SPD Status Reports

Survey of Program Dynamics

Status Report

March 11, 1999

The SPD had a very interesting and productive year in 1998. During the year many decisions were made and the direction of the survey was established. During 1998 the following major decisions were made:

The household sample was set at 20,000 cases annually. We hope to interview approximately 18,000 cases each year.

The instrument/questionnaire design was set as follows:

1998 The basic core modules plus an adolescent Self Administered Questionnaire(SAQ).

1999 The basic core modules plus extended measure of child well-being modules.

2000 The basic core modules plus a child retrospective residential history module.

2001 The basic core modules plus an adolescent SAQ questionnaire.

2002 The basic core modules plus extended measure of child well-being modules.

Incentive payments were approved by The Office of Management and Budget (OMB) for all Type A Non Respondents from the 1998 SPD and all new Type A Non Respondents in the 1999 SPD.

An Attrition Study would be under taken to see if it is possible to bring SIPP attrited cases back into the SPD sample.

The type and number of files that SPD will create was set as follows:

Calendar Year Files

A limited edited calendar year file will be issued every year of data collection. The file will not contain any imputation but will give analysts an early look at the data. The first file, containing data from the 1997 SPD, was issued to the public on February 26, 1999.

Longitudinal Files

Issues Date: Content of Files:

December 1999 A longitudinal file covering the years 1992 through 1997 except 1995.⁽¹⁾

December 2001 A longitudinal file covering the years 1992 through 1999 except 1995.

December 2003 A longitudinal file covering the years 1992 through 2001 except 1995.

Listed below you will find more detailed information about the various activities that have been undertaken by the SPD survey in 1998 and planned activities for 1999.

Data Collection

Year	Instrument	Mode	Interviews	Type As ⁽²⁾	Type Bs ⁽³⁾	Type Ds ⁽⁴⁾	Type Cs ⁽⁵⁾	Response Rate
1997	Current Population Survey (CPS) March	Computer assisted Interview (CAI)	30,125	4,335	302	2,402	1,181	81.7%
1998	SPD	CAI	16,395	2,211	201	637	222	85.2%
1998	SPD-SAQ	Paper	3,259	2320	0	0	0	58.4%

Response rate	=	Interviews
		Interviews + Types As + Type Ds

A team is currently reviewing why the SAQ questionnaire had a low response rate and what needs to done to improve the response rate the next time the SAQ is scheduled for interviewing in 2001.

Subsampling For 1998 and beyond

Because of budget constraints, the SPD survey is only able to maintain a sample of 20,000 cases for data collections each year. To bring the 1997 SPD sample of 30,125 interviews down to the 20,000 level the following sampling scheme was used:

Household Type		Sampling Rate	1998 Sample
Less than 150% of poverty		100%	6,182
150% to 200% of poverty	With Children	100%	1,075
150% to 200% of poverty	Without Children	82%	1,461
Greater than 200% of poverty	With Children	90%	6,623
Greater than 200% of poverty	Without Children	27%	3,707
Institutionalized		27%	81

Sample Loss

The SPD inherited a 26.6 percent sample loss rate from the 1992 and 1993 SIPP samples. After two waves of SPD, the sample loss rate is 50 percent. Previous studies on SIPP sample loss have shown that the sample loss is not uniform. Households in and near poverty attrit at a higher rate than other households. Since poverty households are a key target population in the study of welfare reform, there is concern about nonresponse bias.

Sample Loss - An Average of the 1992 and 1993 SIPP Panels and SPD

Interview Wave	Interviewed HHs	Cumulative Sample Loss Rate (%)	Interview Wave	Interviewed HHs	Cumulative Sample Loss Rate (%)
1	39,446	8.8	7	36,133	23.6
2	37,936	14.4	8	35,761	25.1
3	37,882	16.3	9	35,291	26.6
4	37,477	18.1	10⁽⁶⁾	13,337	26.6
5	36,985	20.3	1997 SPD Bridge⁽⁷⁾	30,125	41.3
6	36,676	21.9	1998 SPD⁽⁸⁾	16,400	50.0

Please note that the SIPP and SPD attrition (sample loss) rate are computed based on the number of households eligible for wave 1 interviewing. Other large-scale longitudinal surveys compute their attrition rates based on the number of households completing a wave 1 interview. In order to understand the comparative response rates of the SPD and other major longitudinal surveys, the Census Bureau staff standardized cumulative response rates for the SPD, the Panel Study of Income Dynamics (PSID), and the National Longitudinal Survey of Youth (NLSY). The results indicate the standardized response rates (adjusted for respondents' mortality) for the entire survey periods are 64 percent for the NLSY, 50 percent for the SPD, and 39 percent for the PSID ⁽⁹⁾.

Incentives

The 1998 SPD survey did not use incentives. The 1997 SPD Survey conducted an incentive test. Its goal of the test was to determine the effect an incentive on bringing retired sample respondents back into a follow-up longitudinal survey. A subset of sample clusters containing low income households in the SPD were designated to receive a $20 voucher. Low income is defined as households that were at or below 150% of the poverty threshold based on previously collected SIPP data. The experimental group consisted of 10,683 households that were at or below the poverty threshold and their neighbors. The control group consisted of 3,343 households who were at or below 150% of their poverty threshold and neighbors. All other households were not eligible.

The incentive test showed that giving a $20 incentive to households has had a positive but not a significant effect on response rates for the experimental group overall. This was true by demographic characteristics, as well. Among the experimental group, the households having received and cashed vouchers had a statistically significantly higher response rate than the response rate of the households having received but not cashed or having not received vouchers⁽¹⁰⁾.

Attrition Study

We are currently conducting an exploratory study on SIPP nonrespondents from the 1992 and 1993 panels. The focus of this study is to measure the possible success rate of locating and interviewing SIPP sample persons that attrited during various periods of the panel. The 1992 and 1993 SIPP panels lost more than 1/4 of the sample by the end of their interview period. The sample of households currently in SPD may not be perfectly representative of the general population because only households interviewed in the last wave of the 1992 and 1993 panels were eligible for the SPD.

The attrition study consists of 360 noninterview (Type A and D) households assigned to 12 regional offices (ROs) with each region assigned approximately 30 cases. All the cases were low-income (200% of poverty threshold) at the time of their last SIPP interview. The cases were also divided between early and late attriters and evenly among three incentive test groups ($0, $50 and $100). The study is currently in the field and will run from mid-January to mid-March 1999.

The sample for the attrition study is person-based, similar to SIPP. All original sample persons (person living in household during wave 1 interview ) currently age 15 or older who were living in the household at the time of the last interview are eligible for interview and must be followed if they have moved. All persons currently living with an original sample person are listed on the control card and all persons 15 or older are eligible for interview.

The attrition study hopes to address the following questions:

1. Can the Regional Offices locate the 1992 and 1993 SIPP nonrespondents (Types A and D)?

2. How difficult and costly is the location process?

3. Once located, will the respondents be willing to complete an interview?

4. Does an incentive help this process and does one incentive amount appear more successful than another?

5. What would be the cost of bringing a selected sample from 1992 to 1993 SIPP nonrespondents into the SPD?

We are planning to release preliminary results from the study by late May 1999. If the study is successful, we will assess whether to institute the practice for the 2000 SPD.

Files

The Census Bureau released a simple calendar file of the 1997 SPD Survey. The file contains only a longitudinal weight for each original sample person. The file provides users with the ability to match SPD data back to the 1992 and 1993 SIPP Panel files.

The Census Bureau and the Social Security Administration have contracted with the Urban Institute to create a March CPS "look- alike" file of the 1992 and 1993 SIPP Panel Files for 1992, 1993 and 1994. These files were released in January 1999. The variables on this file are the bases for the SPD longitudinal files.

The first longitudinal data file for SPD using 1992-1997 data is planned for release in December 1999. The file format is person record and will consist only of persons selected for interviewing in 1998. The file will not contain data for 1995 since the 1992 SIPP did not dat for 1995 and the 1993 SIPP did not completely collect 1995 data.

1999 SURVEY

Data Collection Schedule -- The 1999 SPD survey is scheduled to run from April 29, 1999 through June 30, 1999.

Reinterview -- There will be a Quality Control (QC) reinterview, consisting of approximately 1,500 cases. The reinterview is expected to last about 10 minutes.

Instrument -- The chart below shows the various modules included in the 1999 instrument.

1999 SPD Modules

Section	Module	Section	Module
A	Household Roster		Child-Related Questions
	Adult Questions	P	School enrollment
B	Employment & Earnings	Q	Enrichment activities
C	Income sources	R	Children's disability
D	Independent/dependent Comparison	S	Child health care utilization
E	Amounts	T	Mother's Work Schedule
F	Eligibility and Assets	U	Child care
G	Vehicle Operating Expenses	V	Child support
H	Educational enrollment	W	Contact with absent parent
I	Work training	X	Extended Measures of Child Well-Being
J	Substance Abuse	Y	Positive Behavior/Social Competence
K	Functional limitations and disability	Z	Family Routines
L	Health care utilization		Self-administered Questions
M	Health insurance	AA	Conflict between parents
N	Food Expenditures	AB	Marital relationship and conflict
O	Food security	AC	Parental depression scale

2000 SURVEY

The 2000 SPD will include a new Child Retrospective Residential History module. The module is currently being developed and cognitively tested. The module will replace the extended measures of Child Well-Being Modules asked in 1999.

GENERAL ISSUES

SPD Website/DataFerrett

The SPD survey has a Website and utilizes the DataFerrett system to access and extract data for public use. The site contains an overview of SPD, tables created from the 1992 and 1993 SIPP and the 1997 SPD data, and has the 1997 SPD micro data.

Research and evaluation activities

The following research and evaluation activities are in progress:

1. Incentives: The Census Bureau was given approval from OMB to give a $40 incentive to all Type A households from the 1998 SPD and to all new Type A households during the 1999 SPD. The incentive is needed to help increase the response rate.

2. We are planning to apply cross-sectional weights to the 1997 Bridge data to compare it to the official CPS March income data to help assess the effects of nonresponse bias on the SPD estimates.

Contractor activities

There are four existing contracts funded by the SPD.

An interagency agreement with HHS to develop a profile of state welfare program rules, and to gauge the accuracy of substate welfare profiles (the subcontractor is the University of Wisconsin's Institute for Research on Poverty). Work is ongoing--no deliverables thus far.
An interagency agreement with HHS to fund $150,000 worth of research grants on SPD. The grantees will be selected in December 1999 by the University of Chicago-Northwestern University Poverty Research Center.
ChildTrends, Inc (CTI) continues to give advice on questionnaire design. A new agreement with CTI is being negotiated and will focus on providing research support as requested by Census Bureau on child related issues.
Iowa State University (ISU) has a small contact to develop a methodology for using the SPD questionnaire and a state supplemental sample to make state estimates on the effects of welfare reform. ISU is working on the final sampling/weighting report and hope to have the report released within the next month.

For further information on SPD operations issues, contact the SPD Branch at (301 763-1616 or dsd_survey.of.program.dynamics@census.gov.

The 1992 SIPP has a complete data set for only January 1995 and the 1993 SIPP has a complete data set for only January - September 1995.
Type A A household that is occupied by persons eligible for interview but an interview is not obtained.
Type B A household or person who are temporarily ineligible for interview because they are institutionalized.
Type D A household or person who have moved to an unknown address or have moved more than 100 miles from the nearest SIPP PSU.
Type C A household or person who can no longer participate in the survey because they deceased, moved out of the county, or armed forces members and not living with other sample persons.
A wave 10 interview was conducted for only 3 rotations of the 1992 panel. There was no wave 10 interview for the 1993 Panel.
Only those HHs interviewed in the last wave of the 1992 or 1993 panels were sent to the Field for the SPD Bridge.
Only, those HHs interviewed in the Bridge and selected during the subsampling were eligible for the 1998 SPD. The number 32,800 is derived based on the assumption that sample loss is uniformly distributed among the subsampled groups.
Memorandum for Daniel H. Weinberg from Donald J. Herandez dated January 28, 1999 titled Comparing Response Rates for SPD, PSID, and NLSY.
Memorandum for Karen E. King from Lieu Galvin dated January 21, 1999 titled Preliminary Evaluation of the Survey of Program Dynamics (SPD) Bridge Incentive Experiment by Demographic Characteristics.

SPD Introduction

SPD Overview

Analytic Uses of SPD Data

Contact:(dsd.survey.of.program.dynamics@census.gov)
URL: http://www.sipp.census.gov/spd/