Quota sampling has been introduced by George Gallup in 1935. It was used by him to successfully predict the winner of the presidential elections in the years 1936, 1940 and 1944.
Quota sampling is a method of stratified sampling. The strata’s are decided on the basis of pre-determined characteristics. These characteristics are determined by the researcher. But, the selection within the strata is non random. Selection within the strata is at the discretion of the interviewer. In simpler terms, it is a systematic effort to force the sample fit in a profile by using quotas.
There are two types of Quota sampling: proportional and non-proportional.
Proportional quota sampling: The population is divided into quotas as per the requirement. The responses are recorded as per the quota. If the quota requirement is fulfilled, further responses of that particular quota will not be recorded. This method is restrictive in terms of quotas.
Example: If the population has 40% females and 60% males, the responses will be recorded proportionally. If the quota of 40% females is fulfilled, no more responses of females will be recorded. Rest 60% responses will be recorded from males.
Non-proportional quota sampling: This method is less restrictive when compared to the probabilistic quota sampling method. Here the quota is restricted to minimum sample requirements. This method assures that smaller groups are adequately represented in the sample.
Example: If the quota is fixed as 100 for a sample, then 100 responses from each stratum are collected. Any number of responses above 100 is not restricted.
Let us see how this method was used to predict the results of election polls.
The sample size was determined by the researcher. The number of men, women, blacks, whites, below 40 years age and 40 years age etc., was pre-determined. These numbers are in proportion to the electorate at large.
When every important characteristic of the population is considered while setting up the quotas, the sample will produce a good cross section of the population. This helps in making accurate predictions.
The same method was applied for the 1948 election between Thomas Dewey and Harry Truman. The poll was conducted with a sample size of 3250. Each individual in the sample was interviewed by a professional interviewer. The quotas for each interviewer were pre-determined.
Based on the results, Dewey was expected to win the polls. But the actual results were reversed. Truman emerged victorious. With this, statisticians understood that there were serious flaws in the quota method.
Few identified problems with quota sampling technique are:
- Possibility of missing few criterions which affect the voting decision of people.
- Who is in the sample is determined by human as selection within the strata is at the discretion of the interviewer.
Though the predicted percentage of votes is different from the actual results in the years 1936, 1940 and 1944, Gallup was able to predict the winner correctly merely due to luck. The spread between the candidates was high to cover up the error. But, he ran out of luck in 1948.
The major learning from this is that human intervention in sample selection is always subjected to bias.
Shivani Kolala_ SectionA_pgp13028