Attachment 1

Attachment 1

Response to OMB Comments on SSA Ticket to Work Beneficiary Survey

Question #3

New Beneficiary Survey (2004/2005/2006)

A.      What AAPOR formulas were used to determine the response rates for the NBS phone call surveys? 

The unweighted response rate (completed interviews+partial completes+ineligible cases divided by all released cases) is an indicator of response among the sampled cases, but does not account for unequal sample weights or for the potential for nonresponse bias. Ineligible cases are included in the numerator for two reasons: 1) the cases classified as ineligible are part of the original sampling frame (and hence the study population). We obtained complete information to fully classify these cases (i.e., their responses to the eligibility questions in the questionnaire are complete) and therefore classify them as respondents; 2) incorporating the ineligibles in the numerator and denominator of the response rate is essentially equivalent to the definition of a response rate with these cases excluded if the persons with an additional estimation of the number of eligible cases among those with eligibility unknown. By including the ineligible cases in the numerator and denominator, we avoid using this estimation stage and the response rate computation is more clearly explicated. (This is from the Ticket to Work evaluation, 2^nd report. The appendix C section on this subject is provided below).

Appendix C: National Beneficiary Survey Methodology

People in both the Representative Beneficiary Sample and the Ticket Participant Sample

receive the same survey questionnaire. Round 1 CATI data collection for both samples

began in February 2004. Beginning in May 2004, in-person CAPI interviews were conducted concurrent with CATI interviews. In-person interviews were conducted with sample persons who requested an in-person interview, those who needed an in-person interview to accommodate a disability, and telephone nonrespondents. CATI and CAPI Round 1 data collection was completed in October 2004. A total of 7,603 interviews were completed with individuals in both samples: 6,302 cases were completed by telephone and 1,301 were completed by CAPI. An additional 531 sample persons were determined to be ineligible to participate in the survey.¹ The overall unweighted response rate for the combined sample was 77.2 percent.² The overall weighted response rate was 77.6 percent.³ Proxy interviews were completed for 1,997 sample persons.

Interviews were completed with 6,520 individuals in the Representative Beneficiary

Sample and 1,083 persons in the Ticket Participant Sample. An additional 458 beneficiaries and 73 Ticket participants were determined to be ineligible to participate in the survey. The unweighted and weighted response rates for the Representative Beneficiary Sample were 77.0 percent and 77.5 percent, respectively. The unweighted and weighted response rates for the Ticket Participant Sample were 78.9 percent and 80.9 percent, respectively.

B.      Describe the protocol that MPR uses when they encounter an incorrect phone number.   

The interviewer statuses the case with a code specifying the outcome--that is whether the phone is not in service, is a fax line, or is the wrong person.  The case then goes to the locating department.  Locators use a variety of sources including directory assistance, checking contact information from previous rounds if a longitudinal case, dialing other numbers provided by SSA (payees, authorized reps, etc.), Internet searches, client updates, and Accuriant (a service which maintains various databases for this purpose).  If the address is thought to be valid, a letter is sent informing the sample person of the survey and asking him/her to call in.  Updated address information may be obtained from the letters which can then be used to locate a phone number.  If a new phone number is found, the case goes back for CATI calling. If not, and it is eligible for field treatment, it goes to CAPI.

2