Quota sampling

To ensure that respondents to a random sample are genuinely representative of the population from which they are drawn, it is common practice to calculate the proportions of respondents by age, gender, class, region and so on, and check that these are in line with the known figures of census data. Since this has become routine procedure for checking the accuracy of random samples it is a very short mental step from here to suggest that if applying the proportions of known characteristics of the population to the results of random samples is valid as a check for their accuracy, then why not simply select the sample so as to represent those characteristics in the known proportions in the first place? This reasoning forms the basis for quota sampling. The reasoning is that if the known characteristics of the population are represented in the correct proportions then all other data collected will be represented in the correct proportions. Stating its underlying assumption in this way also illustrates the main reason for criticism of quota samples. It could happen that even though known and measurable characteristics are distributed in the correct proportions, unknown characteristics relevant to the subject of the survey may not be.

A quota sample based on interviewing people at home is likely to underrepresent those who go out a lot. A quota sample based on interviewing people in the street may underrepresent those who do not go out a lot. If behaviour relevant to going out or not going out is important to the context of a quota sample survey - perhaps it is about leisure activities in general, or frequency of visits to pubs, or the amount of TV viewing - then important aspects of the data may be misrepresented, resulting in the collection of biased data which will mislead the decision maker. When the significance of the limitations of the sampling procedure is as obvious as in the example suggested, then common sense can go a considerable way to overcoming this disadvantage. The theoretical danger of quota sampling is that some hidden bias may exist which will not be discovered.

Quota sampling is considerably influenced by the researcher, in ways that random sampling is not. In the first place, the researcher determines which characteristics should be used as a basis for setting quotas. Next, the actual selection of respondents is left to the interviewer rather than determined by the sample selection procedure as in random sampling. In quota sampling the interviewer is presented with a set of target interviews to complete, and the interviews are described in terms of the characteristics of the respondents required. For example, the interviewer may be told that he or she has to complete 30 interviews in three to four days' work, the 30 interviews to be completed as shown in Figure 7.1.

Your as&s/nmeftl is lb interwow SO rasporjdenls to- the sen. age and cVsas quotas detailed Derails

Your as&s/nmeftl is lb interwow SO rasporjdenls to- the sen. age and cVsas quotas detailed Derails

RfKniirTK)

jrttjpt

Mat*

is

FeiTiaTa

T6

50

ftouuinstf

Adtnu/ad

fotai

1B—34

13

3S-54

ti

ss*

A

90

Class

Peqvred

Achieved

Total

ASCI

cs

11

7

Figure 7.1: Quota sheet

Once the interviewer has these quotas it is up to him or her to locate the appropriate number and type of people within the working area in the required time. Most interviewers knowing their area well will know exactly where to go to collect the type and age of person specified. Everyone knows for their own home town where the middle class and working class areas are. One can think of the place one would go to if one were particularly interested in collecting the views of younger people or of older people. Applying exactly that kind of local knowledge, the interviewers quickly and efficiently complete the quota set for them. The problem is that by selecting respondents from such stereotyped areas they are not entirely representative of the whole population, which is more mixed than the method described might reflect. What of the middle-class individual living in a generally working class area, or vice versa? What of the young person not to be found with the majority of young people, and so on?

Recognizing that a possible weakness in the validity of quota sampling lies in the degree to which the interviewer is able to influence the selection of final respondents, hybrid systems of sample selection have been developed. In random location sampling the interviewer is given a quota of interviews to complete, but may only do this within an area of named streets. Other quota controls may be established, perhaps to limit the proportion of respondents with easy-to-interview occupations such as men working on outside jobs or public transport. Quota controls are also used to include the correct proportion of working housewives, and of other characteristics relevant to the purpose of the survey.

The great advantages of quota sampling are its quickness and cheapness, both in generating the sample in the first place and in completing and controlling the fieldwork. In order to select the sample, no sampling frame needs to be devised. This is not only a saving in cost and time, but in certain circumstances it makes sampling possible when a sampling frame cannot be established and yet some important characteristics of the population to be sampled are known. In industrial and trade research, quota sampling is particularly useful since sampling frames are often difficult to construct yet the major characteristics of an industry or trade may be known. Another advantage of quota sampling is that it overcomes one of the problems of random sampling, in that although sampling frames may be quickly outdated, the population characteristics used as the basis of allocating quota samples are far more stable.

In carrying out the fieldwork the most important advantage of quota sampling is that interviewers do not have to find named individuals. This accounts for the greatest cost and time savings of the method. The interviewer simply screens likely individuals with a small number of classification questions and, if they meet the requirements of the quota, goes on to complete the full interview. If they do not meet the quota requirement, the interview is terminated and the interviewer continues knocking on doors or stopping other individuals in the street until he or she finds someone who does. This reduces the costs of quota sampling to about one-third to one-half that of random sampling.

As far as the quality of data is concerned, studies that have tested the results of quota samples against those of random samples have suggested that for the majority of commercial purposes these are perfectly acceptable. In these applications the cost benefits of the method outweigh the theoretical disadvantages, provided adequate control is exercised. This will include adequate training of interviewers for their important role in this type of sample selection procedure. The cost saving made in the application of quota rather than random sampling can be used to improve the quality of the data generated in other ways, say by increasing the sample size. Quota sampling is mainly used as an acceptable but cheaper alternative to random sampling for the purposes of most ad hoc market research surveys. The fact that this is the major method now in use appears to underline its widespread acceptability. Presumably its army of regular users has found that the system works well in practice. For social surveys, however, when it is often the views of minorities rather than majorities which it is required to represent, it is possible that quota sampling may not be sufficiently sensitive.

Online Survey Champion

Online Survey Champion

There are people all over the world trying to find ways to make money online. From stay at home moms looking to make a few extra dollars to college students and entrepreneurs, the allure of making your own hours and working from home or from the local coffee shop is very appealing.

Get My Free Ebook


Post a comment