TECHNICAL APPENDIX Methodology The sample consists of one percent of all claimants who established a new claim in local offices anywhere in the State during the 12-month period beginning in November 1963 and continuing through October 1964. They were selected on the basis of the terminal digits of their SSA numbers. Claimants filing under Federal programs (UCFE or UCX) were excluded from the sample, as were all claimants filing in other states on the basis of California wage credits. At the time he filed his new claim, each sample claimant was asked to answer questions about his personal characteristics and employment status, on a claimant questionnaire (copy attached). Upon completion, each questionnaire was reviewed for accuracy and completeness by a local office specialist familiar with the study requirements. At the end of each survey week, local offices transmitted completed survey questionnaires to Central Office. Here they were matched against accounting records which show base period employers and quarterly earnings (DE 429). Sample controls were maintained throughout the survey period, in order to minimize the proportion of claimants for whom information was not available. Obtaining Employer Information Employer letters requesting the weekly distribution of wages from the start of the claimant's base period up to the week of the claim were mailed to each of the claimant's employers (copy attached). Employers in the claimant's base period were identified from accounting records (DE 429). The claimant's lag period employers and earnings were also obtained from accounting documents (DE 507) as soon as they became available in the wage record file and letters were sent to any additional employers not already listed as base period employers. Controls on employer letters were maintained, and those employers who did not respond to the original request for information were sent second notices. Volume of Employer Letters Processed A total of 28,178 letters requesting weekly wage information was sent to California employers. An additional 4,558 "second requests" for information were sent to employers who did not respond within 60 days to the initial letter. Including replies to second requests, a total of 24,733 replies was received, for a return of 87.8 percent on the original letters. Replies to second requests totaled 2,216 for a 48.6 percent return. Farm employers received 1,225 initial requests. This accounts for only about 4.3 percent of the total, indicating how small a proportion of claimants have earnings in both subject employment and farm employment. The number of base period employers per claimant varied from zero to as many as 45. The average was three employers per claimant, if both the base period and lag period are considered. Data were reviewed and edited by Research and Statistics personnel, and coded schedules prepared for machine processing. Missing data on the distribution of earnings by weok were allocated whenever the resultant number of weeks was not likely to affect the claimant's eligibility under any of the alternative tests. Sample Parameters The theoretical number of claimants expected in the sample is one percent of new intrastate claims recorded by Central Office during the entire sample period............................ Sample claimants included in the Central Office document count of new claims, but subsequently removed from sample: Additional claim transfers ..... Classified within sample period, but date benefit year began not in sample year .... New intrastate claim transferred to interstate category Theoretical Number of Claimants Expected in Sample, Net Non-respondents (missing questionnaires) In addition to the 149 claimants removed from the sample, a considerable volume of questionnaires were excluded through the use of the sample controls. Most of these were for former military personnel (UCX claimants), 305, or former Federal employees (UCFE claimants), 162. An additional 39 questionnaires were excluded because they were for extended duration (ED) claimants, incorrect SSA numbers, or duplicates. Several individuals completed questionnaires but had no claim processed by Central Office. A few questionnaires were received from claimants who filed before the sample period commenced, or after it was completed. Unlike random sample error, which is subject to statistical measurement, sample bias cannot ordinarily be evaluated mathematically, although it has an effect upon the accuracy of the results. The largest source of bias in this experiment is that caused by missing questionnaires. There were 562 of these, which amounted to 5.8 percent of the sample. Since control cards and claims documents are available for these claimants, their actual effect on the accuracy of the survey can be determined for these characteristics. Percent change in eligibility under present law if incomplete and missing questionnaires were included is 1.0 percent, as shown in the following table: Liable interstate claims constitute about five percent of all claims filed against California, and they were deliberately excluded from the survey because of the difficulty of obtaining questionnaires, and getting them completed in a uniform manner, from claimants filing claims in other states. About 22 percent of the interstate claims filed against California earnings are invalid, so that the exclusion of these claims from the sample results in a slight understatement of the number of claimants ineligible under both the present and the alternative eligibility tests. Another source of bias is the errors of response in the information obtained from the sample claimants. The accuracy of answers to different parts of the claimant questionnaire vary. For example, there is a much higher level of accuracy from the response to questions on the claimant's sex, age, and education, than to questions regarding the claimant's work history. It is very difficult for many claimants, particularly those with irregular periods of employment, to reconstruct their work history for the three-year period requested by question 17 of the claimant questionnaire. For this report, the claimant's answers were accepted as given, providing they were complete and logical. Report tables based on question 17 (Tables 10 to 22) are therefore less reliable than are those for the remainder of the survey. Plans are under way for a validation study regarding this portion of the survey data. Comparisons will be made from work histories, Department records, and employer letters. In contrast to this, the level of accuracy for tables presenting data derived from Department records is very high. A ready comparison can be made between weekly wages as reported by the employer and quarterly wage records. Data concerning the critical question of eligibility under the various alternatives is very reliable. Spot checks were made on the accuracy of allocation of missing data by using completed schedules received from employers subsequent to the allocation. In nearly every case, the allocations were accurate to plus or minus one week of $20 earnings. Reliability of Sample Data The results obtained from a sample survey are likely to differ somewhat from those that would be obtained from another sample from the same population, or from those that would result from a complete count or census. These variations occur by chance when the sample is chosen by a random process. The range of sampling variation to be expected varies with the size of the sample in the group or subgroup being analyzed and is relatively larger for smaller sample sizes. The following table presents the approximate variability that may be expected for various sizes of groups or subgroups in this sample: Inter Subgroups with 375 claimants are subject to 10 percent sampling error. pretations based on subgroups smaller than this are less reliable and should be made with caution. The sampling variation for a subgroup of any size may be estimated by interpolation between the values given in the above table. CLAIMANT QUESTIONNAIRE You have been chosen to take part in a study of California Unemployment Local Office Number Interviewer's last nam PLEASE ANSWER THE FOLLOWING QUESTIONS CONCERNING YOU AND YOUR FAMILY 8. If you live with your family, how many persons are in your household? (including yourself) 9. If you live with your family, how many persons in household earn wages? 10. How many persons besides yourself can you claim as dependents for Federal Income tax purposes? (If you have no dependents, write 0). Number of persons Number of persons 11. Do you regularly pay more than half of your family's support? Yes No 12. Circle the highest grade (or year) of regular school you completed. ་ College Number of persons 1 2 3 4 or more 13. What is your usual occupation? 14. List any other occupations in which you were employed during the last two years.. Write name of occupation CODE Claimant do not write in this space ANSWER QUESTION 15b ONLY IF YOUR ANSWER TO EITHER OF THE ABOVE TWO QUESTIONS WAS "YES" 15b. Would you take a course of training if offered? Yes No YOU WORKED IN A JOB NOT COVERED BY UNEMPLOYMENT INSURANCE or 16. List non-covered and out-of-state employers for jobs held since July 1, 1962. Name of firm Street address City DE 3810 REV 1 (1-64) CLAIMANT QUESTIONNAIRE - HR 98 State of California Department of Employment 35-2770-69 42 |