TQB: Teacher Quality Bulletin

In ed research, ‘tis never the season for making a hypothesis and checking it twice

See all posts

How do academic research and world exploration differ? In research, it's good to discover the same thing twice! 

Replication – research that repeats a previous study – is a cornerstone of methodological rigor. But when was the last time you read about a replicated education study? We won't judge you if you can't remember. Turns out, few exist. Of those that do, their integrity may be questionable.

In a recent study, Matthew Makel and Jonathan Plucker assess replication in education science by looking into the publication history of the top 100 education journals. When looking for replication reports specifically, their analysis found that an infinitesimal number of the education articles published in these journals were actual replications (221 out of 164,589). And while this may not seem consequential at first, it matters. Even well-crafted experiments can find anomalous results – replicating a study can confirm that the original findings hold true or can turn up conflicting information that helps us better understand the subject of analysis.

In nearly half of all replications, the same research team that published the original article was also responsible for the replication study. This is slightly worrisome given that the success rate (when the replication has similar findings to the original study) is much higher (71 percent) when conducted by the original team than when there's no author overlap (54 percent). Although same author replications are beneficial to the field, it's clear that third-party authors are needed to increase reliability.

So why is replication not expected in education research? The authors argue that social science publishers and funders want new and exciting findings, so they're more likely to fund, hire, publish and promote groundbreaking studies rather than those that seem like "old news." Naturally, this kills any incentive to attempt replications.

Despite this unfavorable environment, some researchers are bucking this trend. Check out Bacher-HicksKane and Staiger (2014) and Rothstein (2014 working paper) – both of which replicate an earlier study by ChettyFriedman and Rockoff (2014)on the impact of teachers. Be on the lookout for a write-up on those replication papers in next month's TQB.