Ethnic attitudes as a function of ethnic presence

• Published in: Canadian journal of behavioural science Vol. 28; no. 3; p. 171
• Main Author: Kalin, Rudolf
• Format: Journal Article
• Language: English
• Published: Ottawa Canadian Psychological Association 01.07.1996
• Subjects: Cultural identity; Ethnicity; Social conditions & trends; Canada
• ISSN: 0008-400X; 1879-2669
• DOI: 10.1037/0008-400x.28.3.171

Attitudes, assessed through comfort ratings, towards 12 ethnic and racial groups in Canada were measured in a national survey in 1991. The attitudes towards these groups were related to their ethnic presence in the survey respondents' Census Subdivision (equivalent to a municipality), as revealed in the 1986 Census. Separate analyses were carried out in Quebec and the rest of Canada. Small but statistically significant direct relationships were found between most ethnic attitudes and log 10 transformed ethnic presence. The relationships were statistically significant for Italians, Ukrainians, Germans, Jews and Portuguese in Quebec, as well as in the rest of Canada. In Quebec, the relationship was also significant for British. In the rest of Canada, but not in Quebec, further significant relationships were obtained for Arabs and French. Linear relationships were not statistically significant in Quebec, as well as the rest of Canada for Chinese, Native Indians, and West Indian Blacks. The direct relationships between ethnic attitudes and ethnic presence observed for most groups indicate that attitudes towards a particular group are more positive to the extent that the group is well represented in the geographic region of the respondent. The results were explained in terms of the contact and mere exposure hypotheses. Ethnic presence data showed negative skewness as in the earlier study by [Kalin] and Berry (1982) and they were therefore subjected to logarithmic transformations. Pearson correlation coefficients were computed between attitudes towards a particular group and the log10 transformed ethnic presence of that group in the Census Subdivision of the respondent, with each correlation being computed among nonmembers only. In order to have a determinate log10 value for zero, the value of 1 was added to the presence per 1000. Additional checks for linearity of the relationships between attitudes and log transformed ethnic presence were performed as follows. Categories of the ethnic presence variables (in units of per 1000 population) were created such that the midpoint of the categories would form a logarithmic scale with approximately equal intervals. The category intervals were the same as those used in the earlier study by Kalin and Berry (1982), namely 0, just greater than 0 to just less than 5, 5 to just less than 20, 20 to just less than 50, 50 to just less than 100,100 to just less than 300, and 300 and higher. These same intervals were used to categorize the ethnic presence of all groups. When the number of respondents in the lowest and highest categories were less than 20, the extreme categories were collapsed to contain at least 20 respondents in order to yield relatively stable means. The means of the attitude variables within these categories were then computed for nonmembers, as were the means of the log10 transformed presence data. The points formed by the attitude and log10 transformed presence data were then graphed as bivariate scatter plots. The correlation coefficients between the two sets of means were also computed and the linear trend lines plotted. It is apparent consistently in Figures 1 to 3 that the regression lines for Quebec are below those for the rest of Canada, suggesting that residents of Quebec have generally less favourable ethnic attitudes than those from the rest of Canada. To confirm this suggestion, t - tests were run for all the ethnic attitudes among non - members, with the classification variable being region (Quebec vs rest of Canada). Tests for all ethnic attitudes showed that the means were significantly lower among Quebec respondents (p <.0001, two - tailed), except for attitudes toward French, where there was not a significant difference. Despite the overall less favourable ethnic attitudes in Quebec, the relationships between the attitudes and ethnic presence were slightly stronger for Quebec, than the rest of Canada, for Italians, Ukrainians, Germans, and Portuguese. In order to examine the possibility that these relationships were primarily due to non - French non - members, who generally have more positive ethnic attitudes than French Canadians (Berry and Kalin, 1995), the individual correlations between ethnic attitudes and ethnic presence were run in a sample of Quebeckers of French origin (excluding British and other ethnics). These correlations were virtually identical to those for all non - members in Quebec, as reported in Figures 1 - 3. It appears therefore that the positive linear relationship between ethnic attitudes and ethnic presence holds equally well for Quebeckers of French as well as of other origins.