Using propensity score matching to deal with imbalances from hard-to-reach study samples / Thomas Talhelm.

Format:

Book

Author/Creator:

Talhelm, Thomas, author.

Language:

English

Subjects (All):

Sampling (Statistics).

Social sciences--Research--Statistical methods.

Social sciences.

Physical Description:

1 online resource

Place of Publication:

London : SAGE Publications Ltd, 2026.

Summary:

This dataset explains how to use propensity score matching to deal with demographic differences between two samples. Propensity score matching is used to reduce demographic differences between samples of participants. For example, sometimes medical researchers want to compare the effect of a medicine, like a flu antiviral medication. Ideally, we‘d randomly assign people to take and not take the flu medicine, and that randomization would create two groups that are balanced on potential confounding factors like how sick they were, age, and gender. But randomization is not always possible. One way researchers deal with that is to use medical records to compare people who took the medicine versus those who did not. People who took the medicine might be older or sicker, which is a problem. Researchers use propensity score matching to create two samples that are matched on those characteristics—one group that took the medicine and one group that did not. In our example, we’ll use propensity score matching to deal with demographic differences between two samples of participants in a hard-to-reach area of rural China. Our goal is to compare the two areas with samples that have minimal differences in potential confounding factors like age and gender.Propensity score matching creates a new dataset sampled from your dataset to be matched on key potential confounds between sample, such as age, gender, and education. You will learn how to use code in the statistical program R to create matched samples. As an example, the outline uses data from a study of two state-run farms in northwest China—one wheat farm and one rice farm. People were quasi-randomly assigned to these farms starting in the 1950s, creating a unique natural experiment to test whether farming rice versus wheat shapes cultural differences in people’s relationships and how they think. This method is particularly useful for hard-to-reach populations like these two farms, where finding equivalent samples is difficult. The samples from the two farms had differences in demographics like age and gender. The propensity score matching reduced these differences, which decreases the potential for demographic confounds between the two samples. The dataset file is accompanied by a teaching guide, student guide, and how-to guide.

Notes:

Description based on XML content.

ISBN:

9781036253073

OCLC:

1594889591

Publisher Number:

T303042