Practicing Statistics: Notes
Work based on secondary data – mostly time series but occasionally cross sectional
data as well – requires standard statistical methods. You get all the relevant
methods in books. But there are pitfalls in mechanically applying textbook
procedures to available data. There are far too many gaps between what the data
should be and how they are actually generated and given to you. Let's
postpone the discussion on secondary data and turn to primary data collection.
Indeed, a fairly high proportion of Ph.D dissertations in social sciences,
especially economics, these days are based on field work involving small
scale sample surveys. What follows is of course not a systematic exposition of
the principles that should guide such work; it's a collection of ' tips and
tricks' some of which you might find interesting. All this written to
help students learning to do research; those with a little experience in
field work may also find some of the hints useful. The expert will no doubt
find it all too elementary.
Most students associate field work with a so-called questionairre (Q) to be filled up,
and promptly make up one relevant to their study. This is what they show to their research
supervisor or a statistician (who has offered to help in setting up the design of the study) as a first
effort before they take the plunge (which they hope to do after the expert says: this looks OK). Alas,
the big Q is a wrong beginning, although it's absolutely essential.