Attachment 3

Attachment 3

Response to OMB Comments on SSA Ticket to Work Beneficiary Survey

Question #3

Ticket to Work Beneficiary Survey

Response Rate Calculations:

As called for by the Office of Management and Budget’s recently released Standards and Guidelines for Statistical Surveys, this study will use Response Rate 4 (RR₄) as defined by the American Association of Public Opinion Research (AAPOR). This rate is calculated using the following terms:

• Completes – both full and partial as well as screen-outs;

• Refusals – both refusals to begin and break-offs in the middle of the survey;

• Non Contacts – never able to make contact with the identified individual;

• Other non-interviewable individuals – those who are physically or mentally unable to participate as well as those with a language problem;

• Unknown Households – always busy, no answer, call blocked, or other line problem;

• Unknown Other – unable to complete the screening portion of the questionnaire, and

• Not Eligible – fax/data line, disconnected, cell phone, non-residential number, over quota, and any other not eligible category.

The first step in calculating a response rate is to estimate the proportion of cases of unknown eligibility that are eligible. This estimate, e, is calculated using the following formula:

	e =	Completes + Refusals + Non Contacts + Other
		Completes + Refusals + Non Contacts + Other + Not Eligible

Once e has been calculated it is a simple matter to calculate RR₄. This is done using the following formula:

	Completes
	Completes + Refusals + Non-contacts + Other + e(Unknown Household + Unknown Other)

Locating Missing Phone Numbers:

It is anticipated that most of the cases drawn for this study will have a telephone number included in the administrative data on file with SSA. Attempts will be made to find telephone numbers for those cases where this information is missing.

There are two standard ways of locating missing telephone numbers: directory assistance and accessing information on file with various credit bureaus. SRBI has access rights for basic information (address and phone number) with the major credit bureaus. The fees for accessing this information range from 30 cents to $17 per case depending on the amount of information available and the manner in which the request is submitted (batch vs. individual names). Given the limited budget for this phase of the project, it is anticipated that we will use the 30 cent per case function and follow this up with directory assistance for those cases not found.

It should be noted that the budget for this phase of the project was kept purposely low given the previous experience that SSA administrative data has a very high proportion of current telephone numbers for individuals who have applied for programs such as Ticket to Work.

Analysis of Non-Response Bias:

In most cases it is difficult to determine if the characteristics of respondents differs from those of non-respondents and if these differences have an impact on the results of the study. That is, had non-respondents been included, would the results have been different? However, in this study, since the sample will be developed from administrative files, it will be possible to compare differences between respondents and non-respondents on various characteristics already on file. These characteristics could include, but would not be limited to, gender, age, race, ethnicity, level of disability, date of on-set, residence location (urban/suburban/rural), etc. To the extent that the statistically significant differences are found in these areas, and if it is felt important to the analysis, the data collected in the survey can be weighted to compensate for the two or three most important discrepancies. The determination of which discrepancies to compensate for will be made on the basis of relative importance and the number of cases in each cell.

2