Quota Sample

Quota sampling is a method of selecting respondents for surveys. Survey administrators assign quotas to interviewers that define groups of respondents by using a few key demographic characteristics. Interviewers fill their quotas by choosing individuals whose characteristics match these characteristics as respondents. When the responses are combined for all of the respondents, the characteristics of the sample precisely match specified demographic characteristics of the survey population. For example, an interviewer may be assigned a quota of four African Americans, eight whites, two Hispanics, and two Asian Americans, half of which live in the central city and half of which live in the suburbs. When all of the interviews are completed, the aggregate statistics for the sample and the study population will match exactly for the demographic characteristics that were included as interview criteria. However, matching on aggregate statistics does not ensure that the sample will accurately reflect the population on the variables that are of interest in the study.

The most infamous example of the bias that can occur with quota sampling involves polling for the 1948 presidential election. This election pitted Harry Truman, the Democratic candidate who had assumed the presidency when Franklin D. Roosevelt died, against Thomas Dewey, the Republican candidate. Three polling firms—Roper, Gallup, and Crossley—all declared Dewey as the winner based on interviews that had been conducted using quota sampling methods. Of course, Truman won with just under 50% of the popular vote, beating Dewey by nearly five percentage points. The quota samples were carefully matched to census data on age, race, gender, rent paid, and residence, but the interviewers had chosen too many Republicans. Republicans were easier to interview, either more willing to participate or more approachable, and thus Democrats were underrepresented in the interviews. After that election, newspapers were filled with stories about the failure of the pollsters, as well as many commentators. Subsequently, all three polling firms dropped quota sampling techniques.

The failure to accurately predict the 1948 election provided an impetus toward the use of probability samples and very limited use of quota samples for social science research as well as polling. Most pre-election polling relies on random digit dialing (RDD) to select households from which adult respondents are selected at random, or random computergenerated selection of a sample from an automated list of eligible voters. Specific individuals are identified by these survey procedures, and interviewers are not allowed to substitute respondents, thereby eliminating interviewer discretion and selection bias.