Posted by: TomA | 23 October 2008


So thank you to everyone who filled out the online survey posted last month. I had a huge number of responses (250 people completed at least some of it and half of those stuck it out to reach the finish) and it has provided some really useful and interesting results.

The BBC have since provided me with some data (considerably later than hoped admittedly), so I’ve spent my days crunching through 3.5Gb of raw data. It has been problematic and tiresome, to say the least, but the end is firmly in sight now, I’m getting some results out and this time next week should (in fact, need to be) done. The econometrics aspect has done my head in to say the least. Its not an aspect of economics I particularly enjoy; I often feel its abused to some extent to validate results without much thought or consideration. Sadly, the standard of my econometrics isn’t great, so I’ve spent a LOT of time recently slogging through journals and textbooks trying to find appropriate models and estimators to apply to my analysis. For example, running a regression of how the score arising from the survey data influences how much of a programme people watch. With a continuous dependent variable, bounded on the interval [0,1] the choices aren’t obvious. OLS is ignores the boundary; logits, probits and tobits would be fine were the data discrete; and beta-distributions looked initially promising but can’t deal with observations at 0 or unity (or which a considerable proportion of my results are). Hopefully, a generalised linear model approach using a logit link and the binomial expoential covers these key areas, but leaves me open to endless further statistical problems of which my grasp is weak.

Disappointment also arose in finding that one key aspect of the data I was hoping for is missing, which prevents me from investigating three or four of the most important hypotheses in my paper. It feels like such a cop out to hypothetically talk about these – if I had the data I’d do this, the results might suggest that, which in turn challenges this theory or provides support evidence for another. It is vague, bland and completely unsatisfying.

Still, I’ve 8 days to turn it around…. Hm.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: