Random And Quota Sampling

Mather Digital Marketing - Learn about Digital Marketing - Random And Quota Sampling

Random and quota sampling

Thus there are still very many situations where the industrial researcher is faced with vast numbers of potential informants. It is then that our managing director suggests - on the face of it, not unreasonably - some form of random sampling. After all, he has possibly been well served in the past by consumer research SEO Agencies who have provided, at very reasonable cost, market data with statistical validation. Why can't these industrial fellows do the same?

The first difficulty is the statistical requirement that for a true random sample every item in the relevant population must have exactly the same chance of election as any other item. This means, in industrial research terms, that a complete list must be prepared of all establishments relevant to the survey before sample selection can take place.

There may well be many thousands of such establishments, and certain leading directories or credit control services, such as Kompass and Dun and Bradstreet, offer at very reasonable cost a computer service that no industrial market researcher can afford to ignore. The cost of preparation of a complete list of all establish-mess relevant to the survey, however, can still be quite un-economic in many cases. The consumer researcher, in contrast, has access to the electoral register, at negligible cost. Admittedly, this is not itself perfect recently it even included a few horses, according to press reports - but it is more than adequate for all practical purposes. The industrial researcher would give his ears for an industrial counterpart, horses and all.

Just as fundamental a difficulty in the use of random sampling (though one more easily overcome) is the lack of homogeneity in most industrial markets: certain companies offer vastly greater potential than others. Imagine the results obtainable from a random sample of the chemical industry which happened not to include ICI.

If the consumer researcher finds a random sample unnecessary or impracticable, he can turn to quota sampling. The process is simple, and, with adequate field control, reliable: the interviewer simply selects likely informants until he has built up to a pre-determined quota. Building up the quota may not be as easy as it might seem, but there is a wealth of demographic information available as a basis, much of it, such as the Census of Population, provided at vast government expense. There is nothing like the equivalent in industrial fields.

Thus, although industrial researchers do use, when appropriate, sampling methods akin to random and quota sampling, only rarely can any degree of statistical validation be attached to the results. It should not be imagined that the profession is a whole is satisfied with this state of affairs: much research is being undertaken. For the time being, however, our managing director must often be disappointed.

Interested in Application Engineering