SUPPORTING STATEMENT
Part B
Pretest of the Medical Expenditure Panel Survey
Quick Response Survey
January 12, 2012
Agency for Healthcare Research and Quality (AHRQ)
Table of contents
B. Collections of Information Employing Statistical Methods 3
1. Respondent universe and sampling methods 3
2. Information Collection Procedures 4
3. Methods to Maximize Response Rates 4
4. Tests of Procedures 5
5. Statistical Consultants 5
The Medical Expenditure Panel Survey (MEPS) is a household-based survey. Its sample is representative of the non-institutionalized civilian population of the United States. The MEPS sample is drawn from the larger nationally representative sample of households who participated in the prior year in the National Health Interview Survey (NHIS). The sample for the MEPS Quick Response Survey is a subsample of the cooperating households in one MEPS panel.
A key element of the sample design for this data collection is to select a sample that will permit the production of appropriate sample weights for analysis as quickly as possible following data collection. To achieve this goal, the sample will be selected as follows:
Select a sample of reporting units (RUs) from among the set of families in one MEPS panel (Panel 15) designated to receive weights for the full year 2010 consolidated file, based on the Panel 15 Rounds 1-3 weighting effort.
The pool for sampling will include responding RUs that were inscope as of December 31, 2010.
Student RUs, who are tied to their families for family estimation purposes, will not be eligible for sample selection.
Select the sample of RUs with probability proportional to size (pps), where the measure of size is the person-level weight of the reference person
The sampled RUs from round 3 will be linked to the corresponding RUs in Round 5. These Round 5 interviews will be fielded in early 2012.
The sample design uses pps sample selection with the sample weight as the measure of size. The reciprocal of the sample weight will represent the probability of selection of the RU for the QRS so that the design is self-weighting, removing or reducing the variation of the sample weights associated with the NHIS and MEPS oversampling. Appropriate sorting of the sample before sample selection will be done to reduce the effect of clustering on sample variances.
The number of RUs sampled will be expected to yield 1,500 completed QRS interviews after allowance for ineligibility and nonresponse in Rounds 4 and 5 as well as to the QRS. Planning incorporates the following assumptions:
Non-response, Round 4 3 percent
Non-response, Round 5 3 percent
Non-response, QRS 10 percent
The sample for the QRS will be selected, as described in Item 1, at the start of the final round of MEPS data collection for households in Panel 15. As these households complete their Round 5 interviews, all will be informed of their possible selection for the followup QRS. Those actually sampled for the QRS will be identified as their data are receipted, and, with appropriate locating information incorporated into a file that will be used to manage respondent contacts and direct the telephone interviewing. The sampled households will receive a letter informing them of their selection. Telephone contacts to conduct the QRS will follow the mailings. The data collection will be by computer-assisted telephone interviewing. Reports on interview production will be available on an ongoing basis; data from completed interviews will be available for preliminary processing shortly after the interview has been completed.
The instrument for this collection contains a maximum of 17 questions for a single respondent. No questions require respondents to consult records or to review visual aids to assist in responding. No questions call for multiple or extended responses. The instrument is designed to impose minimal burden on respondents.
Once an acceptable household respondent has been contacted, the data collection can be completed in a single, brief telephone call. No followup contacts are planned. Quality control of the interviewing will be implemented through silent monitoring, with a standard minimum of 10 percent of each interviewer’s work monitored by quality control staff.
This effort will not require imputation for missing data.
Notification of the possibility of selection at the time of the Round 5 interview and notification by mail of actual selection for the QRS will serve to establish the legitimacy of the effort and its association with MEPS. To accommodate respondents’ schedules, the advance letter offers respondents a toll-free number to call to complete the interview at time of their own choosing. For those who do not initiate a call, telephone interviewers will place calls at different times of day and on different days of the week. After 6 well-timed and well-spaced contact attempts, the case will be reviewed by a supervisor to determine whether additional contacts should be made and what days and/or times the contacts should be made. This is a cooperative sample, well informed about MEPS, and only limited followup for refusal conversion is anticipated.
This data collection will test procedures that would be used in subsequent iterations of the Quick Response Survey to obtain timely estimates on emerging policy-relevant issues. These specific procedures have not been pretested. Data items selected for this test data collection were selected from items determined to be valid and reliable in other data collection efforts.
Data collection and processing for the MEPS Quick Response Survey will be conductd by Westat.
| File Type | application/msword |
| File Title | SUPPORTING STATEMENT |
| Author | wcarroll |
| Last Modified By | DHHS |
| File Modified | 2012-01-12 |
| File Created | 2012-01-12 |