We are testing a new beta website for gov.scot go to new site

Attitudes to Discrimination in Scotland

Listen

ATTITUDES TO DISCRIMINATION IN SCOTLAND

APPENDIX ONE LOGISTIC REGRESSION

Table A1-1 Logistic Regressions of "sometimes there is good reason to be prejudiced" versus "Scotland should do all it can to get rid of prejudice".

Coefficient

Standard error

(1) Sociological Variables

Educational Attainment

Degree

-.54

.22 **

Professional

-.46

.20 **

Higher

-.24

.19

Standard Grade

.09

.21

Below Standard Grade

.10

.18

Foreign

-.61

.61

(None)

Age Group

18-24

-.30

.25

25-34

-.55

.21**

35-44

-.54

.19**

45-54

-.30

.19

55-64

-.44

.19**

(65 and over)

Constant

-1.07

Nagelkerke R 2 = .04

(2) Economic Variables

Respondent's Economic Position

In Education

.19

.36

Unemployed

.01

.32

Retired

.69

.16**

Other

.55

.18**

(In employment)

Self-rated economic hardship

Living comfortably

-.61

.19 **

Coping on income

-.72

.19 **

(Having difficulty)

Constant

-.88

Nagelkerke R 2 = .04

(3) Psychological Variables

Preferred kind of area

Different kinds of people

-1.18

.14**

Depends

-.65

.17**

(People are the same)

Constant

-1.10

Nagelkerke R 2 = .08

(4) Combined Model

Preferred kind of area

Different kinds of people

-1.14

.15**

Depends

-.62

.18**

(People are the same)

Self-rated economic hardship

Living comfortably

-.75

.18 **

Coping on income

-.78

.18 **

(Having difficulty on income)

Age Group

18-24

-.14

.26

25-34

-.52

.21**

35-44

-.51

.19**

45-54

-.39

.19**

55-64

-.56

.20**

(65 and over)

Constant

-.97

Nagelkerke R 2 = .10

**significant at the 5% level.

This table shows the results of four logistic regression models constructed using a forward stepwise procedure. In the first model all those variables in Table 5-2 (including the full range of categories for each rather than just the end points, but excluding party identification as this itself may be considered in part a product of the other characteristics in the table) were eligible for inclusion if they were significant at the 5% level. In the second and third models all the economic (second model) and psychological (third model) variables were analysed in a similar fashion. Finally the fourth model was constructed by using the same forward stepwise procedure where all those variables that appeared in one of the first three models were eligible for inclusion, with the exception of respondent's economic position as this variable was evidently picking up an age difference rather than one between those in work and those unemployed.