Community Employment Innovation Project (CEIP)
Detailed information for 2004 to 2007 (54 month follow-up)
The Community Employment Innovation Project (CEIP) is a research demonstration project that is testing a new approach to the income security system for individuals on Employment Insurance (EI) or Income Assistance (IA). Individuals who agreed to take part in the study had a chance to be offered an opportunity to participate in community employment for three years. The CEIP is conducted in the Regional Municipality of Cape Breton, Nova Scotia.
Data release - November 24, 2008 (No public use microdata file was produced by Statistics Canada and data will not be made available through the Data Liberation Initiative (DLI).)
Human Resources and Social Development Canada (HRSDC) (formerly Human Resources Development Canada) funds the CEIP. The Social Research and Demonstration Corporation (SRDC) has overall responsibility for the project design, coordination, research and evaluation. The SRDC has contracted Statistics Canada (STC) (Special Surveys Division) to select a sample of potential participants and to undertake various data collection and administrative activities for this demonstration project.
The project will take eight years to complete. The gradual enrolment of volunteers for the study began in July 2000 and concluded in June 2002. In the first year, potential participants were selected from a population of current Employment Insurance recipients. The Nova Scotia Department of Community Services (DCS) supported the inclusion of Income Assistance recipients in the second year of this initiative. The project recruited 1,006 volunteers among eligible Employment Insurance recipients and 516 among Income Assistance recipients.
Reference period: 2002
Collection period: 2000 to 2007
- Employment insurance, social assistance and other transfers
- Workplace organization, innovation, performance
Data sources and methodology
The target population for the Community Employment Innovation Project consisted of Employment Insurance and Income Assistance recipients residing in the Cape Breton Regional Municipality at a specific point in time. The EI and IA target population had distinct characteristics.
The sampling unit for EI recipients was the EI claim and the observation unit was the person who filed the claim. Every four weeks, between July 2000 and April 2002, Statistics Canada created a population file of people eligible to participate in the project using administrative data supplied by Human Resources and Social Development Canada. The target population consisted of EI claimants, at least 18 years of age whose EI claim met the following criteria:
(1) claimant must have received at least $1 in regular benefits in the four-week sampling reference period,
(2) claimant must not be participating in another EI work initiative;
(3) there were at least 12 weeks remaining from a total entitlement of at least 22 weeks; and
(4) the number of weeks paid on the claim must be between 10 and 13 weeks inclusive at the last week processed.
The CEIP also targeted employable Income Assistance recipients aged 19 years and over. The employability criterion was defined by the Nova Scotia Department of Community Services and the target population determined from their administrative files based on monthly IA payments for the period covering May 2001 to March 2002.
All survey instruments were developed in a joint effort between the Social Research and Demonstration Corporation and Statistics Canada. When available, questions used in other Statistics Canada surveys, such as the Survey of Labour and Income Dynamics (SLID, record number 3889), the Labour Force Survey (LFS, record number 3701) and the Canada Survey of Giving, Volunteering and Participating (CSGVP, record number 4430), were used. Other questions dealt with topics unique to the CEIP and were developed and tested by the SRDC, the sponsor. Questionnaires were reviewed internally and tested for readability and duration by experienced STC personnel, including a senior interviewer.
This is a sample survey with a longitudinal design.
The sample design is specific to the "stream", Employment Insurance recipients (EI) or Income Assistance beneficiaries (IA).
Individuals selected by Statistics Canada received an invitation to attend an information session conducted by CEIP staff in Sydney, Nova Scotia. Attendees interested in joining the study were required to complete an enrolment form consisting of a short survey (the baseline) and an informed consent form.
Human Resources Development Canada (HRDC) provided a data file of EI beneficiaries every four weeks between July 2000 and April 2002. The information provided on the file was limited to that which was necessary to apply the selection criteria for CEIP and to allow Statistics Canada to make contact with eligible participants. The survey population was stratified by age (under 30 versus 30+). The sample was allocated so that the younger group would not be over-represented in the group of people enrolling in the project. The number of individuals selected (simple random sampling) each cycle was adjusted at various times to accommodate for changes in take-up-rate and the availability of community-based projects. Over the enrolment period, 5,980 eligible EI recipients were selected by Statistics Canada and sent an invitation to attend an information session. Of this number, 1,622 attended a session and 1,006 volunteered to join the random trial study.
The selection of individuals from the IA caseload required a two-step process. Each month, the Nova Scotia Department of Social Services (DCS) provided Statistics Canada with a file of current IA recipients. The file was void of personal identifiers but contained the information required to apply the eligibility criteria. Statistics Canada would select a sample and communicate the results of the selection to DCS. This enabled DCS to establish the first contact with their clients and asking those interested to identify themselves to Statistics Canada by returning a signed card especially designed for this purpose. Individuals had three weeks to return the card in order to be included in the eligible population. Between June 2001 and May 2002, DCS mailed 3,232 letters to potential eligible IA participants. A total of 1,014 of them returned their signed card to Statistics Canada. From this list, Statistics Canada randomly selected 804 individuals and sent them an invitation to attend an information session at the CEIP Office; 557 attended and 506 volunteered to join the random trial study.
Eligible EI or IA recipients could only be selected once.
Responding to this survey is voluntary.
Data are collected directly from survey respondents.
Data were collected using a computer-assisted telephone interviewing application. The respondent's address and phone number were available from their enrolment form. At this time respondents were also asked to provide two contact names and phone numbers. Interviewers were asked to conduct extensive tracing to contact the respondents. A number of study volunteers from the Income Assistance sample did not have their own phone so interviews were conducted using one of the contact phone numbers.
The sample was introduced gradually over the collection period. Each collection cycle was generally "allowed" three months in the field for completion. However, some units remained in the field for up to five months to complete an interview.
The vast majority of interviews were completed during the first month in the field (83%); an additional 12% were completed in the second month.
Administrative data sources - EI, IA, and CEIP program data - will also be used for estimating program impacts on individuals and for the benefit-cost analysis.
The CEIP survey data was processed with minimal editing. The main type of error treated after collection were errors in questionnaire flow, where questions which did not apply to the respondent were found to contain answers or conversely, when relevant questions were missed. A computer edit automatically eliminated superfluous data. For skips based on answered questions, all skipped questions are set to "Valid skip" (6, 96, 996, etc.). The "Not stated" reserved codes (9, 99, 999, etc.) are used to identify missing data resulting from problems with the computerized questionnaire or questions that were skipped because the respondent could not answer previous filter questions (Don't know or Refusal).
When using a computerized questionnaire, some editing is done directly at the time of the interview. Where the information entered is out of range (too large or small) of expected values, or inconsistent with the previous entries, the interviewer is prompted to seek confirmation from the respondent.
Variables such as date of birth, date of enrolment and even gender were often inconsistent from one source to the other (administrative versus the enrolment form completed by the study volunteer). These variables were carefully validated.
Pay frequency was edited in a very small number of cases using an outlier detection technique.
The relationship of household members was examined for internal consistency and with respect to Statistics Canada definitions. Editing was done with a very conservative approach. For instance, if two persons reported being spouses but answered "Single, never married" to the marital status question, the marital status was changed to "Married or common-law".
No imputation is done for this statistical program.
The estimation strategy consisted of the calculation of a baseline weight and a non-response adjustment factor.
The underlying assumption in the calculation of the baseline weight is that the target population for the CEIP consisted of all eligible individuals who would have been interested in taking part in the program had they been selected by Statistics Canada. Since the interest of eligible candidates was not known at the time of sampling, the target population can only be estimated using the "take-up" rate of those selected.
For the 5,980 Employment Insurance recipients who received an invitation to join the CEIP the baseline weight represents the inverse sampling ratio specific to the sampling cycles and to the age stratum (less than 30 years and 30 years and over). The sum of the weights equals the number of individuals on the survey frame (over 14,000). Of the 5,980 individuals who were selected, only 1,006 voluntarily enrolled in the study. The sum of the baseline weights for this group equals 2,300 an estimate of the number of eligible EI recipients who would have been interested in a full scale CEIP offer over the enrolment period.
The two-phase sampling approach used for the Income Assistance recipients imposed a slightly different definition of the target population for the CEIP. The sum of the baseline weights for this group equals 1,200 an estimate of the number of eligible IA recipients who would have been interested in a full scale CEIP offer over the enrolment period. In this case, the "interest" is measured by the number of Income Assistance recipients who returned signed cards to Statistics Canada.
The non-response adjustment for respondents recruited from the Employment Insurance sample takes random group (program or control), age (less than 30 years or 30 years and over), and gender into consideration. The adjustment factors range from 1.06679 to 1.23334 and are generally higher for females. For respondents recruited from the Income Assistance sample the non-response adjustment controls for random group, the family situation (family or single) and gender of the respondents. The adjustment factors range from 1 to 1.19175 and are generally higher for males.
Individual survey data were examined mainly for completeness and internal consistency of some related items. The largest number of internal inconsistencies is found in dates of absences from a job and in the social capital section of the questionnaire. Issues with questions on social capital were anticipated from the very beginning of collection based on interviewers' report.
Given its random trial design, CEIP data will be used to compare the outcome of program group members to that of control group. Users should be extremely cautious to use the data in any other context.
Statistics Canada is prohibited by law from releasing any information it collects which could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.
No public use microdata file will be produced by Statistics Canada and data will not be made available through the Data Liberation Initiative (DLI).
Revisions and seasonal adjustment
This methodology does not apply to this survey.