A. Consumer Sampling

A. Consumer Sampling

A. Consumer Sampling

The foundation of the sample method in consumer research relies on two factors. The first, to which we have already referred, is that the vast majority of people interviewed in a survey will give answers which are true to the best of their knowledge. The second is that it is usually possible to establish that the sample is, within fairly acceptable limits, representative of the whole.

It is important that one should understand the importance of this second factor. Upon the degree of probability which exists in any survey will depend the credibility of its findings. If there is reasonable doubt that the sample which is being examined is not representative, then it will be better to abandon the project and start again; because a wrong sample will produce wrong results which could be disastrous for those who may decide to act upon them.

How certain can one be that any sample is truly representative of the whole?

Let us consider, for a moment, a typical example of sampling, frequently undertaken by businessmen when entertaining their more important clients. When they order a bottle of wine they will, in the first instance, be shown the bottle with its cork intact. This enables them to be assured that the description on the label coincides with the particular vintage which they have ordered. Subsequently, the wine waiter will uncork the bottle and pour a small sample of its contents into the host's glass for him to taste and test. Now, this conventional method of approving wine is based on the overwhelming probability that the sample which is taken is representative of the whole of the contents of the bottle.

Few consumer surveys are quite so undeniable! Indeed, in an effort to overcome this problem, market researchers have developed a number of methods all aimed to achieve the highest degree of representativeness compatible with reasonable economy.

We have already referred to the possibility that false statements made by individual respondents could have a detrimental effect upon the value of the survey. In practice, however, it has been found that a false statement in one direction is usually cancelled out by an equally false statement in the other. Where it is discovered, in a survey, that a high proportion of false answers have been given to a specific question, it is usually because that question has been badly phrased. Note that WE say obviously false answers. When we come to consider the preparation of questionnaires it will be seen that the series of questions put to respondents can be designed in such a way that untruthful answers can be pin-pointed without much difficulty.

The obvious advantage of the sampling technique compared with a total survey of the whole market is that the cost of the sample survey is much smaller. However, since the size of the survey will have a direct bearing on its cost, it will be dependent upon the budget which has been allocated. The ideal sample is a microcosm of the entire population. It should be noted that the words population or universe refer, in market research parlance, to the sub-section of the actual population which one is investigating. A universe, therefore, may be the total number of schoolteachers, or the total number of families with two cars.

It is the composition of the sample that is more important than its size. It should be remembered that the researcher is not trying to establish a large or a small sample upon which to carry out his survey but one of optimum size. If, for example, one intends to survey a whole country, one would arrange about 2,000 interviews for the sample. This figure, however, would apply equally to countries with such divergent population sizes as, let us say, the United Kingdom, the United States of America, or Germany or Holland. Where the universe is uniform it is often possible to work with a smaller sample, but it must remain a true cross-section of the population.

Where the universe is not uniform, the researcher will probably divide it up into sections-or strata-which are uniform. It will be appreciated that when this occurs the size of the sample which is decided upon will depend upon the size of the smallest sub-section. This is because no sub-section can contain too few respondents. However, this could lead to the necessity of interviewing such a large number of respondents that the entire survey would be hopelessly uneconomic.

Let me explain this with an example. Let us say that we select a sample containing 2,000 interviews in order to study the buying habits of a universe of, say, 750,000 people. In order to represent adequately the views of the various categories of people which make up that universe, 50 persons have been included to repre-sent those with an income in excess of �10,000 per annum. Now it must be obvious that 50 is too few to reflect adequately the variety of viewpoints, personal interests and buying habits of the highly paid section of the universe.

The problem is, that if we were to double the quantity of the �10,000-a-year people the whole sample would have to be doubled. That would be very expensive. The way round this problem is to double only the number of highly paid people and increase the sample to 2;050 interviews. A correction is made to the final results when the figures which relate to the answers given by the wealthy respondents are halved. Thus the proportions are maintained and the final results derived from the survey are not distorted. This device is known as weighting the results.

There are various methods of selecting samples for consumer surveys. Those most frequently employed are the probability, the random and the quota systems. In practice, researchers often use hybrid methods to suit the type of research in which they are engaged.


Read on - Sales Force Communications 2012



Please Note

The Trade is, of course, a major source of product ideas. All manufacturers examine, with avid interest, the new products of their competitors.

  • Latitude - performance led digital marketing agency

    www.latitudegroup.com/
    Latitude is a leading digital marketing agency providing services including PPC management, SEO, web analytics, social media optimisation, online display.

    Brighton Digital Marketing Festival (BDMF) - from Pure360

    www.brightondigitalmarketing.com/
    The Brighton Digital Marketing Festival (BDMF), is a one day event of interactive sessions, bringing together the brightest talent in the digital marketing industy.

    Top SEO in the UK

    www.benedict.co.uk
    Benedict Digital Marketing Expert is available on Friday afternoons to give free advice for people doing web projects or students