National Long-Term Care Survey

Overview of NLTCS Documentation

The 'Documentation' section of the NLTCS website contains detailed background information helpful in understanding how data have been derived and/or are used in analysis.  Documentation is grouped into four general categories: 'Sampling Memos,' 'Weights Memos,' 'Source and Accuracy Statements' and 'Other Documentation.'

Sampling Memos

To start the survey in 1982, a sample covering all age groups was drawn from the Medicare eligibility files of all persons 65 and older.  The most important of the sampling memos, 'Sampling Specifications for the Long Term Care Survey (LTC)' dated February 26, 1982, gives the sampling plan in great detail.  Sampling memos in later years reveal the same philosophy but only address the aged-in component and any supplement that was drawn from Medicare files (e.g., 95+ supplements). 'These are typically produced by the Census Bureau before the sample is drawn.

Weights Memos

Base weights are produced at the time the sample is drawn as the reciprocal of the probability of selection into the sample. 'Final weights are described in the weights memos as part of the final documentation. 'These weights account for factors that cannot be known at the time of the survey (e.g., certain kinds of loss-to-follow-up and non-response). 'The weighting memos are typically produced by the Census Bureau with input from Duke researchers.

Source and Accuracy Statements

The Source and Accuracy statement lists the known sources of non-sampling errors present in the data and provides an estimate of their likely magnitude. 'Also listed are the quantities and proportions that can be estimated and the sampling errors, or equivalently the size of the confidence intervals for each estimate. 'These are typically produced by the Census Bureau with researcher input after data collection.

Other Documentation

Other documentation provided includes several items: Overview and Use of the Public Use Data Files of the 1982 and 1984 Long Term Care Surveys; and an eight-part 'Final Report; of the 1999 Supplemental Surveys prepared by the Research Triangle Institute and documents on assisting pseudostrata and half sample codes and calculating design variances.  These are described briefly from the navigation link "Other' to the left.

For links to the individual documents provided in the above categories, see the navigation menu to the left.

We would welcome your suggestions for additional documentation products that would enhance your effective use of the NLTCS. Please contact Kenneth G. Manton, Ph.D. at 919-668-2703 or kgmanton@duke.edu