Skip Navigation

Academic Libraries Survey (ALS)



4. SURVEY DESIGN

TARGET POPULATION


The libraries of all institutions in the 50 states, the District of Columbia, and the outlying areas that have as their primary purpose the provision of postsecondary education. Branch campuses of U.S. institutions located in foreign countries are excluded. Through 1996, ALS distinguished between libraries in postsecondary institutions accredited by agencies recognized by the Secretary of the U.S. Department of Education and libraries in nonaccredited institutions that had programs of 4 or more years. Starting with the 1998 collection, the major distinction has been whether or not the library is part of a postsecondary institution that is eligible for Title IV funds.

SAMPLE DESIGN

ALS surveys the universe of postsecondary institutions. 

Data Collection and Processing

For the 1990, 1992, 1994, 1996, and 1998 data collections, state IPEDS Data coordinators collected, edited, and submitted ALS data to the U.S. Census Bureau, using the software package Input and Data Editing for Academic Library Statistics (IDEALS). An academic librarian in the state assisted with the collection and submission of the data.

When the ALS transitioned back into IPEDS in 2014, NCES became responsible for survey administration and data processing and dissemination. IPEDS keyholders are responsible for collecting ALS data from librarians and submitting it to NCES. Between 2000 and 2012, ALS was not a component of the IPEDS survey system. The 2000 through 2012 ALS surveys were web collections. In 2010 and 2012, the web application included a user guide, a tutorial explaining features and operation, the survey instrument, and an edit check tool. The U.S. Census Bureau was the collection agent. State-level library representatives were available to promote responses from librarians and to assist in problem resolution when anomalies are discovered in responses.

Reference Dates. Most ALS data are reported for the most recently completed fiscal year, which generally ends before October 1 of the survey year. Information on staff and services per typical week are collected for a single point in time during the fall of the survey year.

Data Collection. In the 2000 through 2012 ALS data collections, library respondents submitted data directly to the Census Bureau through the Web. For the 2008 web-based data collection, state-level library representatives were available to promote prompt responses from librarians. The web-based survey is the latest in a number of steps to improve ALS collection.

In July 1990, NCES initiated an ALS improvement project with the assistance of the National Commission on Libraries and Information Science (NCLIS) and the American Library Association’s Office of Research and Statistics (ALA-ORS). The project identified an academic librarian in each state to work with the IPEDS coordinators in submitting their library data. During the 1990s, many of these library representatives took the major responsibility for collecting data in their state. Others were available to assist in problem resolution when anomalies were discovered in completed questionnaires.

The ALS improvement project also led to the development of the microcomputer software package IDEALS, which was used by states in reporting their academic library data from 1990 through 1998. Along with the software, NCES provided state IPEDS Data coordinators with a list of instructions explaining precisely how responses were to be developed for each ALS item. Academic librarians within each state completed hard-copy forms, as they had previously, and returned them to the state’s library representative or IPEDS coordinator. States were given the option of submitting the paper forms, but were encouraged to enter the data into IDEALS and submit the data on diskette to the Census Bureau; a majority of states elected the diskette option.

ALS was mailed to postsecondary institutions during the summer of the survey year, with returns requested during the fall. Any survey returns from institutions that did not have an academic library were declared to be out of scope, as were institutions that did not have their own library but shared one with other institutions.

Editing. The web-based data collection application features internal edit checks. An edit check tool alerts the respondent to questionable data via interactive 'edit check warnings' during the data entry process and through edit check reports that can be viewed on screen or printed. The edit check program enables the respondent to submit edited data to NCES which usually required little or no follow-up for data problems. Over the years, ALS has made use of seven types of edits: summations, relational edit checks, range checks, current year/prior year comparisons, ratios, item comparison, and missing or blank items.

After responses are received, the U.S. Census Bureau reviews the data and contacts respondents with questionable data to request verification or correction of that data. Data records are then aggregated into preliminary draft tables, which are reviewed by NCES and the U.S. Census Bureau for data quality issues. Once all edits have been performed and all corrections have been made, the data undergo imputation to compensate for nonresponse (see below).

Top

Estimation Methods

Imputation is used in ALS to compensate for nonresponse. In 2014, when ALS was reintegrated into IPEDS, the imputation procedures changed. In 1994, the procedures were changed to use data from the previous survey, if available, and to only use imputation group means (see below) if prior-year data were not available. Before 1994, only imputation group medians were used.

Imputation. ALS imputation is based on the response in each part of the survey. Most parts go through either total or partial imputation procedures, except for the following items: (1) Number of branch and independent libraries; (2) Library staff information - contributed services staff; and (3) Library operating expenditures - employee fringe benefits. These items are imputed only if reported prior-year data are available (contributed services staff and employee fringe benefits apply to only a few institutions). Items (1) Electronic Services, and (2) Information Literacy do not go through imputation.

The imputation methods use either prior-year data or current-year imputation group means. The procedures are slightly different depending on whether an institution is totally nonresponding or partially nonresponding in the current year. If prior-year data are available, the imputation procedure either carries forward the prior-year data or carries forward the prior year data multiplied by a growth factor. If prior-year data are not available, the imputation procedure uses the current-year imputation group medians or means as the imputed value.

Medians/means and ratios are calculated for each of 11 non-mutually exclusive imputation groups based on 27 imputation cells that were formed by the various combinations of FTE and institution sector. The use of FTE to determine imputation cells was not employed until 2002. In 1998 and 2000, the strata were based upon the highest level of degree (doctor's, master's, bachelor's, and associate's) and control and size of institution. The four control/size imputation categories were (1) public, less than median number of degrees for institutions in that category; (2) public, equal to or greater than the median; (3) private, less than the median; and (4) private, equal to or greater than the median.

After imputation, if a total was missing or known to need adjustment, then the total was readjusted to equal the sum of its detail items.

Using a ratio adjustment to prior-year data represented a change from the imputation procedures followed in cycles prior to 1996, and may have resulted in some small differences in estimates. While checks indicate that the effect of the change was not large, caution should be exercised in making comparisons with pre-1996 or earlier reports. Using FTE to determine imputation cells and using medians instead of means for imputation also represents a change from the procedures followed in cycles prior to 2002. While research indicates that the effect of the change in imputation procedure was not large, caution should be exercised in making comparisons with reports from 2000 or earlier.  

Top

Recent Changes

Before 2000, ALS was a component of IPEDS; the state IPEDS Data coordinators collected, edited, and submitted ALS data to the Census Bureau, using the software package IDEALS. From 2000 to 2012, ALS data was collected over the Internet via a web-based reporting system, with the Census Bureau as the collection agent. Beginning in 2014, ALS was reintegrated back into IPEDS and was collected via the IPEDS web-based reporting system, managed by the U.S. Department of Education.

Several changes were made to the survey instrument in 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, and 2012. These are summarized below.

In the 1996 instrument, the data items in part E (Library Services) were expanded to request separate reporting for returnable and nonreturnable, as well as totals. In addition, a new section, part G, was added to collect information about access to electronic services, both on and off campus.

In 1998, a substantial number of changes were made to the ALS survey instrument, especially in the collections and expenditures sections. These changes reflect the extensive changes in library services that occurred with regard to the development of electronic media. The definition of a library was not changed, but was moved to the cover page and reformatted as a checklist. Other substantial changes were made to the Library Services section (parts E and F), the Library Collections section (part D), the Library Expenditures section (part C), and the Electronic Services section (part G).

In 2000, a new set of questions on “consortial services” was added to the questionnaire. In 2002, several questions were dropped and one was added. In 2004, a number of items pertaining to “information literacy” were added to the questionnaire. The 2006 through 2012 data collections did not include any major changes to the questionnaire with the exception of “reference sources per typical week” being changed to “total information services to individuals” in 2010. To assist data users in managing the variable changes over time, crosswalks between current-year and previous-year surveys are available in the file documentation materials of many of the surveys.

In 2012, several questions were removed from the survey. Current serial titles and electronic reference sources and aggregation services were removed from the library collection section. The information services to individuals questions were removed from the library services section. The five information literacy section questions were replaced with two questions involving student learning/student success outcomes. Also for the 2012 web-based data collection, state/jurisdiction-level library representatives were available to provide prompt responses from librarians and to assist in problem resolution when anomalies were discovered in responses from the academic librarians. For more information about the 2012 ALS, refer to http://nces.ed.gov/surveys/libraries/.

In 2014, ALS was integrated into the IPEDS data collection and the survey became a mandatory, annual survey for all degree-granting Title IV institutions. Only institutions with library expenses equal to or greater than $100,000 were required to report additional details about their expenses to IPEDS. Questions related to library staff were moved to the IPEDS Human Resources component. Many questions from the 2012 ALS collections and services sections were removed or revised. For a list of revisions, see the 2014–15 IPEDS archived changes at http://nces.ed.gov/ipeds/InsidePages/ArchivedChanges.

Future Plans

The National Center for Education Statistics will continue to evaluate the potential for changes to the ALS survey and revise as-needed to reflect changes that have occurred in the services, collections, expenditure accounting, and staffing of academic libraries.

Top