Unreliable Statistics (Part 2)


Mike Rigby is right that we should treat all published statistics with caution ('Unreliable Statistics', Glazine 1st July). The comparison of the OFT report to which he refers with the annual installation rate of conservatories is a case in point.

The OFT report itself is not new. It was published over three years ago (February 2000) and is based on a survey of 2,200 adults in Great Britain and Northern Ireland. It is the latest in a series of Consumer Detriment surveys that the OFT carry out at five yearly intervals.

It is not specifically about conservatories or even the glazing industry as a whole, although interestingly it did also estimate1,007,000 consumer problems with windows, double glazing in 1999. In all it estimated 85.8 million consumer problems, in 60 categories ranging from food and drink products to pension schemes, for the year 1999. The 85.8 million problems nationally were estimated by grossing up from the sample.

Conservatories made up just 0.22% of the problems reported by the 2,200 respondents. Unfortunately research details from the survey have not been allowed into the public domain but it is evident that small changes in that percentage could make a large difference to the published number of conservatory problems (185,000). For instance if the percentage had been 0.12% the number falls to 100,000 but if 0.32% then the number shoots up to 275,000.

A further difficulty is that the report does not make clear when the conservatories were actually installed. Our estimate is that between 1980 and 1999 about 1.5 million were installed in Great Britain.

But in any case the purpose of the OFT report is to get a measure of different types of consumer problems and how they are changing over time, not as a way of deducing the size of any particular industry.

So, yes, let's be careful about using published statistics, especially when they are being used for purposes other than for which they were intended!

Yours etc.

Robert Palmer
Palmer Market Research