Hennen's American Public Library Ratings
The ratings are based on federal data but only the author is responsible for ratings on this site.
haplr-index.com  6014 Spring Street, Racine, WI  53406   USA

Share knowledge, seek wisdom. 

Home    Ratings      Top Libraries    Orders     Samples    Press    
Consulting      FAQ     Book     Presentations     Miscellaneous  


Imputed Data

In this edition, we have stopped using imputed data. 

What is imputed data and why does not using it matter? 

Imputation provides educated guesses about what the data would have been if a library had reported it.  The Institute of Museums and Libraries (IMLS) imputes data for individual libraries when the libraries themselves fail to report data elements.  They do the imputation based, among other things, on past reports of an individual library or based on the average for libraries within a population category.  Eliminating libraries that have imputed data for any measure in the HAPLR rating system means eliminating 1,287 libraries in the IMLS dataset (13.9% of all libraries in the U.S.)

Imputing data serves a useful purpose for library data as a whole.  When libraries fail to report a data element, annual visits, for instance, imputation is used to "guesstimate" what the value would have been if reported.  This allows us to put together reports about annual visits at a macro level.  We can then talk about annual visits per capita for all libraries in the U.S. with a fair degree of assurance that the averages are close to the mark. However, for individual libraries the imputation, based on prior year activity or population group averages, may be far from the mark.  In the HAPLR system that will not drastically effect the final score, but it will do so. 

Eliminating results for libraries with imputed data adds to the precision of the HAPLR scores. 



Home Ratings Order Samples Press FAQ Misc

April 2010

Webmaster: thennen@haplr-index.com