and
working against the cultural prejudices of moving inner-city people of color into "mainstream"
jobs, their graduates too often found the same kinds of employment that they were attempting to
escape. This pattern is most acute for the nursing assistant programs. Fifty percent of the people
in part-time nursing assistant positions without health insurance came out of free nursing
assistant programs. Twenty-one percent after program one and 44 percent after program two of
the people in full-time low end service jobs graduated from free nursing assistant programs.
Forty percent of the people in low paid, part time blue collar jobs came out of free nursing
assistant programs. This pattern suggests that the programs found nursing assistant positions,
either full or part time, for as many graduates as possible. Since they encouraged their
participants to find jobs in order to meet placement goals, those who could not find related work
returned to low end service positions or blue collar jobs through their existing networks. The
pressure for quick placement would encourage this trend. The fact that most agencies could not
afford to work with former clients after the 90 day placement period was up also meant that
people who did not find related jobs were left to their limited social networks after failing to find
work through agency job developers.
Participants with better preparation were placed in better jobs. Eighty-seven percent of the
people in well paid, full time nursing assistant positions had high school diplomas. Qualitative
data suggests that this finding is due to the fact that employers are more willing to take a risk on
someone with basic educational credentials. However, participants without high school diplomas
placed into lower wage or part-time nursing assistant programs out of free programs were able to
move up into the moderate wage jobs. Only 68 percent of the people in these good nursing
assistant programs as their second job out of a free training program had a high school diploma.
This suggests that placing people without high school in a part-time nursing assistant job may be
a better long term strategy than placement into unrelated positions.
The tuition based nursing assistant programs often led nowhere. Twenty-three percent of the
people in part-time, low-end service jobs came out of these programs and few were in nursing
assistant jobs. In some cases, trade school program graduates could move into nursing care, but
into the worst positions. Thirty-eight percent of the people in part-time, low wage, no health
insurance nursing assistant positions as a second job came out of the paid programs.
Both free and tuition based clerical programs were often able to serve as career paths for their
participants. Starting with people with better basic skills, they were able to place participants
into entry-level, full time jobs without health insurance. Many of these people were able to move
up into better jobs. The same pattern was true for the greater than six month tuition based
programs and free programs, showing the variability in quality and participant characteristics
among tuition based programs.
Overall, none of these programs could escape the overwhelming importance of the local socio-economic system and their participants place within it. Previous work experience, and the social
characteristics of attitude, speech patterns and neighborhood based social networks played a
profound role in participant outcomes. Over 70 percent of the people in low end service jobs
before entering training returned to similar positions. People in clerical positions before going to
school were able to upgrade their employment in that field. Nursing assistants remained in
similar helping professions.
These findings on job placements show performance based contracts do impact on the
performance of contractors. This study suggests that the free programs put much more energy
into placements. They also suggest that letting the market rule the training choices for low wage
workers often leads to disaster as people lacking the social networks and other tools to evaluate
training programs often simply end up with huge debts which place further burden on
government guarantee associations. But limiting non-profit agency payments and demanding
short-term placement goals for populations needing long term assistance and follow-up also has
unintended negative effects. The types of performance goals stipulated in these JTPA and JOBS
funded programs work for populations prepared to benefit from training, but do not help those
most needing to escape poverty.
The social network study found several clear patterns for employment and training for workers
from disadvantaged populations. The first clear finding is that the population on public
assistance is extremely diverse. The dramatic differences in career and training paths across
race/nationality and gender reveals that patterns of discrimination, as well as socialization toward
certain kinds of employment, persists in the 1990s. Part of this is due to the extreme segregation
of Philadelphia and the poor quality of its public schools. The fact that many of the people who
had never worked or were in low end service jobs despite training had finished high school
shows the quality of the education for many low wage workers in Philadelphia.
Part of this difference in career and training paths is also due to social networks and agency
mandates. Friends and family can only provide advice based on their own experience and world
view. Public and non-profit agencies are constrained by the available pool of programs and their
own training about appropriate career and training paths for the people they serve. Both agency
rules and the training program system is further shaped by commonwealth and federal policy or
foundation trends. This means that most people go to short term programs or trade schools
funded through student loans because this is what is readily available through existing systems.
The study also highlights the fact that in many cases the population working in low wage or even
working class jobs is interchangeable with the population on welfare. Given that 94 percent of
the people in this study had been on public assistance at some time in their lives, and that 87
percent had worked for wages as adults, policy makers and program developers can not assumes
that this population simply needs training for appropriate jobs or work experience.
Since most of the people in this study had gone through multiple training programs, it is also not
safe to assume that training alone is the answer. The second major finding is that appropriate
training, combined with appropriate work experience and solid basic skills leads to long term,
stable employment. This finding suggests that the best strategy for all programs is to create
experiences that combine all three factors. The best option is not work or training, but work and
training.
The choice of type of training for different individuals reflects a combination of individual
choice and available options. A statistical study like the social network study can not easily
measure the process of personal choice. Nor did this study collect information which could
evaluate such intangible factors as attitude, ability to work in different kinds of environments or
the family or personal problems which can shape the ability of any individual to successfully
complete training or find and keep a good paying job. However, this study does point to some
differences that program designers should keep in mind.
People without a high school diploma need to finish a GED as a first step toward stable
employment. This study shows that those without high school diplomas have limited training
options and even more limited employment opportunities. However, given the difficulties in
completing a GED, combining GED preparation with appropriate work experience and related
skills training may be a better option than simply expecting people to complete a GED before
entering training or finding a good job.
The high drop out rate for college shows that many of the people in this population are not
prepared to succeed in a college environment. The fact that many college graduates found work
in fields unrelated to their training also shows that college may not be the best choice for all low
income people. However, college did make a significant difference for some people in this
population. This suggests that college should be one option in a menu of training programs.
Job specific skills programs helped some people but not others. Four factors in combination
influenced the ability of a job specific skills program to provide a career path for its students: 1)
the quality of the program, 2) the basic skills and life conditions of its students, 3) the nature of
the work available for people with that particular type of training and 4) the ability of the
program to successfully place its students in those jobs. Since these four factors can not be easily
measured, finding the "best" programs requires thoughtfully looking at the ways that these
factors work together for different kinds of employment.
Given that many people who found good jobs relied on friends and family as a referral source to
work, developing mentoring programs which connect people who lack these ties to people
working in that field is one strategy which could help some people move from welfare to work.
Mentoring or buddy programs provide training program students with direct advice on many
important aspects of work like appropriate dress, workplace behavior, how to write an
appropriate resume and cover letter as well as insider connections to employers.
This study shows the importance of policy makers designing more flexible and long term
outcome measures geared toward multiple populations. They also need to take into account the
consequences of funding requirements and performance standards on program design. Most
policy is based on assumptions that contractors are simply making a product, be clear enough
about the design specifications and monitor ability to meet those standards, and the required
results will follow. But people are not widgets. They come with a variety of background
characteristics which affect their progress through programs and have the free will to choose if,
when and how they use their training. Policy needs to reflect this dynamic understanding of
social structure. Performance standards need to reflect evaluation procedure that accounts for
long term outcomes and uses more sophisticated models with different characteristics for
populations with varying education and work experience. Contracts which hold back funds
based on short term, rigid performance standards are particularly corrosive in giving agencies the
funding they need in order to meet the intent of policy. Instead, these contracting mechanisms
should be replaced with incentive structures developed to reflect appropriate strategies for
differing populations. These types of mechanisms are best developed with agency
subcontractors.
This report recognizes that the PIC has little control over performance standards. However, there
is room for creativity if the agency is able to develop a longer term strategy for contract
performance which views participation in PIC programs as only one step in a career path for
participants. The system in place when this study was conducted did seem to work reasonably
well for the population with a high school diploma and the social skills to succeed in the
workplace. It is far less successful with people who lack this basic credential or who have
graduated from high school but have poor math and reading skills. A large part of the problem is
the nature of the work available for this population rather than the training itself. Nursing
assistant and foodservice positions with poor working conditions do not provide the money to get
a family out of poverty or the schedule to allow a single mother to manage her household and
children.
If these programs are the only option, it would serve the PIC well to take conscious efforts to
work with the health care community and appropriate non-profits and foundations in order to
ensure that people placed into this kind of work receive longer term social supports and
encouragement to finish a GED and move on to other education systems. Given that career paths
depend on gaining experience in that particular field, greater emphasis should be placed on
finding related employment in an appropriately supportive environment rather than placement at
a given wage rate. It is also essential that the training community find ways to work with the
business community to find ways to make sure that low wage workers get insurance for the entire
family after the year of free medicaid is over. It is also essential that subsidized childcare and
social supports remain available for those who need them. This implies that the PIC and other
agencies need to play an active role in efforts to address the quality of life issues which make it
difficult for people to juggle work and family.
Finally, this study primarily includes people who are in training either because they lost jobs or
are dissatisfied with their previous employment. The fact that dislocated workers with good
work histories, solid basic skills and sometimes good training could not find adequate
employment again suggests that even the best training system will not work without connections
to employers in a vibrant economy offering stable employment. Workforce development systems
are only as good as the economies in which they function. Effective policy and program
development must pay equal attention to local economic development.
Gordon, David M., Richard Edwards and Michael Reich. (1982). Segmented Work, Divided
Workers: The historical transformation of labor in the United States. Cambridge,
Massachusetts: Cambridge University Press.
Herr, Toby and Robert Halpern. (1994). Lessons from Project Match for Welfare Reform.
Chicago: Project Match.
Herr, Toby and Robert Halpern with Aimee Conrad. (1991). Changing What Counts: Re-thinking
the Journey Out of Welfare. Chicago: Project Match.
Oliker, Stacey J.(1995). Work Commitment and Constraint Among Mother on Workfare.
Journal of Contemporary Ethnography, 24,2, 165-194.
Stack, Carol. (1974). All Our Kin: Strategies for Survival in a Black Community. New York:
Harper and Row.
Wilson, William Julius.(1996). When Work Disappears. New York: Alfred A. Knopf.
Appendix A: Differences Between Agencies Participating in the Study:
This study drew participants from eight PIC funded training programs and community college.
Percentages in each program were as follows:
Institute for the Study of Civic Values unemployed parents initiative: 8%
Community Women's Education Project: 10%
People's Emergency Center: 13%
Community College of Philadelphia: 12%
Arbor-Cite 20%
OIC 8%
1199c 4%
ISCV program prep project 5%
PIC Upfront 17%
PHDC 2%
For the purposes of the analysis, the ISCV Alternative Work Experience Program was divided
into two groups. The unemployed parents program served an older population, largely male and
more white which was oriented toward finding immediate employment. The program
preparation group included people who had recently graduated from PIC funded feeders and
were waiting to enter job specific skills programs. This population was classified as a feeder. In
the more general analysis, the ISCV unemployed parents program was labeled "mandatory
community service", the second ISCV program and CWEP were grouped together as feeders, the
PEC population was labeled as shelter because they differed significantly from other feeder
programs, Arbor and 1199c were job specific skills programs primarily serving women and OIC
and PHDC were job specific skills programs serving men. The PIC upfront program was
classified as mandatory job development.
General Demographic Differences
In general, the PEC population was the youngest, female, the least educated, and had more
limited work experience. The ISCV AWEP-UP population showed the opposite trend with the
oldest, better educated and more male population.
Age: PEC had the youngest group with 26 percent between 18 and 21, 40 percent under 25 and
another 36 percent between 25 and 36. ISCV had most of the over 45 population at 32 percent of
that group over 45. This percentage was much higher than for any other program. The bulk of
the population for other programs were in the prime working ages of 22 to 45. Arbor and OIC
had a higher percentage in the older end of that range. Both had 36 percent of their populations
at 36 to 45.
Race:
Mandatory and Feeder Programs:
ISCV AWEP-UP had the most white (Black 25%, white 42.9%, Hispanic 17.9%, Asian 10.7%)
PIC Upfront (Black 82.4%, white 11.8%, Hispanic 3.9%)
CWEP also much more white(Black 42.9%, white 37.1%, Hispanic 17.9%)
PEC (Black 90.5%, white 7.1%)
ISCV Program Preparation (100% Black)
CCP (Black 55.3%, white 26.3%, Hispanic 7.9%, Asian 5.3%)
Job Specific Skills
All mostly Black:
Arbor 85.5%
OIC 88.5%
1199C 92.9%
PHDC 100%
Gender:
ISCV 67.9% male, PIC Upfront 8.9% male
All feeders 89-100% female
JSS female 100% female
JSS male OIC 64% male, PHDC 50% male
CCP 76.9% female
Ever Married:
ISCV 93%
PIC upfront 41%
CWEP 38%
PEC 11%
ISCV program prep 29%
Arbor 28%
1199c 21%
OIC 31%
PHDC 50%
CCP 45%
Immigrant Most were in the ISCV AWEP-UP program(43.5%) and CCP (16.2%). The rest
were 8% or less.
Education: ISCV had most college (36%), next PIC upfront (13%), not counting CCP (53%)
Finish high school:
ISCV 75%
PIC Upfront 61%,
CWEP 55%
ISCV program prep 53%
PEC 20%
Arbor 65%
1199c 54%
OIC 29%
PHDC 33%
CCP 79%
Finish GED: The people in technical training programs and CCP had finished GEDs
Diploma (GED or Finish High School):
ISCV AWEP-UP 68% (the difference here is caused by missing data for finish high school)
PIC Upfront 66%
CWEP 60%
ISCV program prep 78%
PEC 35%
Arbor 84%
1199c 71%
OIC 46%
PHDC 67%
CCP 95%
Attended Training Programs: There was no significant difference in number of programs taken
or completion rates among programs.
Current Student Loans:
ISCV AWEP-UP 17%
PIC upfront 65%
CWEP 50%
PEC 46%
ISCV Program Prep 50%
Arbor 81%
1199c 46%
OIC, PHDC samples too small to measure
CCP 62%
Types of Training Programs Taken Before Current Program:
Data processing(p000): CWEP(45.5%), PEC (39%), Arbor (71%), CCP(23%),
1199c(21.4%), CCP (23%)
Skilled trades(p000): Men's programs ISCV (27.3%), CCP (23.3%), OIC (39%),
PHDC (60%)
College (p000) :ISCV (31%), CCP (58%)
Nursing Asst(p000): 1199c (91%), CWEP (23%), PEC (16%),OIC(23%),
PIC Upfront (19%), PHDC (40% , but only 2 people)
Business(p001): ISCV (17.4%), PEC (47.8%), CCP (21.9%), Arbor (18%)
OIC (18%), ISCV Program Prep (20%)
Foodservice(p000): OIC (39.1)
Completed program:
Yes: Mandatory, good completion rates: ISCV(50 % completed something 22.7% 1
program, 13.6% 2, 13.6% 3) , Upfront (65.8 % completed something 32.6% 1 program,
23.3% 2)
Feeder less good: CWEP (32 % completed 1 program), PEC (54.8 % completed
something 51.6% 1 program, 3.2% 2), ISCV Program Prep (64.3 % completed something
42.9% 1 program, 21.4% 2)
JSS women, more programs completed: Arbor (55.6 % completed something 41.3% 1
program, 12.7% 2, 1.6% 3), 1199c (64.3 % completed something 42.9% 1 program,
21.4% 2)
JSS male, similar: OIC (54.2 % completed something 29.2% 1 program, 25% 2)PHDC
sample too small
CCP most programs: (70 % completed something 46.7% 1 program, 16.7% 2, 6.7% 3)
Did not complete:
Feeder population more likely to drop out: CWEP (72.7 % not completed something 50%
1 program, 13.6% 2, one each 3 and 4) PEC (64.8 % not completed something 48.4% 1
program, 6.5% 2), ISCV Program Prep (35.7 % not completed something 28.6% 1
program, 7.1% 2)
CCP also high (33.3 % not completed something 26.7% 1 program, 6.7% 2)
JSS Programs: between 20 and 25%, mostly 1 program
Number of jobs: Largest percentages never worked are Arbor (23.9%) and 1199c (30.8%).
Programs with older people have more jobs.
How Hear about program:
Friends and Family (p01):Low for ISCV (27.3), CCP (16.7%), Arbor
(22.6%),1199c(28.6%), PIC upfront (25.6%), rest between 40% and 50%
DPW(p005): PIC Upfront (30%), PHDC (60%), rest under 20%
Agency(p000): Female JSS high: Arbor 32.2%, 1199c 71.4%
CCP 26.7%, rest 15% or less
School(p03):CCP 29%, rest less than 15%
Media no difference
Military only ISCV
Types of Jobs Held: There was no statistically significant difference for many types of
employment. Significant differences occurred for the following jobs:
Cashier (p002): Feeder lots of experience: CWEP (31% yes, 20.7% 1, 6.9%2, 3.4% 3),
PEC (63% yes, 34.3% 1, 5.7%2, 2.9% 3), ISCV Program Prep (86.7% yes, 46.7%
1,26.7%2, 13.3% 3)
Mandatory less: ISCV AWEP-UP (14.8%), PIC Upfront (23.5%)
JSS women most: Arbor (54.9% yes, 37.3% 1, 5.9%2, 5.9% 3, 5.9% 4), 1199c (44.4%
yes, 22.2% 1, 22.2%2)
JSS male, less likely: OIC(16.7%), PHDC (40%)
CCP (25.7%)
Nursing Assistant(p02): CWEP (24%), OIC (21%), 1199c(33.3% but only 3), PHDC
20% 1 person)
Had a Job which Provided Health Insurance (p03): high is CCP (54% ), PIC upfront (47%),
PHDC (60%), rest 40% or less
Reasons for leaving work: Only Business closed significant: ISCV (59.3%), Arbor (31.4%),
OIC (29.2%), ISCV Program Prep (40%), PIC upfront (43.1%), PHDC (60%), rest less than 20%
Reason left last job (p001):
ISCV AWEP-UP 40% laid off
CWEP: 19% childcare, laid off
PEC: 34% childcare, 12.5% temp/part time and other
CCP: 21.7% still there, 13.6% laid off, went to school
Arbor: 31.8% maternity, 27% child care, 13.6% laid off
OIC: 21% temp/part time, 15.8% laid off
1199c: 33% maternity
ISCV Program Prep: No pattern
PIC-upfront: 28.6% temp/part time 21.4% laid off
PHDC: No Pattern
Length of last job(p02):
Under 1 year: CWEP 51.7, PEC 70.6%, Arbor 46.6%, OIC 52.4%, ISCV Program Prep 60%,
Pic-Upfront 42.8%
Over 2 years: ISCV 48%, CWEP 31%, CCP 45.4%, OIC 33.3%, 1199c 42.9%, PIC upfront
28.6%, PHDC 40%
SOCIAL NETWORKS, CAREER AND TRAINING PATHS
FOR PARTICIPANTS IN EDUCATION AND TRAINING PROGRAMS
Technical Report
Jo Anne Schneider
May 1997
Report Prepared for the Philadelphia Private Industry Council
Acknowledgments
Research for this report was conducted under the auspices of the Institute for the Study of Civic
Values. I would like to thank the institute staff for their help, support and advice on this project.
Partial funding came from the Private Industry Council of Philadelphia and the 21st Century
League. I would also like to thank Robin Ebright, Melissa Smiley, Mathew Wickens and other
Bryn Mawr, Haverford and University of Pennsylvania students for their help in developing the
questionnaire and carrying out the research project. Finally, I wish to thank the agencies and
their students who participated in this study.
If you have questions about this report, please contact the author at:
Department of Sociology and Anthropology
University of Wisconsin-Parkside
900 Wood Road
P.O. Box 2000
Kenosha, WI 53141-2000
414-595-3418
joanne.schneider@uwp.edu
GED/Job Prep
JSS, Free
JSS, Cost
College
Friend/family
45%
37%
37%
17%
Welfare
23%
18%
3%
4%
Agency
13%
29%
6%
4%
School
0
2%
6%
39%
TV/Radio
13%
8%
41%
13% 
