In order to find out which items should be used and how they could be adjusted to fit into the questionnaire, preliminary research in the form of an expert review was held. Expert review is an empirical pre-testing method because it involves the systematic collection and processing of reviews from sources outside the immediate design team (Ramirez, 2002). As a pre-test the questionnaire was given to 5 experts, most of them working for Tactus and also to my tutor at the University. The questionnaire was given to a general questionnaire design expert who has considerable experience in writing and evaluating questionnaires and in communicating with the target group. The expert reviewed the wording of the questions, the structure of questions, the response alternatives, the order of questions, instructions to interviewers for administering the questionnaire, and the navigational rules of the questionnaire. The questionnaire was also given to three additional subject matter experts who know about the topic under investigation. Those experts assessed whether the content of the questions was appropriate for measuring the intended concepts (Groves et al., 2004).
The experts were instructed to fill in the questionnaire, thereby using the plus-and-minus method (De Jong & Schellens, 2000). Putting down plus and minus symbols where something positive or negative was noticed it was possible to keep the normal flow of filling in the questionnaire and still get feedback on the construction of the questionnaire. In the follow-up interview, a couple of comments were made - mostly positive about the content and statement of the questions, and mostly negative about the phrasing. This was not surprising since the Dutch questionnaire items were either translated from an English questionnaire, or self-constructed by a native German. With the help of the pre-test feedback, the questionnaire was revised. The instructions and outcome protocol of expert review can be found in appendix D and E.
As pre-test the adapted questionnaire was send to 10 people from the waiting list and 10 people who had already started the treatment. There were five additional questions added at the end of the questionnaire to evaluate the clarity and structure of the questions. There was also the opportunity given to comment on positive or negative aspects of the questionnaire and report about expected or missing questions. The pre-test can be found in Appendix C. The pre-test was filled in by 9 participants in total and there were no negative comments made about the questionnaire. Respondents rated the questionnaire as understandable and easy to fill in. Respondents rated the phrasing of the questions as clearly and did not miss an important question.
Different constructs were measured in the questionnaire which needed to be tested as to their reliability. A Cronbach’s alpha of minimal .60 for constructs of minimal three items was hereby required (De Vellis, 2003). Constructs with less than three items were tested on their correlation instead of Cronbach’s alpha. The results are presented in table 3.2.
Note. For constructs with less than 3 items correlation was analyzed instead of Cronbach’s alpha
**p < .001
The results are shown in relationship with the constructs from the model and the research questions. Parts of the analysis are based on Codd and Cohen (2003) who measured relationships between the “Theory of Reasoned Action” variables and the intention of professional help seeking services for alcohol abuse.
In order to know central tendencies, frequency tables about antecedents were made. Most of the data in this paragraph are nominal, which facilitates the differentiation into groups like “males” and “females”. Additionally to analyzing the frequencies of the demographic variables, the data of the anonymous participants was compared to the data of non-anonymous participants. The aim of this analysis was to detect the possible differences between the demographic details of participants between the two treatment options.
Most participating clients (N=62) had Dutch origin (91,9%). The participants’ gender is female for the most part. About 70% of the participants were aged between 36-55 years and 15% are older than 55 years. The results show that the highest percentage of the respondent s had higher education. Concerning the occupation status it can be seen that most respondents are employed. The exact distribution can be seen in table 4.1
The groups are compared using the Chi-Square test to detect the differences between the anonymous and non-anonymous group in demographics. The differences will be confirmed if a statistically significant (p<0.05) difference is found between the two groups.
Table 4.2 shows the different percentages for participants in the anonymous and non-anonymous groups after having applied a Chi Square test for independent samples. The results indicate that there is a statistically significant difference between the anonymous and non-anonymous treatment option in gender distribution (X²(1) =7.323; p=0.007). The gender distribution differs significantly between the anonymous and non-anonymous treatment variant. There are significantly more women and less man participating in the anonymous treatment in comparison to the non-anonymous treatment where the gender distribution is reversed. For the remaining variables there could not be found a significant difference between the anonymous and non-anonymous variant.