Attachment 2

Attachment 2

Response to OMB Comments on SSA Ticket to Work Beneficiary Survey

Question #3

Survey of Disability Report Form Users (Nov. 2003 - Jan. 2004) (Detailed information for the older surveys is not readily available.)

We don’t specifically follow AAPOR guidelines – actually, the standard definitions document doesn’t address telephone surveys of specifically named individuals (the only telephone surveys they address are RDD). Basically, the way we compute response rate is:

Completed Surveys + Nonresponders (see definitions used below)

Sample – Ineligibles (see definitions below)

Our approach is based in large part on guidance we received in 1997, when SSA engaged Pacific Consulting Group (PCG) and SRBI to give us advice on how to improve response rates. At the time, PCG and SRBI actually consulted with OMB regarding the issue of response rates and obtained their concurrence. PCG/SRBI felt that we were overly stringent in our approach and emphasized that we should remove “ineligibles” such as disconnected telephones, deceased individuals, etc. identified at the time of the survey from the denominator when we compute the response rate. For mail surveys, PCG/SRBI advised us to remove undeliverables (obviously, if no new address is available).

In a telephone survey of specifically named individuals, we define ineligibles as:

• Disconnected or otherwise non-working telephone number (e.g., fax/modem);

• Sampled individual is permanently unavailable (e.g., deceased);

• Sampled individual is incapacitated and cannot participate (e.g., mentally incapable); and

• Telephone number on SSA records is incorrect and no new number is reported during the survey period. This is the equivalent of an undeliverable mail survey.

In some surveys where the target population is English-speaking only (e.g., for a survey related to Internet services that are available only in English), we consider non-English speakers to be “ineligibles.” Other surveys might have different categories of ineligibles depending on the topic – one typical example is in an initial disability claim survey where at the time of the survey, the individual has a representative payee who was not involved in filing the claim. When we discovered this through screening, we would categorize the case as ineligible. The individual himself would be considered incapacitated and unable to participate.

We define non-response as:

• Refusals and

• Non-contact despite repeated attempts (15 attempts over the course of different days at different times) and there is no indication of ineligibility. This includes encountering busy signals, no answer, no response to voice-mail message, failure to keep appointment, sampled individual not available at time of call.

Using the most recent survey on your list, the Survey of Disability Report Form Users, as an example, this is how we arrived at the response rates shown.

Traditional Sample Internet Sample Partial Sample

Original Sample 998 999 952

Ineligibles 93 56 59

Nonresponders 206 242 238

Completions 699 701 655

Response rate after removing ineligibles 77% 74% 73%

Response rate before removing ineligibles 70% 70% 69%

When our contractors encounter an incorrect phone number, we require that they contact us so we can check SSA records for an updated phone number. For surveys involving people with a current connection to SSA, such as beneficiaries, we are able to locate some new telephone numbers (not always). The new telephone numbers are usually productive.

2