Volunteer Bias

Differences between Volunteers and Nonvolunteers

An underlying problem in volunteer bias is that volunteers and nonvolunteers are different in important ways. In an extensive review, Robert Rosenthal and Ralph Rosnow examined differences between volunteers and nonvolunteers. They reported that, in general, volunteers are more educated, come from a higher social class, are more intelligent, are more approval-motivated, and are more sociable. Volunteers are also more arousal-seeking, unconventional, nonauthoritarian, and nonconforming, and they have higher levels of empathy and lower levels of trait-anxiety. Females are more likely to volunteer than males, and Jewish people are more likely to volunteer than Protestants or Catholics. Medical research also has found that volunteers are generally healthier and are more likely to follow the treatment plan provided by their physician. Volunteers also have lower morbidity and mortality rates, are less likely to smoke, and are less likely to abuse alcohol.

Overcoming Volunteer Bias

Volunteer bias can emerge in any research study in which sampling is required. As the refusal rates to volunteer increase, the potential for volunteer bias is increased. A reduction of this bias can therefore be made by methods intended to increase rates of volunteering. For example, subjects are much more likely to volunteer for studies that they are interested in, and so researchers should ensure that the volunteer request captures the interest of the [Page 1609]subject as much as possible. Also, subjects are less likely to volunteer if the topic is sensitive or threatening. Researchers can try to make the study as nonthreatening and comfortable as possible and ensure anonymity and confidentiality to the subject. In addition, subjects are more likely to volunteer if the study is perceived to be theoretically or practically important or if the recruitment is made by a person that the subjects are familiar with. Volunteer rates also increase with the perceived level of authority of the recruiter. Increasing the incentive to participate is also a good way to improve volunteer rates. Furthermore, researchers should be careful to make their studies as short and simple as possible; volunteering tends to decrease when the study requires a major commitment from subjects or if the study is not clear. The volunteer bias can be difficult to overcome, but researchers can reduce its likelihood with diligent planning and recruitment practices.

Example: Volunteer Bias in Sexuality Research

Although the volunteer bias can emerge in many research areas, it seems to be particularly problematic with topics that might be deemed sensitive by subjects. One such area is human sexuality research. To understand sexual behavior, researchers might want to sample a large group of subjects and ask them questions about their sex lives. Many potential subjects might be uncomfortable answering personal questions and will choose to not volunteer. Other subjects will have no reservations when talking about their sex lives. If the researcher gets data only from subjects that are more open about sex, the results might not represent the population as a whole. Research on the volunteer bias in sexuality research has confirmed several differences between volunteers and nonvolunteers. It has been reported that men are much more likely to volunteer for sex studies than women. In general, volunteers are more sexually experienced, less sexually inhibited, have less sexual guilt, and have more positive attitudes about their sexuality than nonvolunteers. Some studies have found that male volunteers are more likely to have sexual dysfunctions than male nonvolunteers and that female volunteers are more likely to report history of sexual trauma than female nonvolunteers. Other studies find volunteers have more positive attitudes about erotica, masturbate more frequently, engage in more noncoital sexual experiences, have increased reports of homosexual behavior, and have more experience with less traditional sexual interactions.

...