0990-0223

B. Collection of Information Employing Statistical Methods

1. Respondent Universe and Sampling Methods

The sampling frame will be a list of Medicaid beneficiaries in New Jersey who are eligible to enroll in the state’s Cash and Counseling program, Personal Preference, but who did not enroll. MPR will obtain the frame upon resumption of outreach to, and enrollment of, new prospects in New Jersey. Those who were part of the Cash and Counseling demonstration evaluation (in either the treatment or control group) will be excluded. MPR will consult with New Jersey administrators to determine whether the sample frame of current eligible nonparticipants should be drawn from current Medicaid personal care claims data, from contact lists the state used during outreach activities, or from some other source. If using claims data as the frame, they will identify those with claims for cashed-out personal care benefits while it was possible to enroll. MPR will work with state program staff to ensure that they are in compliance with HIPAA regulations when forming the sampling frame, selecting the sample, and contacting sample members.

MPR will select a stratified random sample large enough to yield 600 completed telephone interviews. The assumption is that 90 percent of the sample will be eligible, and that 75 percent of the eligible sample members will respond. That would mean the initial sample release should be about 889. Stratification variables may include age group, gender, race/ethnicity, tenure in Medicaid personal assistance program, and county. MPR may also stratify by whether the beneficiary ever expressed any interest in the program. For these stratification variables, MPR will work with the ASPE project officer to decide whether to oversample certain subgroups, or just to ensure that they are all proportionately represented in the sample, using either explicit (cell-based) or implicit (sort-based) stratification. MPR will most likely use a sequential sampling technique based on a procedure developed by Chromy and available as a procedures in SAS (SurveySelect).¹

2. Procedures for the Collection of Information

MPR will be collecting primary data using an interviewer-administered telephone survey over a three-month period approximately five months after OMB approval. The survey includes original questions designed to measure factors related to nonparticipation. It also contains some questions similar or identical to those asked in the evaluation of the Cash and Counseling demonstration surveys. Those questions will allow comparisons between current participants and people who were part of the Cash and Counseling demonstration (participants).

MPR will use a hard copy survey instrument to conduct the telephone interview, and then enter the data into an electronic database. As part of this process, they will check for errors in skip patterns and out-of-range response values. As noted, with only 600 completed interviews, and a relatively straightforward instrument, it is not worth the cost of programming the survey instrument into a computer-assisted telephone interview.

3. Methods to Maximize Response Rates and Deal with Nonresponse

MPR will use several techniques to maximize response rates in the survey sample. As noted, it will distribute advance letters to explain the study and solicit the cooperation of eligible respondents. Then, when it attempts to interview sample members by telephone, MPR will employ proven procedures for locating sample members, and preventing and converting refusals. They expect to use standard locating resources for about three-quarters of the sample members. Interviewers will be given specific suggestions for addressing respondents’ reluctance to participate, thus avoiding refusals in the first place. When sample members are reluctant to participate in an interview, the interviewer will enter the reason for a respondent’s reluctance on the contact sheet. In addition, all refusals will be evaluated by a supervisor who will judge whether an attempt to convert the refusal is warranted and appropriate. If it is, refusal conversions will be conducted by specially trained, highly experienced, persuasive interviewers who enjoy the challenge of overcoming a refusal. When appropriate, MPR will also send out refusal conversion letters targeted to specific situations. (Example of refusal conversion letter included in Appendix 2.)

Inappropriate items and phrasing can lead to break-offs, thus reducing completions and response rates. Different versions of the questionnaire will be developed for different respondents, inserting appropriate pronouns (second person or third person) as appropriate to avoid interviewer errors that may confuse the respondent and/or collect information about the wrong person. In addition, barriers to completion include sample members’ being physically or cognitively unable to use a telephone, linguistic preference, and, less commonly, living in a household with a telephone. To overcome these barriers, MPR will, respectively, conduct interviews with knowledgeable proxy respondents, conduct interview in Spanish (the most common non-English preference for this sample) or make use of interpreters, and provide a toll-free number so that respondents can participate in interviews from a neighbor’s house or public telephone.

4. Tests of Procedures or Methods to be Undertaken

During the first week of October 2004, MPR conducted a pretest with nine Arkansas Medicaid beneficiaries who were eligible for a Cash and Counseling program, IndependentChoices, but who did not sign up. Some had heard of IndependentChoices, and others had never heard of it. The nine respondents were people who called in and volunteered to do the “practice interview” after being sent a letter about the pretest from Arkansas’s Department of Human Services (Division of Aging and Adult Services). As a result of the pretest, MPR found several questions that required wording changes, or additional interviewer probes, to be clearer to respondents. No questions were removed as a result of the pretest. A fuller description of the pretest was sent to the ASPE project officer. The pretest with nine English-speaking respondents revealed that the interviews ranged from 19 to 37 minutes each, averaging 26.8 minutes.

5. Individuals Consulted on Statistical Aspects and Individuals Collecting and/or Analyzing Data

The person responsible for sample acquisition and survey administration is:

• Barbara Lepidus Carlson, Senior Statistician, Mathematica Policy Research, Inc., (609) 275-2374.

The person responsible for overall research design, data analysis, and reporting is:

• Leslie Foster, Researcher, Mathematica Policy Research, Inc., (609) 936-3265.

C. Attachments/Appendix

1. Copies of applicable sections of laws or regulations [ASPE TO PROVIDE]

2. Data collection instrument

3. A copy of the 60-day Federal Register notice [ASPE TO PROVIDE]

4. Introductory and follow-up (refusal conversion) letters to respondents

1 This procedure offers all the advantages of the systematic sampling approach but eliminates the risk of systematic, list-order bias by making independent selections within each of the zones associated with systematic sampling, while controlling the selection opportunities for units crossing zone boundaries.

4