In a previous analysis of this news series, we presented a table showing how the percentage of those who do not categorically rule out "if Germany had a strong leader who is above the law" changes with attitudes towards immigration and authority as an educational goal. From this previous analysis (heading "Respect", December 14, 2024), we already know that the percentage in question is higher the more critical the attitude towards the form of immigration in question is or the more positively the principle of authority is viewed as an educational goal. It also showed that the two attitudes examined have their effect independently of one another.
We come to the same conclusion if we also take into account another variable: the degree of satisfaction with how our democracy works. If you click on the table to enlarge it, you will see that the percentage in question - viewed row by row - is higher the more positively the principle of authority is viewed as an educational goal (x1), regardless of the attitude towards the form of immigration in question (x2) and the satisfaction with the way our democracy works (x3). At the same time, a comparison of the percentage in question in the right-hand column for each of the three color-bordered sub-tables shows that the more critical the attitude towards the form of immigration in question is, the more likely it is that a strong leader above the law will not be categorically ruled out.Thirdly, a column-by-column comparison of corresponding percentages across the three sub-tables (right-hand column) shows that the percentage of people who do not categorically rule out a strong leader above the law also increases with dissatisfaction with the way our democracy works; The main difference lies between those satisfied with it (green sub-table) and the other two groups (blue and orange bordered sub-tables).
The structure of the percentages tends to suggest the same conclusions if we do not refer to the column-wise comparisons to the column on the far right, but to the columns within the sub-tables. Analogous to the row-wise (x1-related) comparisons of the percentages, the x2-related comparisons would take into account that x1 and x3 also affect the said percentage; and at the same time the x3-related comparisons would take into account that x1 and x2 also have a corresponding influence.
To confirm our conclusion that the observed structure of the percentages is compatible with the assumption that x1, x2 and x3 influence the variation in y independently of one another, we "translated" this contingency table analysis into a regression model that tests precisely this assumption. Since y is a binary variable, we calculated a logistic regression. And to test the assumption of an independent influence, we compared a main-effects model with a model that also takes into account statistical interaction effects between x1, x2 and x3 in their effect on y. The result shows that we can assume a pure main-effects model here.
Legend to the table:
y indicates: “How acceptable for you would it be for Germany to have a strong leader who is above the law? In each case, the percentage of people who would not categorically rule out such a leader is shown.
x1 indicates: “To what extent do you agree or disagree that obedience and respect for authority are the most important values children should learn? 5pt scale recoded to 3 categories: 0=disagree (strongly); 1 = neither agree nor disagree; 2=agree (strongly). x2 indicates: “…, to what extent do you think Germany should allow people from the poorer countries outside Europe to come and live here?” 4 pt scale recoded to 3 categories: 0= allow many to come and live here, 1=allow some, 2= allow a few /allow none. x3 indicates: “ … on the whole, how satisfied are you with the way democracy works in Germany?“ The 11pt scale ranges from 0=extremely dissatisfied”, over 1, 2, 3, …, to 10=”extremely satisfied”. Here, this scale is recoded to 3 categories: (0, 1, 2, 3, 4 = 0)(5, 6, 7 = 1)(8, 9, 10 = 2). “anweight” – weighted frequency distribution. In each cell, the number of respondents to whom a percentage refers is indicated in grey (percentage basis). All figures rounded to 0 decimal places. *Includes a small rounding inaccuracy.
For information purposes only: The two logistic regressions mentioned above yielded the following estimates. We show the respective estimate of the b effect and – in brackets – this value divided by its standard error, b/s.e. Main-effects model: b0= -1.378 (-17.02); x1: b1=0.691 (18.66); x2: b2=0.213 (5.40); x3: b3= -0.281 (-8.05); Pseudo-R2 = 0.068 (n=8,135)
Model that additionally estimates statistical interaction effects: b0= -1.308 (-9.28); x1: b1=0.689 (8.07); x2: b2= 0.237 (2.48); x3: b3= -0.417 (-4.57); x1x2: b4=-0.048 (-0.93); x1x3: b5=0.054 (1.11); x2x3: b6= 0.048 (0.96); Pseudo R2 =0.069 (n=8,135).
Due to the direct reference to the table shown above, the variables were included in the calculations in recoded form, just as there.
References
- Source questionnaire: European Social Survey (2020). ESS Round 10 Source Questionnaire. London: ESS ERIC Headquarters c/o City, University of London
- Data: European Social Survey European Research Infrastructure (ESS ERIC). (2023). ESS10 Self-completion - integrated file, edition 3.1 [Data set]. Sikt - Norwegian Agency for Shared Services in Education and Research. doi.org/10.21338/ess10sce03_1
- Documentation: European Social Survey European Research Infrastructure (ESS ERIC). (2022). ESS10 Data Documentation. Sikt - Norwegian Agency for Shared Services in Education and Research. doi.org/10.21338/NSD-ESS10-2020
On the German sub-study of the European Social Survey, Round 10
- Scientific management: GESIS Leibniz Institute for the Social Sciences (www.gesis.org)
- Implementation: infas Institute for Applied Social Sciences (www.infas.de)