We need more—exactly. In particular, for example, when it comes to those individual retrofits, if we want to identify the most effective retrofits.... We had difficulty in what we would term the “confidence interval”. The smaller the sample you have, once you start breaking down the data of the individual retrofits, the more variation you'll have statistically, and it will be more difficult to say what is the most effective retrofit to recommend to a given household, given that there's so much household heterogeneity that the models, almost by definition, cannot capture. The benefit of getting large samples of data is they reflect the heterogeneity of Canadian households.
