Should I favor better matching or a larger number of included patients?

Modified on Sat, 21 Jan 2023 at 09:01 PM

When using propensity score matching (PSM), there is a trade-off between the number of included patients and the quality of the matching.


Having a larger number of included patients will increase the precision of the estimates of treatment effects. However, if the matching is poor, the estimates of treatment effects will still be biased. Therefore, increasing the number of included patients should not be at the expense of the quality of the matching.


On the other hand, better matching can be achieved by including fewer patients, but this reduces the sample size, which can decrease the precision of the estimates of treatment effects. Therefore, the goal should be to find a balance between having a sufficient number of included patients to increase precision and having a good match to ensure the validity of the estimates of treatment effects.


A practical strategy to achieve this balance would be to first include a larger number of patients and then perform the matching, and then, remove any outliers or participants with poor matches so that the quality of the matching is not compromised.


Another strategy to consider is using a more lenient match tolerance; this increases the chance of finding a match, although it may not be as close as the ideal match, it may allow increasing the sample size without sacrificing too much the quality of the matching.


By default, EasyMedStat picks this latest option: a balanced choice that favors matching but with a large tolerance. You could however change these settings manually if you have knowledge regarding PSM.


It is also important to keep in mind that propensity score matching is not the only method to handle confounding variables, and sometimes other methods such as propensity score weighting or adjustment might be more appropriate depending on the characteristics of the study. Researchers should consider the strengths and limitations of PSM and weigh it against other methods before making a decision.

Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select atleast one of the reasons

Feedback sent

We appreciate your effort and will try to fix the article