I. Charges to the Committee
Review the 1997 Biennial Survey and advise GPO on the wording of the questions.

Make a final report by April 1998 with recommendations on the content and methodology of future biennial surveys and other data collection instruments in the Depository Library Program

II. Activities
The committee reviewed the 1997 Biennial Survey and made recommendations for revisions and clarification that were incorporated into the 1997 Biennial Survey.

For the longer term we looked at the data needs of the Program as demonstrated in the annual appropriation hearing testimony, information needed in the "Study to Identify Measures Necessary for a Successful Transition to a More Electronic Federal Depository Library Program," data required during the process of revising Title 44, and data needed by GPO for managing the Program. We also looked at data gathered by libraries in the Program for their own management purposes and for reporting to other organizations.

III. History and Process:
Depository libraries are required to "report to the Superintendent of Documents at least every two years concerning their condition." (U.S. Code, T. 44, Sec. 1909) The Biennial Survey has been conducted every two years since 1947, except 1949. The content of each biennial survey has been guided by the information needed at the time by GPO. Only one question is legally required: "Do you wish to remain a depository library?"

GPO uses the data:

Each Biennial Survey has reflected the particular needs of its time. As a result there is a lack of coherence and the data are not very useful to study the development of depository libraries over time. At the same time there have been categories of information needed by the GPO and by program libraries in promoting and administering the program that were not available through the Survey.

In order to give the Biennial Survey more weight as a research and management tool, we subscribe to the following: General Principles for Data Collection in the FDLP

[1] The data that are gathered should meet the needs of GPO and the program libraries:

[2] The data should complement and be standardized with other major data surveys, insofar as possible, e.g., IPEDS, ARL, Public Library.

[3] Each data element should be sufficiently defined so that FDLs can complete the survey instrument with accuracy and consistency.

[4] Data elements and definitions should be consistent over time. There must be compelling reasons to add or change data or definitions.

[5] Program libraries should have sufficient advance notice of new data or definitions to allow them to prepare, e.g., if annual data are required, the libraries should know at least a year in advance.

[6] The content, wording and methodology of the biennial survey should be reviewed and tested in advance by experts in statistical and survey methodology.

[7] The content, wording and methodology of the biennial survey should remain relatively constant or should change only with advance warning. If other data are needed on an occasional or emergency basis, GPO should use other means or special surveys. With the assistance of Depository Library Council GPO should review the survey instrument biennially in the off years, looking at old responses and suggesting substantive and reasoned revisions. In general a question should be of sufficient lasting import to stay on the survey for at least three surveys.

IV. Recommendations and Rationale
Recommendation:
We recommend that GPO formalize and standardize the Biennial Survey. Issues to consider in this process include the general principles for data collection outlined in the report of the Statistics Committee. We further recommend that:

Rationale
Although the Biennial Survey, as it currently exists, has yielded useful information for the GPO, the information gained has not equaled the investment in time and energy that it takes to create, answer, and the compile the data. GPO and the program libraries have frequently needed data in the past that were not available, e.g., data on users, data on the cost of being a depository. We need to begin to gather data on a consistent and methodologically defensible basis. We need to gather the data that experience has shown are necessary for our purposes. We need to be able to compare the data against other survey instruments currently used in library data collection. GPO and the FDLP would be improved by tracking data over time and determining trends in service, collections, etc. FDLPs should be able to compare and contrast data over time and use this information to improve service and planning locally. GPO and the FDLP will be able to use the data to promote and defend the program.

V. Content of the Biennial Survey
We are not prepared at this time to specify the exact content of the Biennial Survey. We recommend that a group composed of representatives of Council, Program libraries, and GPO staff develop the survey instrument with the help of an expert consultant. A methodical approach might be to:

[1] consult the interested parties about the information they need on a regular basis for good management, including:

[3] look at the professional literature on the topic;

[4] create and test the questions, definitions, and methodology before Fall 1999.

Considerable work has been done on these issues in the past. The published professional literature should be mined for useful information. Among those we have consulted are:

American Library Association. GODORT Statistics Interest Group. "Statistics Guidelines for Government Documents Librarians," Documents to the People 9(November 1981): 279-284.

Morton, Bruce. "Random Thoughts on Numbers: the Need for Minimum Uniform Statistical Reporting Standards for U.S. Depository Libraries," Government Publications Review 11 (May-June 1984): 195-202.

Seavey, Charles. "Measurement and Evaluation of U.S. Federal Depository Collections," Government Publications Review 18 (March-April 1991): 147-155.

Having said that we will not specify content, we do have some suggestions that might be considered. We do not want to suggest that these are the only questions, merely the ones that, as representatives of Program libraries, we have thought of so far.
* the basic legal question -- Do you want to continue as a depository?
* the depository library profile
* questions that might get at the cost of being a depository library (The question of the cost of being a depository is one that has come up again and again. This information has eluded us so far. We suggest that GPO explore with experts the feasibility of constructing a cost model based on average and typical costs of depository libraries, in the manner of the "market basket" used in the Consumer Price Index.)

24 March 1998
Submitted by the Statistics Committee of the Depository Library Council