Guhn – population-based data collection

In the School of Population and Public Health and specifically in my research unit, The Human Early Learning Partnership here at UBC, we often conduct data collection in a manner that we refer to as population level based data collection. And commonly, a lot of research is done with large samples, sometimes even with small samples. And in order to come up with correlations, associations, even causal mechanisms, samples can be very effective. When you want to be effective in terms of reaching a population for conversation, if you want to make impact in relation to policies, decision making, having data on the entire population can be extremely powerful because it allows you to put things on the map, literally speaking. You don’t have to say, “Oh this is representative for all of Canada,” but you can for each local unit, neighbourhood by neighbourhood, city by city, town by town, however you want to cut it up you can show the variability. And when you speak to parents, and I have this experience many many times, you go to a local community and you talk about research. The moment you talk about the children in their neighbourhood the discussion changes. If I say, “Oh yeah a study in New York city found… a study in Germany found… a study in Australia found… that here’s a connection between physical activity and obesity” maybe some of the generic message stands. The moment I talk about our children in our respective neighbourhoods it creates a whole different set of meaning and it connects emotionally, socially to the audience you’re trying to reach. So one of the main advantages is for communication purposes, engagement purposes, by mapping our data we can engage with the local communities in a much more differentiated manner.  

Another advantage of population level data is that you can look at subpopulations in a much more differentiated manner. If you have a representative sample within Canada of two thousand children often the conclusion is, “Well that’s maybe true for Canadians in general but we don’t really know how some of the subgroups are affected by this because our sample is too small for that to make a meaningful inference.” While with the population level research, you have data on everyone.  

The third example is; a third benefit of population level data, especially if you integrate into a data base that’s linkable, is every time you have administrative school records, you have birth records, you have all kinds of administrative data, you have data on everyone. And the moment you want to look at associations with things like well-being, social relationships, after school time, the way we do on our MDI survey you can link it to other existing resources and still have a large enough sample to make meaning inferences from bringing all these different data sources from different disciplines together. So there are multiple and methodological, communication related and policy related advantages of collecting data at the population level.  

We also have to acknowledge that population level data collection is a trade-off. You get all the benefits we talked about but of course you can’t go into the same depth as you can with certain other types of research designs. If you really want to get into very, very deep meaning, yes qualitative interviews, talking to people for several hours, doing document analyses is of course critical. So in that sense it’s one methodology that compliments a lot of other methodologies. You cannot necessarily get the same depth as elsewhere but it gives you coverage for the entire population so it has that benefit but it of course is a trade-off for other things.