A difficult year
From colleagues of Karl Jöreskog I received an invitation for a contribution to a book in honor of this important statistician.
It seems that my long term devotion to the statistical approach he had introduced (SEM) was appreciated.
I have used the opportunity to summarize in my contribution the procedures we had developed to tackle measurement problems in the social sciences.
It seems that my long term devotion to the statistical approach he had introduced (SEM) was appreciated.
I have used the opportunity to summarize in my contribution the procedures we had developed to tackle measurement problems in the social sciences.
The first program to predict the quality of survey questions (SQP.Dos)
Last year the characteristics of the many questions of the MTMM experiments were coded. Now it was possible to determine the effects of these different question characteristics on the quality of the questions. This analysis was very important for me because it was the final phase of a research that started 10 years before with the organization of the IRMCS group. The purpose was to create a prediction program for the quality of questions on the basis of characteristics of the questions. The analysis was not simple because of the large number of characteristics (some numeric, others categorical) and the high correlations between some of these characteristics. In the end I could report that the explained variance for the validity of the questions was quite high (.61) and for the reliability somewhat less (.47). This was better than I expected. With our IRMCS friends we finished this period of working together during our last meeting in Gent. It was as always a pleasant meeting as you can see.
Last year the characteristics of the many questions of the MTMM experiments were coded. Now it was possible to determine the effects of these different question characteristics on the quality of the questions. This analysis was very important for me because it was the final phase of a research that started 10 years before with the organization of the IRMCS group. The purpose was to create a prediction program for the quality of questions on the basis of characteristics of the questions. The analysis was not simple because of the large number of characteristics (some numeric, others categorical) and the high correlations between some of these characteristics. In the end I could report that the explained variance for the validity of the questions was quite high (.61) and for the reliability somewhat less (.47). This was better than I expected. With our IRMCS friends we finished this period of working together during our last meeting in Gent. It was as always a pleasant meeting as you can see.
I did not think that this would be the last result, improvement was certainly possible with more data from more countries but it was shown at least for data in three languages that this approach made sense and could be used to suggest which questions had to be improved, given that the predicted quality of some questions was too low. Soon this program turned out to be very helpful to detect bad questions in the European Social Survey as we will describe later.
A new course for Communications science
The second issue that I like to mention is that I received a message of the commission for the curriculum in Communication Science of the faculty. They expected me in the next academic year to teach a course which should be more statistically oriented. The course should be more selective, with special attention to the “reliability of sampling procedures”.
I had two problems with this proposal. The first was that it is not the task of the statistics course to select the students which can or can´t get a degree in Communications Science. The second problem was that the commission did not know what they were talking about because “reliability of samples” is not a topic that can be taught because it does not exist as such. Therefore I wrote them back that I was not willing to give a course on a topic that does not exist.
I had spent a lot of time to prepare a new course and now I got this reaction. I thought that it was enough and I decided to refuse to prepare any course.
Although I refused to teach they could not fire me because I had, as one of the last professors, obtained a permanent position as professor at the university. My reaction made me free of teaching new courses in this faculty but I also realized that this was more or less the end of my work at the university.
I had to look for other activities.