Many of these perceptions are consistently shared by the respondents in both surveys.
More than 70 % of the respondents felt that most people were basically dishonest.
Information about the respondent's social network was also collected over tly in the interview.
Although these results must, to some extent, reflect the metalinguistic awareness of the respondents, no attempt was made to evaluate metalinguistic awareness systematically.
The higher the score the more politically sophisticated respondents are.
As with the religious context measure, procedures of summing within households, subtracting the respondents' values, and recoding were utilized to create the measure.
Consequently, the respondents were asked first about the group used in the tolerance persuasibility experiment in 1996.
A much more complicated picture of elite change emerges, however, when we analyse the relationship between attitudes to change and respondents' personal characteristics.
