Wednesday, Nov 25
	PSID Guide | Data and Documentation | Publications and Conferences | Child Development Supplement | Search CDS Guide | Time Diary | Bibliography | Contact Us

Household Pretest 1 - June, 1996

Household Pretest 1 was not a normal pretest, because its major purpose was not primarily to test the questions. The items in the questionnaires for this study have been used extensively in other studies. While we did test the questions, the main goal of Pretest 1 was to test the procedures for this study. Specifically, we were trying to find ways to help minimize the burden on both the interviewer and respondent and to make things go as smoothly as possible to help achieve a high response rate. The pretest was used to find ways to improve the order of administration, to develop tools to help keep track of things, and to develop procedures for production data collection. The pretest consisted of 23 households with 45 children.

Teacher Pretest 1 - July, 1996

The respondents for Teacher Pretest 1 were teachers who were attending summer school classes at Teachers College, Columbia University. Presentations were made in four summer school classes, four professors were given flyers, and flyers were posted around the campus. In addition, to recruit more preschool teachers, local preschools were called. The results were as follows:

TABLE 8. Number of Interviews, Teacher Pretest

Type of school	# completed by mail	# completed by phone	# refused	Total complete
Elementary	3	2	1	6
Middle	3	2	0	5
Preschool	2	0	0	2
Totals	8	4	1	13

As a result of this pretest, we learned that the respondent burden was very high. The instruments were scaled back in length to help alleviate this burden. We also learned that only three of the responding teachers had any concerns regarding confidentiality. The majority of the teachers felt comfortable completing the questionnaire.

Pretest 2 - October - November, 1996

The number of household interviews completed in Pretest 2 was 51. For the following tables data from up to 49 households (with 85 children) were used; the information from 2 households was not complete.

N=19 households (30 children) from households that were included from the immigrant PSID pretest.

N= 9 households (15 children) from households that were included from the Core PSID pretest.

N=21 households (40 children) from Achild-only households (where no PSID interview was taken).

In preproduction, up to three children could be selected from each sample household. The following table shows the percentage of households where one, two, and three children were included in the study.

Table 9. Number of Household Interviews, Pretest 2

# of Children	% of Households	# HH	# Children
1 child	36.7%	18	18
2 children	53.1%	26	52
3 children	10.2%	5	15
Total	100.0%	49	85

The next table shows the distribution of the children's ages for Pretest 2.

Table 10. Distribution of Children's Ages, Pretest 2

Age of Children	% of Children	(N)
0-2 years	18.8%	16
3-5 years	22.4%	19
6-9 years	37.6%	32
10-12 years	17.6%	15
DK Age	3.5%	3
Total	99.9%	85

The following table shows the average length of each interviewer-administered instrument in Pretest 2.

Table 11. Length of Interviews, Pretest 2

Questionnaire	Avg Minutes	(N)	Range
Child Interview	13.7	58	2-42
Prim. Caregiver-Child	42.0	69	5-81
Father Out-Child	24	1
Father Out-HH	18	1

The next table shows the eligibility rates and response rates for each questionnaire.

Table 12. Eligibility Rates, Pretest 2

Questionnaire	# Eligible	% Eligible*	(N)	Response Rate (% of eligibles)
Primary Caregiver-Child	85	100%	84	98.8%
Primary Caregiver-HH	49	100%	44	89.9%
Other Caregiver-Child	66	76.5% of kids	65	98.5%
Other Caregiver-HH	39	77.6% of HH	38	97.4%
Father Out-Child	15	17.6% of kids	4	26.7%**
Father Out-HH	8	25% of HH	2	12.5%**
Elementary Teacher	55	64.7% of kids	24	43.6%
Elementary Diary	49	57.6% of kids	16	32.7%
Preschool Teacher	13	15.3% of kids	6	46.2%
Preschool Diary	13	15.3% of kids	7	53.8%
Elementary Admin	56***	65.9% of kids	22	39.3%
Preschool Admin	13	15.3% of kids	3	23.1%

*Base N = 49 Households; 85 children
**2 refused to give information; 2 did not know where father was.
***Elementary Admin includes duplicate administrators (1 admin for 2 kids); Don't know (DK) why the number eligible is different from elementary teachers.

The following table summarizes the contact attempts for the Pretest 2 households.

Table 13. Contacts, Pretest 2

Household Contacts	Total contacts	In person visits
Average number of contacts	5.4	2.5
Minimum	1	1
Maximum	15	6
(N)	34	34

The second test of the teacher questionnaires occurred concurrently with the full pretest of the household questionnaires. After the household portion of the interview process was completed, the field interviewer asked for the respondent's signed permission form allowing us to contact the child's teacher and administrator. The interviewer then mailed self-administered questionnaires to both the teacher and administrator. In Pretest 2, the Survey Research Center followed up to increase the response rate. This consisted of mailing a reminder postcard to the teacher and administrator one week after the field interviewer mailed the questionnaire packets. If the teacher or administrator had not returned the questionnaire one week after the postcard had been sent, the research technicians began to conduct reminder calling. Up to six calls total could be placed to each teacher and administrator.

Table 14 presents the results of Pretest 2. There were 85 eligible children in this pretest. Sixty-eight children were either in elementary/middle school or preschool/daycare. Teachers of middle school children did not receive the time diary because the children change classes and they would not know what the children do all day. Instead, the middle school children filled out the time diaries as part of the household interview.

TABLE 14. RESPONSE RATES, PRETEST 2

Questionnaire Type	# of questionnaires sent	# of questionnaires received	Response Rate
Elementary/Middle School Teacher	55	28	50.9%
Elementary School Time Diary	49	20	40.8%
Preschool/Daycare Teacher	13	7	53.8%
Preschool/Daycare Time Diary	13	7	53.8%
Elementary/Middle School Administrator	56	23	41.1%
Preschool/Daycare Administrator	13	3	23.1%
Total Questionnaires sent & received	199 sent	88 received	44%

Reminder/thank you postcards were sent to all teachers and administrators. Forty-five schools were called. The average number of calls per school was 2.5 calls. The minimum number of calls to a school was zero and the maximum number of calls was six calls.

As a result of this pretest, the teacher time diaries were combined into the teacher questionnaires so that the teachers would be more willing to complete both instruments. A question was added to the elementary/middle school teacher questionnaire to skip middle school teachers over the time diary portion of the questionnaire. A separate questionnaire for home-based care providers or "baby-sitters" was developed for use in production to capture this type of care arrangement as well. A corresponding administrator questionnaire was not needed for this type of care arrangement.

The respondent burden still proved to be high in the second pretest. In response, the instruments were further scaled back to help alleviate respondent burden and the questionnaires were tested to determine if they could be administered over the phone. It was thought that administering the questionnaire over the phone would be easier for the respondent. The questionnaires were successfully completed over the phone but were even more time-consuming. The time diary in particular was cumbersome when administered over the phone. As a result, we decided not to change the data collection method. The procedure for the teacher and administrator questionnaires remained the same for production.

Results of Pretest 2
The following procedures changed after Pretest 2 for production data collection.

• Pretest 2 interviewed up to three eligible children in a household; production is limited to two children per household.
• The teacher time diaries were incorporated into the teacher questionnaire for production; they were separate instruments for pretest.
• We added a home-based care questionnaire for babysitters for production, based on our pretest experience.
• The child time diaries were reformatted for production.
• The self-administered questionnaires were reformatted for production.
• The partner of the primary caregiver was redefined as other caregiver for production to include grandmothers of sample children and other adults living in the target child's home.