A new wrinkle on an old concern: are the new ethics review requirements for explicit warnings in consent forms affecting the results of sexuality research?

Abstract: Consent forms required by Institutional Ethics Review Boards have changed historically to protect universities and participants but may inadvertently affect sex research through the provision of detailed warnings. We tested this contention using an experiment (N = 135) and a survey (N = 75). Three types of consent forms were used to manipulate the degree of warning ("procedures only," "mild warnings," "explicit warnings"). We measured participants' mood and expectations after reading the consent form but prior to any study tasks. Following the study, we assessed post-test mood and participants' expectations and experience in the study. Consent forms did not affect mood or survey data. However, warnings influenced participants' evaluations of experimental stimulus materials, with more negative evaluations produced by warnings and detailed content than by procedures only wording. Moreover, more detailed content in consent forms did not provide participants with a better sense of what the studies would entail. We discuss the need to provide information to participants while minimizing demand effects.

Key words: Research ethics review, consent forms, bias, REB requirements.

Introduction

Sexuality researchers across disciplines explore important and sensitive aspects of human experience and well-being. None of these topics are easy to investigate. The researchers who do this work spend many hours thinking through the ethical issues involved in their studies and carefully constructing their studies to balance the need for knowledge and the desire to protect participants from harm. For example, in our research on the effects of pornography (Senn & Radtke, 1990) participants must be exposed to sexual imagery including some which may be upsetting. Study stimuli must be described sufficiently well in recruitment and consent procedures to ensure that participants are adequately prepared to provide informed consent while not using pejorative labels or other descriptors that would inadvertently "suggest" to participants what they should think or feel about the images. In research on violence against women and sexuality (Senn, Desmarais, Verberg, & Wood, 2000; Senn & Desmarais, 2004; Simpson & Senn, 2003), participants must be asked highly personal questions about their sexual experiences. The general content of survey questions must be alluded to in recruitment but without suggesting that only some participants need volunteer.

Researchers have long been aware of the potential participant selection biases associated with sexuality research and have explored how biases vary according to the procedures used (Saunders, Fisher, Hewitt, & Clayton, 1985; Wolchik, Braver, & Jensen, 1985; Wolchik, Spencer, & Lisi, 1983). In general, the more intrusive or involved the study, the more strongly selection biases operate. In our own research, we have compensated to some degree for subject selection biases by using random sampling instead of volunteering (Khoo & Senn, 2004) and, where selective volunteering still appears to be occurring, by comparing our samples to the population to ensure representativeness (Senn et al., 2000). However, as anyone doing academic research in the past ten to 15 years can attest, the consent forms we are required to present to participants have changed dramatically. In the 1990s, universities seemed to become more concerned about liability issues related to the conduct of research in Canada (and the U.S.) and these began to influence ethics review (Palys & Lowman, 1998). When the Tri-Council Policy Statement was adopted in 1998 and the format and process of university Institutional Ethics Review Boards (IRBs) was regularized, requirements regarding consent forms noticeably changed. Consent forms have become longer in North America (Mann, 1994) with substantially more content than the focus on the tasks and procedures that participants would encounter in the single-page consent forms of the past.

We believe that some of these changes have not necessarily had the effect the review panels intended. We argue, as did Breckler (2005) in a recent article, that researchers must determine whether the procedures used to protect participants or to inform them of their rights are affecting the research data. This is one instance in which we are hoping to support the null hypothesis but if we do not, we will need to find ways to eliminate these effects or compensate for them when drawing conclusions from our research studies.

Historical roots and precedents