HSR Reports (1995) Profile: William P. Fisher, Jr. on Rasch Measurement. Rasch Measurement Transactions, 9(2), p.427.
William P. Fisher, Jr., Ph.D., specializes in tailoring quantitative and qualitative health information systems to particular practice environments for use in clinical research and practice, utilization management, quality assessment and improvement, patient education, preventative medicine, treatment planning, health economics, and marketing.
What if you and a colleague in another institution were measuring patients' health signs using different instruments? Until recently, comparing the scores from the two patient groups would have been analogous to comparing apples and oranges, making bench-marking impossible. New mathematical modeling techniques are transforming rating scale, assessment, and test-based measurement. Probabilistic models originally developed by the Danish mathematician Georg Rasch, and extended and refined by the University of Chicago's Benjamin D. Wright, offer substantive improvements over currently popular methods of instrument validation.
Imagine easily being able to compare results of two studies that
have used different measuring scales, or testing the reliability
of a health status measurement tool much more efficiently than
ever before! In addition to expediting the above, Rasch
measurement constructs:
* equal interval units of measurement for statistical
comparison;
* expression of scale values of survey/test items and people's
health/abilities/attitudes in the same quantitative unit;
* individual scale value, error, and data quality estimates
(model fit statistics) for each item and person;
* no problem with missing data;
* increased data quality;
* decreased data volume;
* validity assessment and reliability study built into analysis;
* the resources required to design self-scoring report forms
that combine the rating worksheet with the measurement analysis
for instant, on-the-spot qualitative analysis and quantitative
results, obviating the need for expensive and time-consuming
computerization;
* multi-faceted models that adjust measures for uncontrolled
variations in judges' perceptions of, or environmental effects on
health and performance;
* computer-, clinician-, or self-adaptive administration of
instruments, which increases measurement efficiency by posing
only the questions relevant to the patient; and
* the co-calibration (equating) of alternative instruments
intended to measure the same construct, making it possible for
them to measure in the same unit, even if they have different
numbers of items, or are based on different observational
frameworks, such as clinical examination, performance assessment,
or patient self-report.
Abstracted with permission from HSR REPORTS, A Newsletter from the Department of Health Systems, Research & Public Health, Louisiana State University, School of Medicine in New Orleans, July 1995, Vol 1, No. 1.
Profile: William P. Fisher, Jr. on Rasch measurement. Fisher WP Jr. … Rasch Measurement Transactions, 1995, 9:2 p.427
The URL of this page is www.rasch.org/rmt/rmt92f.htm
Website: www.rasch.org/rmt/contents.htm