Students' rates of disclosure on sensitive sexual behaviours : a comparative study using methods of the Unmatched Count Technique 1 (UCT 1), Unmatched Count Technique 2 (UCT 2) and Self-Report Questionnaire.
Sexual behaviour can be seen as sensitive behaviour because it is highly private to the participant and can often be “laden with negative evaluation” (LaBrie & Earleywine, 2000, p. 321). As a consequence, participants may then give socially desirable answers to surveys, interviews or self-reports about sexual behaviours (LaBrie & Earleywine, 2000). However, there is another way of analysing sensitive behaviours that is different from methods used before such as the Self Report Questionnaire, Audio Computer-Assisted Self-Interview, Face To Face Interview and Informal Confidential Voting Interview. This method is known as the Unmatched Count Technique (UCT). The UCT provides a way of asking about sensitive behaviours in an indirect way. If the UCT results in higher base rates, this information could help any researcher or persons working in the health industry to “assess risk within a given population” in order to target interventions, which may be needed for that population (LaBrie & Earleywine, 2000, p. 322). In research, survey questionnaires have effectively been used to investigate constructs of concern as defined by the researchers and what was considered relevant and problematic within a certain community. Some survey questionnaires measured key constructs such as trait, attitude and behaviour (Abowitz & Toole, 2010). Furthermore, surveys have enabled researchers to study large populations thereby retrieving a large amount of information about that population (Abowitz & Toole, 2010). However, these surveys rely on participants’ self-reports rather than their nonverbal behaviour (Abowitz & Toole, 2010). The reliability in survey methods can be found when answers of the participants remain consistent and if the data converges “on a common pattern or result” allowing the researcher to make generalizations about the results (Abowitz & Toole, 2010, p. 110). However, the very nature of how the survey is constructed can be a threat to the validity of the survey and the outcome of the data. For example, if the format of the questions and how the questions are worded influences how the participant will interpret and answer the question, these can result as being problematic to the validity of the survey especially when the survey is not clearly worded (Abowitz & Toole, 2010). In addition, one of the biggest threats to reliability in survey methods is when participants are self-reporting and whether the answers they are reporting are considered to be true or not (Walsh & Braithwaite, 2008). According to Kays, Gathercoal and Buhrow (2012), it was found that many participants “skew their presentation in order to enhance social desirability” and this is especially the case when being surveyed about sensitive information or behaviour that is considered to be socially undesirable (p. 252). Furthermore, when participants are surveyed about sensitive information, item non-response will be evident as the respondent is concerned “with confidentiality of disclosure, particularly regarding highly sensitive information” (Kays et al., 2012, p. 525). In summary, the survey method of analysing behaviour that is operationalised as sensitive and seen as socially undesirable behaviour, cannot be considered to be completely valid or reliable due to the risk of underreporting or item non-response and several other reasons later discussed in this thesis (Droitcour, Casper, Hubbard, Parsley, Visscher, & Ezzati, 1991). In past research it has been found that base rates for risky sexual behaviour have been underestimated and many techniques have been researched to combat the underreporting that participants give when asked about risky sexual behaviour (Droitcour, et al., 1991). Survey methods have been viewed as problematic in combating this underreporting of sensitive information as they have been unsuccessful in inspiring trust, resulting in non-response and social desirability bias and therefore regarded as unreliable (Coutts & Jann, 2008). However, one technique that established “higher rates of truthful self-reporting” is the Unmatched-Count Technique 1 (UCT 1) (Walsh & Braithwaite, 2008, p. 49). However, there is another version of the UCT (which, for the purpose of this study, will be referred to as the UCT 2) that could also be more effective and yield higher base rates for self-reporting. This method is proposed by Chaudhuri & Christofides (2007). The UCT 1 is constructed by assigning participants to two independent groups. The one group receives a series of statements that consist of non-sensitive innocuous items and the participants are asked to report how many of these items are true or apply to them (Chaudhuri & Christofides, 2007). The second group will receive the same non-sensitive innocuous items as in the first group; however, one additional statement is added. This statement will be considered as the sensitive item (Dalton, Wimbush, & Daily, 1994). The participants in the second group will also be asked how many of these items are true or apply to them. From these two groups an “estimate of the base rate for the sensitive behavior can be obtained” (Dalton et al., 1994, p. 818). In contrast, the UCT 2 works somewhat differently to the UCT 1. Chaudhuri & Christofides (2007) argues that the participants need to have an increased sense that the items in the list serve a meaningful purpose and thus, the participants will increase their “level of cooperation” (p. 592). Therefore, Chaudhuri & Christofides (2007) suggests that the innocuous items should not be unrelated to the sensitive item (as in the UCT 1) but should be similar to the sensitive item. This is how the UCT 2 will be constructed within this thesis. This study forms part of the PhD of the supervisor of the current study. Therefore, there were also several other methods investigated to conclude which one seemed more reliable and valid in assessing what is operationalised as sensitive behaviour. These methods were the Audio Computer Assisted Self-Interview (ACASI), Self-Report Questionnaire (SRQ), Informal Confidential Voting Interview (ICVI) and the Face to Face Interviewing (FTFI). This study will mainly focus on the SRQ and the UCT 1 and UCT 2.