Convenience Sampling

This sampling technique has been responsible for some celebrated errors in research. When the Hite report claimed that 75% of women had extramarital affairs, or the newspaper report in 1948 that predicted Dewey would defeat Truman. The errors in these claims was due to a sampling process that biased the results.

A convenience sample is either a selection of participants that are easily available, like standing on a corner and asking passerby’s their opinion on a subject, or the self selection of individuals who choose to participate in a survey, such as one in a magazine where you complete a card, tear it out, and mail it. In both types of convenience samples, the problem is that the participants are likely to represent a biased portion of the population as a whole, at least we can not claim that they are representative.

Lets say you stand on a corner during rush hour and stop people as they pass by asking them to complete your survey. What biases would you expect if you were trying to judge say, women’s views on the role of mothers in society? To start with who is likely to be on the street at rush hour, working women or stay at home moms? Who are you likely to approach, the angry looking woman charging down the street, or the pleasant looking one casually standing at the corner?

These subtle, and not so subtle biases will influence the results of your study in unknown ways. As a result you will not be able to say anything about women’s views in general. You will only be able to report what the specific women you spoke with said, and then you will have no way of knowing whose views those results represent.

The care with which we sample is related to what we want to know and how accurate the results need to be. For some questions, we may want to get an idea of what some of the issues are related to a particular topic, such as merchant’s views of panhandling, or panhandler’s views of new laws affecting them. In this case, we are not trying to generalize with any authority to the whole population, and the results of the survey would not make that claim. Rather, we want to be able to describe some of the issues that are of concern to participants, without being able to describe what percentage of the population is concerned with a particular issue, how strongly the population feels, etc.

In this case we can use a convenience sample, while recognizing that who we talk to may cause us to emphasize a different set of issues than if we had talked with someone else. As long as we do not claim that the results are representative of the group as a whole, we can safely use this sampling technique to illuminate what a convenience sample of respondents had to say.