You are here

How not to analyse history

Tuesday, January 27, 2015 - 22:57

The Institute of Policy Studies has a remarkably poor study on the perceptions of Singapore history.

How bad is it? This is one of the top findings: Recent events are more likely to be remembered.

No kidding? We had to survey 1,516 Singaporeans to find that out? Tell us more!

The study asked Singaporeans aged 21 and up about a list of "historical" events, including "WWII Japanese Occupation", "Caning of Michael Fay" and "Opening of two Casinos", and sought repondents' ability to recall those events as well as their views on how important those events were.

The entire report is a case study of how not to do analysis.

To begin with, many of the findings are self-evident. Of course more recent events are more easily remembered, in general. Because memory, you know, fades. And of course events that one finds important are more likely to be regarded as important for future generations. We're talking about historical events, which by definition are important across time. If I think it is important now, I should think that it will be important next year.

There is also a problem with fundamental assumptions. Such as the finding that the ability to recall an event is not correlated with its perceived importance. But that conclusion is only meaningful if the list given to respondents contained objectively equally historically significant events. That clearly was not the case. Nobody is going to think the opening of casinos was as significant as the Japanese Occupation, whether they can recall it or not. It's like, we gave 10 different types of candy to people at different times of the day, they did not like the different candy equally. Duh.

The study also tries to extrapolate from responses rates to derive factors that it believes determines perception of historical importance. Woah, somebody is getting cause and effect and correlations all jumbled up. People's ability to recall historical events and their perception of those events are shaped by policy. What history we teach in schools and how we teach it determines the responses given to that survey, not the other way around.

This study was probably done with Singapore's 50th year of indepedence in mind. It's SG50, let's do something that looks at 50 historical events. Unfortunately, somewhere in the enthusiasm to do something special, a critical assessment of whether said activity is actually meaningful or correct was lost. That should be a lesson to all of us. A lot of organisations want to commemorate this special year, and with good reason, because the nation's openness and willingness to support major projects is probably at a high right now. But not every idea is worth the same. Precious resources (polling 1,516 people!) and publicity that was given to a flawed project could have been better used on something more meaningful.