Let’s say that marketing research studies the behavior of wolfs (bear with me, please!). You play with one to see how it responds. The wolf seems quite firendly and composed. You do the survey with 100 different wolfs so that you get a more representable sense of how wolfs are.
You do focus groups, bringing in groups of random wolfs. This time the wolfs seem a bit more agitated and hostile. But hey, all is fine – they mostly fight among themselves. Wolfs are friendly and composed, they’re just bad neighbours to each other, right?
You lock up your office and head into the woods. You run into a pack of wolfs and they tear you to pieces. Go figure!?
Possible conclusions:
1) Don’t f___ with wolfs!
2) Studying individuals is not enough! We (often) run in packs.
3) The important stuff happens in the woods. Study behavior in context!
Indeed. There is much irony in having to forget the micro if you want to study microfinance properly
By: Domen Bajde on June 2, 2011
at 6:32 pm
Very funny. In a way this makes me think of microfinance experiments (of the RCT type) where the evidence is in fact collected in situ – an improvement on those studies where “games” were played in order to “prove” microfinance works – but still the collective aspect is ignored. If you study 10 microfinance clients individually who sell mangoes and they all (miraculously) see an increase in their income, you might completely oversee the 10 other mango sellers they replaced on the same street, and whose income went down! You might also mistake the rising incomes of some mango sellers as evidence of “development” or “poverty reduction”. Alas, the trouble is, if we leave the clean world of experiments and study the whole picture, we have to deal with messy and contradictory realities… what a drag!
By: Anonymous on June 2, 2011
at 1:39 pm