Related Experiment Video
Updated: Jun 13, 2026

Inverse Probability of Treatment Weighting (Propensity Score) using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
Development and evaluation of weighting methods for the 2024 national pharmacist workforce study
David A Mott1, Vibhuti Arya2, Brianne K Bakken3
1University of Wisconsin-Madison, School of Pharmacy, Madison, WI, USA.
Background:
Post-stratification weighting adjusts sample respondent characteristics to align with population-level characteristics. Developing and evaluating weighting methods to reduce bias between sample and population characteristics can improve estimates of work and work-life characteristics.
Objectives:
(1) To examine differences in distributions of demographic variables between respondents to the 2024 National Pharmacist Workforce Study (NPWS) and a population-level data source and calculate proportional weights for each variable, (2) To develop and apply five weighting methods using proportional weights, and test the fit between weighted sample and population proportions, and (3) To quantify differences between weighted and unweighted sample estimates of work and work-life characteristics.
Methods:
Using American Community Survey population characteristics (region, gender, five-year age categories, race, two-way age category x gender), we evaluated two weighting approaches using proportional weights alone and three using proportional weights and raking. Using a fit index, we evaluated the fit between weighted sample and population proportions. We compared unweighted versus weighted estimates for six work and two work-life characteristics.
Results:
The unweighted NPWS respondent sample underrepresented younger, male and non-White pharmacists. Raking that used two-way gender x age categories, region, gender, and race produced weights that provided the best fit. Weighting shifted estimates of work and work-life characteristics consistent with hypotheses about the effects of age and gender.
Conclusions:
Post-stratification weights calculated by raking improved the representativeness of the 2024 NPWS respondent sample and refined estimates of pharmacist work characteristics. Data analysis of NPWS data could use weighting to improve estimate accuracy.
Related Concept Videos
Analysis of Population Pharmacokinetic Data
Drug Dosing: Obese Patients
Dosage Regimens: Partial Pharmacokinetic Parameters
Dosage Regimens: Designs and Approaches
Drug Dosing: Geriatric Patients
Dosage Regimen Designs: Nomograms and Tabulations
