+ All Categories
Home > Documents > FEDERAL AND PROVINCIAL POLITICAL PARTICIPATION IN CANADA: SOME METHODOLOGICAL AND SUBSTANTIVE...

FEDERAL AND PROVINCIAL POLITICAL PARTICIPATION IN CANADA: SOME METHODOLOGICAL AND SUBSTANTIVE...

Date post: 29-Sep-2016
Category:
Upload: mike-burke
View: 215 times
Download: 3 times
Share this document with a friend
15
Research note I Note de recherche FEDERALAND PROVINCIAL POLITICAL PARTICIPATION IN CANADA: SOME METHODOLOGICAL AND SUBSTANTIVE CONSIDERATIONS MIKE BURKE / University of British Columbia HAROLD D. CLARKE / UniVerSity Of Windsor LAWRENCE LeDUC / University of Windsor Se basant sur I’ktude des donnkes de I’klection nationale de 1974, les kchelles mesurant les taux de participation dans la politique au niveau du fkdkral et du provincial sont dkveloppkes. Les analyses utilisant ces kchelles rkvtlent que - bien qu’il existe des variantes inter-provinciales dans la participa- tion au niveau fkdkral, particulitrement dans la politique provinciale - au niveau individuel, il y a de fortes correlations positives entre le taux au niveau de la participation fkdkrale et celui de la participa- tion provinciale dans toutes les provinces. En plus, il semble que les Canadiens soient plus actifs politiquement que les ktudes antkrieures ne I’indiquent. En ce qui concerne les corrklatifs de partici- pation, les donnkes suggtrent que les variables conventionnelles socio-dkmographiques et les variables d’attitudes ne donnent aucune explication adkquate quant B la participation soit la politique f6dCrale ou B la politique provinciale. Des analyses supplkmentaires indiquent qu’il pourrait y avoir des diffkrences rkgionales significatives dans la culture politique canadienne, ce qui expliquerait les variations dans les taux de participation. Employing the 1974 national election study data, the authors develop scales measuring participation rates in Canadian federal and provincial politics. Analyses using these scales reveal that, although there is discernable interprovincial variance in participation in federal, and especially, provincial politics, at the individual level, there are strong positive correlations between federal and provincial participation rates in all provinces. Further, it appears that Canadians are somewhat more active politically than earlier studies have indicated. Regarding correlates of participation, the data suggest that conventional sociodemographic and attitudinal variables do not provide adequate explanations of participation in either federal or provincial politics. Additional analyses indicate the possible significance of regional differences in Canadian political culture for understanding variations in participation rates. Few students of Canadian politics would dis- guishing features of democratic politics. With agree with the proposition that the opportunity regard to the study of political participation in for widespread and meaningful participation in Canada, existing inquiries have focused almost politicial processes is one of the crucial distin- exclusively on voting and campaign-related Rev. canad. SOC. & Anth./Canad. Rev. SOC. & Anth. 15(1) 1978
Transcript

Research note I Note de recherche

FEDERAL AND PROVINCIAL POLITICAL PARTICIPATION IN CANADA: SOME METHODOLOGICAL AND SUBSTANTIVE CONSIDERATIONS

MIKE BURKE / University of British Columbia HAROLD D. CLARKE / UniVerSity Of Windsor LAWRENCE LeDUC / University of Windsor

Se basant sur I’ktude des donnkes de I’klection nationale de 1974, les kchelles mesurant les taux de participation dans la politique au niveau du fkdkral et du provincial sont dkveloppkes. Les analyses utilisant ces kchelles rkvtlent que - bien qu’il existe des variantes inter-provinciales dans la participa- tion au niveau fkdkral, particulitrement dans la politique provinciale - au niveau individuel, il y a de fortes correlations positives entre le taux au niveau de la participation fkdkrale et celui de la participa- tion provinciale dans toutes les provinces. En plus, il semble que les Canadiens soient plus actifs politiquement que les ktudes antkrieures ne I’indiquent. En ce qui concerne les corrklatifs de partici- pation, les donnkes suggtrent que les variables conventionnelles socio-dkmographiques et les variables d’attitudes ne donnent aucune explication adkquate quant B la participation soit la politique f6dCrale ou B la politique provinciale. Des analyses supplkmentaires indiquent qu’il pourrait y avoir des diffkrences rkgionales significatives dans la culture politique canadienne, ce qui expliquerait les variations dans les taux de participation.

Employing the 1974 national election study data, the authors develop scales measuring participation rates in Canadian federal and provincial politics. Analyses using these scales reveal that, although there is discernable interprovincial variance in participation in federal, and especially, provincial politics, at the individual level, there are strong positive correlations between federal and provincial participation rates in all provinces. Further, it appears that Canadians are somewhat more active politically than earlier studies have indicated. Regarding correlates of participation, the data suggest that conventional sociodemographic and attitudinal variables do not provide adequate explanations of participation in either federal or provincial politics. Additional analyses indicate the possible significance of regional differences in Canadian political culture for understanding variations in participation rates.

Few students of Canadian politics would dis- guishing features of democratic politics. With agree with the proposition that the opportunity regard to the study of political participation in for widespread and meaningful participation in Canada, existing inquiries have focused almost politicial processes is one of the crucial distin- exclusively on voting and campaign-related

Rev. canad. SOC. & Anth./Canad. Rev. SOC. & Anth. 15(1) 1978

62 / Mike Burke, Harold D. Clarke and Lawrence LeDuc

activities, but even in this area, research has been extremely limited, and many important substantive and methodological topics remain virtually unexplored. I Specifically, a review of the relevant literature (Scarrow, 1961, Laponce, 1967; Van Loon, 1970; Laskin and Baird, 1970; Sproule-Jones and Hart, 1973; Kornberg et al., 1973; Clarke et al., 1975; Curtis and Lambert, 1976) suggests the need for basic descriptive in- formation relating to the following questions: 1/ (a) Is it possible to construct valid scales of federal and provincial electoral participation, both for Canada as a whole, and for each of the several provinces?

(b) If such scales can be constructed, what do they tell one about levels of federal and pro- vincial political participation in Canada? 2/ To what extent d o patterns of federal and provincial political participation vary between and within provinces? 3/ (a) In the Canadian context, what is the total explanatory force of the conventionally used sociodemographic and attitudinal corre- lates of political participation?

(b) Which variables (if any) have particu- larly strong independent effects on political participation? 4/ Does it appear that regional variations in political culture play a role in explanations of political participation in Canada?’

To answer these questions, the present paper uses data gathered in the 1974 Canadian National Election Study. This study involved conducting structured interviews (averaging ninety minutes in length) with a national proba- bility sample of 2562 Canadians eligible to vote in the 1974 federal election.’ Although the pri-

mary purpose of the study was an investigation of attitudes and perceptions related to voting in federal and provincial elections, the interview schedule contained a much larger battery of political participation items than previous Canadian election surveys. In particular, the se- quence of items was designed to encompass a wider range of activities which might fall near the lower end of a participation scale - reading about politics, discussing politics, attending meetings, and so forth - or that might tap some non-electoral modes of participation such as working in the local community. Also, it was decided t o employ this larger sequence with reference to the 1974 election and to each level of the political system independently of specific elections. The study therefore contains four complete batteries of nine participation items - one for the 1974 election, another for the federal level in general, another for the provin- cial level, and a fourth for the localievel.4 These batteries provide a considerable number of alternatives in the construction of measures as well as in the investigation of modes and levels of participation.

MEASUREMENT

Table 1 presents the frequency distribution for thefederal set of items.’ The relative frequency with which individuals report engaging in various political acts is not unexpected. Thus, while fully 87.7 per cent of the respondents recall voting in all or most of the federal elec- tions in which they were eligible to cast a ballot, only 10.5 per cent report that they have spent time working for a federal political party or

1 For a discussion of the lack of studies of political participation in Canada, see Elkins and Blake (1975: 317). 2 Recently, scholars such as Schwartz, Simeon and Elkins, and Wilson have provided evidence documenting the reality of regional variations in Canadian political culture (Schwartz, 1974; Simeon and Elkins, 1974; Wilson, 1974). The work of Simeon and Elkins is particularly suggestive in that they found that regional variations in political efficacy, one of the most important explanatory variables in many studies of political participation, cannot be completely accounted for by regional differences in such variables as socioeconomic status (1974: 430-7). Thus, an exploratory examination of federal and provincial political participation such as the present study might usefully examine regional differences in the strength of various commonly used correlates of participation, and thereby attempt to assess regional variations in the over-all impact of unmeasured variables. 3 The questions concerning participation were asked of a random half sample of the total number of respondents. This half sample contains 1262 respondents which is weighted to an effective national sample size of 1203. For a more detailed description of the sample design, see LeDuc, Clarke, Jenson and Pammett (1974). 4 The set of items used to measure participation at each level is listed in Table 1. 5 That is, the set of items referring to participation in federal politics generally, not the 1974 election in particular.

Political participation in Canada / 63

TABLE I

FREQUENCY DISTRIBUTION OF FEDERAL PARTICIPATION ITEMS (row percentages)

(Positive (Negative response response

Often Sometimes total)’ Seldom Never total) (N)?

Vote frequency[ 59.6% 28.1% (87.7%) 7.6% 4.7% (12.3%) (1184) Read newspapers 40.7 28.7 (69.4) 17.7 12.9 (30.6) (1197) Discuss politics 23.7 37.2 (60.9) 22.0 17.1 (39.1) (1196) Convince friends 8.4 12.7 (21.1) 10.1 68.8 (78.9) (1195) Work in community 5.1 15.3 (20.4) 12.4 67.1 (79.5) (1192) Attend meetings 4.9 14.9 (19.8) 11.7 68.5 (80.2) (1195) Contact officials 3.0 11.1 (14.1) 13.8 72.0 (85.8) (1192) Sign, sticker 4.5 9.1 (13.6) 4.4 82.0 (86.4) (1190) Campaign activity 3.7 6.8 (10.5) 6.1 83.4 (89.5) (1192)

* Dichotomization point for scaling purposes t National half sample N’s reported in all tables are weighted and may vary because of missing data on certain items. The national half sample size is 1203 (weighted from 1262). Since the weights employed do not vary within provinces, in this and subsequent tables, provincial N ’ s reported are unweighted. 5 For the ‘vote frequency in federal elections’ question, response categories are ‘all,’ ‘most,’ ‘some,’ and ‘none.’ For scaling purposes, the dichotomization point is between the ‘most’ and ‘some’ categories. Response categories for all other items are as indicated.

candidate during a campaign. In between these extremes are clusters of respondents who read about politics in the newspapers, and discuss politics, followed by approximately 20 per cent of the electorate who attend political meetings, work with other people in the community to try to solve problems, or attempt to convince others how to vote.

The nature of the distribution is itself sug- gestive of the likelihood of forming Guttman scales from various subsets of items. Although a number of viable scales could be formed from subsets of four or five items, the strongest scales were obtained through the inclusion of a t least one composite item constructed from two or more activities at the same level. The strongest scale of the federal participation items (as mea- sured in terms of coefficients of reproducibility and scalability) is composed of the vote fre- quency, discuss politics, convince friends, at-

tend meetings, and campaign activity items, but is obtained by combining the convince friends and attend meetings items to form a single inter- mediate item in the scale. The use of this com- posite variable as a scale component yields a stronger scale than if either convince friends or attends meetings is used, individually, as scale items.6 As can be seen in Table 2 the modal cate- gory of federal political participation in Canada (as revealed by the scale described above) is voting plus ‘something else’ (in this case, dis- cussing politics).’ Only 11 per cent of the elec- torate remain totally inactive in politics, while an even smaller percentage (9.4 per cent) partici- pate in all types of activity.

Table 2 also demonstrates that the scale coefficients of the measure of federal partici- pation are well above the conventional lower bounds that define an acceptable scale. * More- over, this measure tends to scale in a similar

6 The distribution of this variable is as follows: positive response = 32.3 per cent; negative response = 67.7 per cent. If convince friends is substituted for the composite variable, the scale coefficients of the federal participation scale are CR = .93, cs = .70. If attend meetings is substituted for the composite variable, CR = .95, cs = .76. By using the composite variable as a scale component, the scale coefficients are increased to CR = .95, cs = .77. Non-perfect scale types were assigned scale scores on the basis of a probability distribution. 7 Verba and Nie also found that a large proportion of citizens do not restrict their political activity to voting. Almost 70 per cent of their sample reported engaging in at least one participatory act other than voting (1972: 33-4). 8 A CR of .90 and a cs of .60 are the lower bounds that define an acceptable scale. See Nie et al. (1975: 528-33). The participation scales were also assessed using the coefficient (Loevinger’s H) suggested by Mokken. H values were sufficiently high (for example, H = .89 for the federal participation scale) to indicate that the items formed strong scales. See Mokken (1971).

64 / Mike Burke, Harold D. Clarke and Lawrence LeDuc

manner for different subgroups of the Cana- dian p ~ p u l a t i o n . ~

Although some questions have been raised regarding the scalability of similar items in past studies (Van Loon, 1970: 378, n2). an accepta- ble replica of this particular scale could be con- structed from items in the 1965 election study as

well (see Table 3).1° Also, while alternative scales can be found, these particular items form a sufficiently robust scale for each of the four batteries in the 1974 study to enable cross-level comparisons." While no attempt is made here t o test the scalability of comparable items in other national studies, the nature of the items is

9 As can be seen from the following table, the federal participation measure is an adequate scale within each province:

~~~

Scale coefficients of the measure of federal participation by province

Newfoundland Prince Edward Island Nova Scotia New Brunswick Quebec Ontario Manitoba Saskatchewan Alberta British Columbia

CR

.88

.93

.96

.94

.95

.92

.91

.89

.89

.90

cs .59 .75 .85 .75 .81 .68 .59 .65 .62 .60

-

10 Van Loon (1970) was doubtful of the ability of the 1965 participation data to form a single scale. The reader, however, is cautioned against a direct comparison of the distributions of the 1965 and 1974 participation data. Initially, the two scales are composed of slightly different items. Secondly, the 1965 data measure participation within a specific election while the 1974 participation measure (data shown in Table 2) is a general one. Finally, even where the 1965 and 1974 items appear equivalent, differences in the wording of the actual questions render exact equivalence uncertain. 11 For the country as a whole, the CR and cs for the provincial participation scale equals .93 and .70, respectively (see Table 4). The provincial participation measure also forms an adequate scale within each province:

Scale coefficients of the measure of provincial participation by province

Newfoundland Prince Edward Island Nova Scotia New Brunswick Quebec Ontario Manitoba Saskatchewan Alberta British Columbia

CR

.89

.93

.95

.96

.95

.95

.96

.92

.91

.95

cs .59 .75 .82 .84 .78 .79 .84 .71 .69 .I3

-

Once again, the use of either convince friends or attend meetings as individuol scale components produces a weaker scale than if the composite variable is used. The scale with convince friends as a component has a CR = .92 and a cs = .65. The scale with attend meetings as a scale component has a CR = .93 and a cs = .68. Non-perfect scale types were assigned scale scores on the basis of a probability distribution.

Political participation in Canada / 65

TABLE 11

FREQUENCY DISTRIBUTION OF FEDERAL PARTICIPATION SCALE (column percentages)

Inactive (0). 11 .O% Vote (1) 29.0

Discuss politics (2) 29.8

Convince friends or attend meetings (3) 20.9

+ + +

Campaign activity (4) 9.4 N = 1203 Coefficient of reproducibility (CR) = .95

Coefficient of scalability (cs) = .77

* Number in parentheses refers to scale score.

TABLE 111

PARTICIPATION IN THE 1%5 FEDERAL ELECTION

Panel one: frequency distribution of participation items

(percentage positive response) (N)

Vote 85.3% (2566) Readabout politics 73.9 (265 7) Convince friends 22.8 (2676) Attend meetings 13.7 (2720) Campaign activity 5.1 (2 707)

suggestive of scalability in other political set- tings. Further inquiry will be required, how- ever, to establish the validity of the measure in other polities.12

FEDERAL A N D PROVINCIAL PARTICIPATION

Much of the past research in Canada has com- pared participation rates between the provinces only by means of inter-provincial comparisons of voting turnout. By itself, however, voting turnout provides neither an adequate concep- tual nor a sound operational basis on which to infer rates of political participation. The Gutt- man scales used in this analysis, while made up of only a subset of those acts included under the rubric ‘electoral participation’, are a more comprehensive measure of provincial political activity than voting turnout alone.

As can be seen from the data in Table 4, the

Panel two: frequency distribution of 1965 federal participation scale (column percentages)

Inactive (O)* 6.1% Vote (1) 22.4

Read about politics (2) 44.1

Convince friends or attend meetings (3) 23.2

Campaign activity (4) 4.2

+ + + N = 2467 CR = .94

cs = .70

* Number in parentheses refers to scale score.

distributions of the provincial participation items are similar to the distributions of the federal participation items discussed above, and the scale is of comparable strength to its federal counterpart. A large proportion of the elector- ate vote in provincial elections while a much smaller percentage report working for a politi- cal party or candidate during a provincial campaign. Once again, about one-fifth of the sample report that they attend political meetings and/or try to convince others how to vote. Furthermore, there is a striking similarity in the distributions of the federal and provincial parti- cipation scales. The same proportion of indi- viduals participate in the same acts at both levels of government. The most salient distinc- tion is in the greater number of persons inactive in provincial politics as compared to the propor- tion inactive federally; but even here the differ- ence is less than 6 percentage points.”

12 Virtually identical items to those employed here have been used in a recent eleven-country study of political culture and political participation coordinated by Professor Samuel Barnes of the University of Michigan. When these data are made available, it will be possible to test the cross-national strengths of the scale employed in this paper. 13 Eleven per cent and 16.8 per cent are inactive in federal and provincial politics, respectively.

66 / Mike Burke, Harold D. Clarke and Lawrence LeDuc

TABLE IV

PROVINCIAL PARTICIPATION

Panel one: frequency distribution of provincial participation items (row percentages)

(Positive (Negative response response

Often Sometimes total)’ Seldom Never total) (N) ~ ~~

Vote frequency** 52.8% 27.6% (80.4%- 1 10.1% 9.5070 (19.6%) (1188)

Read newspapers 41.6 29.0 (70.6) 16.4 13.0 (29.4) (1198) Discuss politics 26.0 36.5 (62.5) 20.1 17.4 (37.5) (1197) Convince friends 9.1 13.5 (22.6) 9.4 68.0 (77.4) (1192) Work in community 5.7 15.2 (20.9) 11.1 67.9 (79.0) (1192) Attend meetings 5.3 14.6 (19.9) 11.4 68.7 (80.1) (1193) Contact officials 3.1 11.8 (14.9) 11.4 73.8 (85.2) (1192) Sign, sticker 4.9 9.4 (14.3) 4.0 81.8 (85.8) (1187) Campaign activity 3.6 7 .O (10.6) 5.5 83.9 (89.4) (1194)

* Dichotomization point for scaling purposes ** For the ‘vote frequency in provincial elections’ question, response categories are ‘all,’ ‘most,’ ‘some,’ and ‘none.’ For scaling purposes, the dichotomization point is between the ‘most’ and ‘some’ categories. Response categories for all other items are as indicated.

Panel two: frequency distribution of provincial participation scale (column percentages)

(N)

Inactive (0)’ 16.8% (202)

Vote (1) 25.2 (303) + + + N = 1203 CR = .93

cs = .70

Discuss politics (2) 26.7 (321)

Campaign activity (4) 9.5 (114)

Convince friends or attend meetings (3) 21.9 (263)

* Number in parentheses refers to scale score.

Table 5 presents the mean distributions of the federal and provincial participation scales by province. The table shows participation in federal politics to be highest in Prince Edward Island (X = 2.10) and lowest in Alberta (X = 1.74).14 These difference, however, are not statistically significant. Therefore, the hypothe- sis that the provincial mean values (for federal participation) are in fact equal cannot be rejected. Mean values of the provincial partici- pation scale are highest in Nova Scotia (X = 2.18), Prince Edward Island (X = 2.00), and Quebec (X = 1.99) but relatively lower in Ontario (X = 1.67) and Alberta (X = 1.61). These mean differences in the rate of provincial activity are statistically significant. In sum-

mary, the mean rates of provincial, though not federal, participation, as measured by the Gutt- man scale, vary significantly by province.

The conclusion that interprovincial rates of provincial participation vary more than federal participation is buttressed by examination of cross tabulations of the two participation scales by province. Chi-square values are significant in both instances, but are somewhat greater in the case of the provincial scale. Inspection of pro- vincial scale percentage differences across prov- inces illustrates this variance. Alberta (mean = 1.61), for example, has the highest proportion of non-participants, while Manitoba, also scor- ing low on the summary measure (mean =

1.73), has a substantial proportion of

14 Other authors have discovered that federal participation rates are high in Prince Edward Island and low in Alberta. See, for example, Van Loon (1970: 388); Scarrow (1961: 355).

Political participation in Canada / 67

TABLE V

MEAN DISTRIBUTION OF FEDERAL AND PROVINCIAL PARTICIPATION BY PROVINCE

Federal x Provincial x T Value (N)

Canada 1.89 1.82 2.64* (1203) Newfoundland 1.80 1.76 0.50 (49)

Nova Scotia 2.03 2.18 -2.07. (87) New Brunswick 1.91 1.82 0.80 (66) Quebec 1.92 1.99 -1.54 (341) Ontario 1.90 1.67 5.04’ (344) Manitoba 1.85 1.73 0.98 (59) Saskatchewan 1.90 1.85 0.55 (52) Alberta 1.74 1.61 1.37 (87) British Columbia 1.83 1.92 -1.10 (127)

Prince Edward Island 2.10 2.00 1 .OO (50)

F value .62 2.68’

* pG .05

respondents at the ‘discuss politics’ point on the scale, but few at higher levels. Again, Prince Edward Island has a high provincial mean score (2.00) because it has few non-participants and the largest proportion in the campaign activity category. British Columbia, on the other hand, (mean = 1.92) has a relatively large proportion of respondents in the ‘convince friendslattend meetings’ category but fewer engaging in actual campaign activities. It thus falls near the middle of the distribution of means. More generally, the difference between provinces in rates of pro- vincial and (to a lesser extent) federal participa- tion are based on a number of such important differences in the style and substance of partici- pation in the several provinces.

Additionally, it is possible to compare intra- provincial rates of federal and provincial activ- ity. The T values presented in Table 5 above are measures of the significance of the difference between the means of the federal and provincial participation scales. Within each province, save Nova Scotia and Ontario, the mean rates of federal and provincial participation are remark- ably alike. In Ontario, the mean rate of federal participation is significantly higher than the mean rate of provincial participation; in Nova Scotia this pattern is reversed. Perhaps surpris- ingly, in Quebec, the rate of provincial activity is not significantly higher than the federal parti-

cipation rate.” Over-all, these data indicate that the populations of most provinces are not in- clined to participate a t significantly greater rates a t one level of government.

Comparing mean rates of federal and provin- cial activity may, to some extent, obfuscate individual differences in the nature of partici- pation. Therefore, in order to investigate individual variations in the rates of political activity, the federal and provincial participation scales are cross tabulated in Table 8. It may be seen that there is a strong relationship between the federal and provincial participation scales in the country as a whole. Indeed, the bulk of the sample is contained along the main diagonal. Over four-fifths of the respondents who vote and discuss politics provincially also vote and discuss politics federally. Fully three-quarters of those who engage in all types of activity in provincial elections also participate in those same acts at the federal level. There is a some- what poorer fit between the levels for inactive respondents. For example, 47 per cent of those inactive in provincial politics particpate in some way in federal politics. As might be expected, however, the majority of these respondents are active at the lower levels (vote, discuss politics) of federal participation. Table 9 demonstrates that this strong association between federal and provincial activity obtains within each province.

15 Scarrow found a similar pattern in Ontario and Nova Scotia but, in Quebec, his analysis revealed that turnout in provincial elections is consistently higher than the turnout in the preceding federal election (1961 : 362). Van Loon’s analysis suggests higher turnout rates in federal elections (1970: 388).

TABL

E V

I

FED

ER

AL

PA

RT

ICIP

AT

ION

BY P

RO

VIN

CE

(col

umn

perc

enta

ges)

Federal

part

icip

atio

n Nfld.

PEI

NS

NB

Q

ue.

Ont

. M

an.

Sask.

Alto

. B

C

Inac

tive

10.2%

6.0%

12.6%

4.5%

9.7%

10.8%

10.2%

15.4%

19.5%

9.4%

Vot

e 38.8

34.0

27.6

42.4

29.9

27.9

27.1

26.9

27.6

27.6

Discuss

polit

ics

20.4

24.0

20.7

22.7

28.7

31.4

39.0

26.9

19.5

37.8

Con

vinc

e fri

ends

or

22.4

16.0

21.8

18.2

22.6

20.1

15.3

13.5

26.4

21.3

Att

end

mee

tings

C

ampa

ign a

ctiv

ity

8.2

20.0

17.2

12.1

9.1

9.9

8.5

17.3

6.9

3.9

(49)

(50)

(87)

(66)

(341)

(344)

(59)

(52)

(87)

(127)

@ .02

v= .ll

TA

BL

E V

II

PRO

VIN

CIA

L P

AR

TIC

IPA

TT

ON

BY P

RO

VIN

CE

(col

umn

perc

enta

ges)

Prov

inci

ol

part

icip

atio

n

Inac

tive

Vot

e D

iscu

ss p

oliti

cs

Con

vinc

e fri

ends

Att

end meetings

Cam

paig

n act

ivity

or

Nfld.

18.4%

28.6

20.4

24.5

8.2

(49)

PEI

6.0%

38.0

24.0

14.0

18.0

(50)

NS

8.0%

26.4

21.8

26.4

17.2

(87)

NB

12.1%

33.3

24.2

21.2

9.1

(66)

Que

.

10.6%

28.2

24.9

Ont

. M

an.

Sask.

20.9('10

15.3%

19.2%

24.1

27.1

26.9

29.1

33.9

21.2

18.6

16.9

15.4

7.3

6.8

17.3

(344)

(59)

(52)

Aka

.

26.4Vo

20.7

25.3

20.7

6.9

(87)

BC

17.3%

18.1

26.8

30.7

7.1

(127)

p;c .ool

v= .12

Political participation in Canada / 69

TABLE VIII

FEDERAL PARTICIPATION BY PROVINCIAL PARTICIPATION (column percentages)

Provincial participation

Convince frienh Federal Discuss or Campaign participation Inactive Vote politics attend meetings activity

Inactive 52.3% 4.4% 0.7% 2.7% 3.5% Vote 22.1 78.1 11.4 10.3 3.5 Discuss politics 18.5 9.0 80.8 11.0 4.4 Convince friends

or 5.7 6.6 5.6 71.5 11.9 Attend meetings Campaign activity 1.5 1.9 1.6 4.6 76.7

(N = 1203) (202) (303) (321) (263) (114) pG.001 V = .67

TABLE IX

FEDERAL PARTICIPATION BY PROVINCIAL PARTICIPATION BY PROVINCE

Zero-Order r (N)

Canada Newfoundland Prince Edward Island Nova Scotia New Brunswick Quebec Ontario Manitoba Saskatchewan Alberta British Columbia

.72*

.89

.84

.86

.68

.69

.73

.65

.84

.77

.63

* All coefficients significant at .001 level (F test)

In all areas of the country, individuals tend to ical participation. Van Loon, for instance, participate a t the same rates in both federal and found that individuals with high levels of provincial elections. I 6 political efficacy and political interest are more

active than less efficacious and less interested CORRELATES OF PARTICIPATION persons, and that socioeconomic status was

positively related to participation. Similar As noted earlier, numerous studies have sug- correlations have been found in other studies gested the relevance of attitudinal and social (Van Loon, 1970: 397; Sproule-Jones and Hart , background variables as determinants of polit- 1973: 190; Welch, 1975: 558). In this analysis, a

16 In some provinces the correlation between federal and provincial participation is below the national average. The data in British Columbia, Manitoba, and New Brunswick correspond to the national pattern in that most of the slippage occurs at the lower level of participation. That is, in these provinces there is a comparatively large proportion of respondents who are inactive in provincial politics but who vote andlor discuss politics federally. On the other hand, in Quebec, most of the instability is in terms of the upper level of participation. A large percentage of Quebec respondents who engage in all types of activity in provincial politics participate at lower levels federally. Although available data do not permit one to test the hypothesis, it may be suggested that the patterning of political participation in Quebec detected in these data might be accounted for, in part at least, by the growth of separatism among certain segments of the Quebec population.

70 / Mike Burke, Harold D. Clarke and Lawrence LeDuc

number of attitudinal and sociodemographic variables are employed as predictors of political activity in a multiple regression analysis.”

Before entering the regression equation, the independent variables were standardized, through a series of Z transformations, to a mean of zero and a standard deviation of one. This strategy, by controlling for the contami- nating influence of disparate units of measure- ment, allows the comparison of beta coefficents both within and across subgroups (Irvine, 1971, 1972; Chi, 1972). The dependent variables are not expressed in standard units as this would result in the elimination of the intercept from the regression equation. The intercept will prove useful in that it can be employed as an indicator of ‘unmeasured effects’ on the dependent variable.

Across the country as a whole, attitudinal variables are the strongest correlates of federal and provincial participation. The beta coeffi- cients arrayed in Table 10 demonstrate that political interest, perception of a difference be- tween parties, strength of party identification, and political efficacy are significantly and posi- tively related to participation.I8 In addition, it appears that French Canadians are more likely to be active in politics than either English or ‘other’ Canadians.I9

Contrariwise, variables such as age, sex, and socioeconomic status have insignificant inde- pendent effects on the extent of electoral parti- cipation at both federal and provincial levels.’O

Regarding the over-all explanatory power of the several predictor variables, the multiple R’s are .43 for the analyses of both federal and pro- vincial participation. These correlations are quite similar in magnitude t o those reported by Nie and Verba in studies of participation in the United States and elsewhere (Nie and Verba:

19). Assuming measurement error is not a seri- ous problem, the modest size of these multiple correlations indicates the inadequacy of explan- ations of political participation in Canada and other political systems cast strictly in terms of conventional sociodemographic and attitudinal variables.

It is possible that the effects of certain variables on political participation are not uniform in all areas of the country. Indeed, other studies have demonstrated that there are in fact marked regional differences in the rela- tionships among variables related t o mass political behaviour. Blake, for instance, in a suggestive analysis of regional differences in federal voting patterns, found that the associa- tion of religion and ethnicity with partisan preference varied significantly by region (Blake, 1972). Also, as noted previously, Simeon and Elkins have discovered that the relationship be- tween attitudinal variables such as political effi- cacy and social background characteristics varied from province to province (Simeon and Elkins, 1974: 397-437). Therefore, to explore the determinants of political participation in Canada in more detail, the relationships be- tween electoral participation and its correlates was examined within each region.

The beta coefficients in Tables 11 and 12 reveal that the significant and relatively strong relationship between political interest and parti- cipation reported above for the country as a whole obtains within each region. Interestingly enough, however, there is considerable regional variation in the impact of the remaining varia- bles. Strength of party identification is a better predictor of federal activity in British Columbia and the Atlantic region than in Quebec, Ontario, or the Prairies. The relationship be- tween political efficacy and provincial partici-

17 In the regression analyses political interest, perceptions of differences between parties, strength of party identification and ethnicity are entered into the regression equations as ‘dummy’ variables. On the use of dummy variables in regression analysis see Hilton (1976: 187-92). Concerning the substantive interpretation of regression analyses using dummy variables see Blake (1972: 60). 18 An individual’s degree of political interest, perception of differences between parties, and strength of party identification were ascertained using responses to single questions. Political efficacy, on the other hand, was measured using a four-item scale. Details of the construction of this scale can be found in LeDuc (1976). 19 In the regression analyses ‘English-Canadian’ is the suppressed category of the dummy ethnicity variable. The betas in Tables 10, 11, and 12 measure the effects of having ‘French-’ or ‘Other’ as opposed to ‘English- Canadian’ ethnicity. 20 Sociodemographic characteristics were measured as follows: (a) age, age in years; (b) educution, number of years of formal education; (c) sex, a dichotomous variable with men scored ‘1,’ women ‘0’; (d) ethnicity, a trichotomous dummy variable with the categories English, French, and Other; (e) socioeconomic stutus, the revised Blishen scale of occupational prestige. For details concerning the Blishen scores, see Blishen and McRoberts (1976).

Political participation in Canada / 71

TABLE X

BETA COEFFICIENTS (AND ZERO-ORDER R’S) FOR PREDICTORS OF FEDERAL AND PROVINCIAL PARTICIPATION, CANADA

Federal Provincial

b r b r

Political interest (moderate) Political interest (high) Perception of difference

between parties (moderate) Perception of difference

between parties (high) Strength of party identification

(low) Strength of party identification

(moderate) Strength of party identification

Political efficacy Socioeconomic status Education Sex Age Ethnicity (French) Ethnicity (other)

Multiple R = Regression intercept = (NC = 1

(high)

.26*

.36*

.01

.14*

-.06

.04

.11*

.08*

.04

.05 -.01 -.01

.12* -.07* .43

1.91 (1008)

(.12*) (.26*)

(-.ll*)

(.20*)

(-.13*)

(-.Ol)

(.17*) (.Is*) (.12*) (.09*)

(.03) (.03)

(-.05*)

.21*

.33*

-.01

.13*

-.16*

-t .12* .08* .03 .06* .04

-t .18*

-.03 .43

1.89 (945)

(-07.) (.24*)

(-.12*)

(.20*)

(-.21*)

(-.03)

(.20*) (.12*) (.09*) (.06*) (.05*) (.W (.09*)

(-.Ol)

* Coefficient significant at or beyond .05 level (F test) t Coefficient less than .01 5 Listwise deletion of missing data

pation is statistically significant only among Quebec respondents. Perception of a difference between parties has a much stronger influence on federal and provincial participation in the Maritimes and the Prairies than elsewhere. Al- though sociodemographic variables tend to be relatively weak predictors of political activity, their impact on participation also varies by region. In the Maritimes, for example, older persons are more likely to be active in federal and provincial politics than younger persons. In contrast, in British Columbia, younger people participate a t higher rates than the older age groups. Over-all, while the ‘total’ within-region relationships (as indicated by the R’s) between the several predictor variables and federal and provincial participation are only marginally greater than those for the country as a whole, it appears that the national results presented in Table 10 above mask, to some extent, substan- tial regional differences in the relationships among variables. Although it is impossible to assess adequately the implications of these find- ings within the context of a brief paper, the data

are suggestive of the existence of significant regional variations in the impact of sociodemo- graphic and attitudinal variables on political participation.

Since the intercept is a measure of the depen- dent variable when all the independent variables are zero, it may be interpreted as the ‘starting point’ of participation within a given region. In essence, the intercept represents ‘unmeasured effects’ on the dependent variable (Irvine, 1971). As Tables 11 and 12 indicate, the value of the intercept varies from one region to the next. The base level of federal and provincial participation is highest in the Maritimes and lowest in Ontario. Substantively, variance in the intercepts is significant in that it provides an additional indicator of the potential signifi- cance of regional effects.

SUMMARY AND CONCLUSION

An important aim of this paper has been the development and testing of Guttman scale mea- sures of federal and provincial participation in

TAB

LE X

I

BET

A C

OEF

FIC

IEN

TS (A

ND

ZER

OO

RD

ER R

'S)

FOR

PR

EDIC

TOR

S OF

FED

ERA

L PA

RTI

CIP

ATI

ON

BY

REG

ION

~

__

_~

~

~ ~

Atla

ntic

Q

uebe

c O

ntar

io

Prai

ries

B

ritis

h C

olum

bia

b r

b r

b r

b r

b 1

Polit

ical

inte

rest

(mod

erat

e)

.21*

(.1

5*)

.32*

(.2

1*)

.19*

(.0

2)

.34*

(.2

1*)

.25*

(.09)

Polit

ical

inte

rest

(high)

.38*

(.2

3*)

.a*

(24

')

.36*

(.3

4*)

.33*

(.2

1*)

.24*

(.l

5*)

Perc

eptio

n of

diff

eren

ces

.17*

(-.

07)

-.lo

(-.

02)

betw

een

part

ies (

mod

erat

e)

.25*

(-.M)

-.05

(-.14

*)

-.M

(-.14

*)

Perc

eptio

n of

diff

eren

ces

.37*

(.2

8*)

-.14

(--0

3)

betw

een

part

ies (

high

) .3

8*

(.26*

) .1

1*

(.19*

) .1

0 (.2

3*)

Stre

ngth

of p

arty

iden

tific

atio

n -.0

1 (-.

20*)

.1

5 (lo

w)

-.09

(-.23

*)

-.lo

(-

.ll*)

-.1

2*

(-.ll

*)

Stre

ngth

of p

arty

iden

tific

atio

n .1

4 (-

12)

.14

(-.09)

(mod

erat

e)

.35*

(.1

3*)

-.04

(-01

) -.04

(-.lo

*)

(hig

h)

.ll

(.08)

-.0

3 (.l

o*)

.ll

(.21*

) St

reng

th o

f pa

rty

iden

tific

atio

n .1

7 (.1

7*)

.a*

(22

9

Polit

ical

eff

icac

y -.0

7 (.W

.14*

(.1

9*)

.lo*

(.1

9*)

.IS*

(.17*

) -.0

1 (.0

8)

Soci

oeco

nom

ic st

atus

-.0

5 (.MI

-.01

(.12*

) .M

(.16*

) --I

(.M)

.22*

(.2

1*)

Edu

catio

n .3

1*

(.11)

.0

1 (.1

2*)

.09

(.lo

*)

.04

(-10

) -.2

3 (-

08)

sex

.09

(.08)

.0

1 (.0

1)

-.05

(.01)

.04

(.01)

-.0

9 (-.Ol)

Age

.1

5*

(.14*

) -.M

(-.03

) .0

2 (.o

n .0

8 (.0

7)

-.19*

(-.09)

Ethn

icity

(Fre

nch)

.1

8*

(.02)

.0

7 (.0

2)

-.03

(-.03

) -t

(-1 -t

(-)

Ethn

icity

(oth

er)

-t (-1

-t (-1

-.0

3 (-.

02)

-.11*

(-

.lo)

-.09

(-.0

8)

Mul

tiple

R =

.5

5 .4

5 .4

8 .5

2 .44

Reg

ress

ion

inte

rcep

t =

2.

33

1.98

1.

80

1 .%

2.

09

(N.* =)

(173

) (2

81)

(290

) (1

65)

(113

)

Coe

ffic

ient

sign

ifica

nt a

t or

beyo

nd .0

5 le

vel (F

test

) t Too fe

w cases fo

r an

alys

is

4 C

oeff

icie

nt le

ss th

an .0

1 **

Lis

twis

e de

letio

n of

mis

sing

dat

a

TAB

LE X

I1

BET

A C

OEF

FIC

IEN

TS (A

ND

ZER

O-O

RD

ER R

'S)

FOR

PR

EDIC

TOR

S O

F PR

OV

INC

IAL

PAR

TlC

lPA

TlO

N BY

REG

ION

Atla

ntic

Q

uebe

c O

ntar

io

Prai

ries

B

ritis

h C

olum

bia

b r

b r

b r

b r

b r

Polit

ical

inte

rest

(mod

erat

e)

Polit

ical

inte

rest

(hig

h)

Perc

eptio

n of

diff

eren

ces

betw

een

part

ies (

mod

erat

e)

Perc

eptio

n of

diff

eren

ces

betw

een

part

ies

(hig

h)

Stre

ngth

of

part

y id

entif

icat

ion

(low

) St

reng

th o

f pa

rty

iden

tific

atio

n (m

oder

ate)

St

reng

th of

par

ty id

entif

icat

ion

(hig

h)

Polit

ical

eff

icac

y So

cioe

cono

mic

stat

us

Edu

catio

n Se

x A

ge

Ethn

icity

(Fre

nch)

Et

hnic

ity (o

ther

) M

ultip

le R =

R

egre

ssio

n in

terc

ept

=

(P*

=)

.16*

.2

7*

.10

.32*

.10

.39

.34

-.03 .09

.ll

.24*

.1

9*

.ll

-t .51

2.28

( 1

70)

(.lo)

(.2

7*)

(-.12

)

(.30*

)

(-.20

*)

~0

3)

(.12*

) (.0

7)

(.W

(-.O

l) (.1

8*)

(.19*

) (.W

(-)

.23*

.3

3*

-.06

-6 -.22*

-t .04

.20*

.08

-.01 .19*

-.0

7 .17*

-t .4

8 1.

98

(273

)

(.IS*

) (.2

0*)

(-.ll

*)

(.IS*

)

(-.26

*)

(.W

(.13*

) (.2

3*)

(.16*

) (.1

2*)

(.15*

) (-

.08)

(.0

8)

(-1

.15*

.3

8*

-.01 .06

-.11 .01

.26*

.0

3 .0

2 .1

4*

-.01 .02

.04

-.01 .49

1.73

(2

60)

(-.0

5)

(.37*

)

(-.ll

*)

(.20*

)

(-.21

*)

(-.12

*)

(.30*

) (.1

4*)

(.12*

) (.W

(.02)

(.

ll*)

(.0

2)

(.02)

.28*

.3

8*

.14

.28*

-.08 .ll

.13

.02

--t .21

-.07 .06

-t .05

.44

1.97

(1

52)

(.14*

) (.2

4*)

(-.07

)

(.19*

)

(-.IS*)

(-.03

)

(.17*

)

(.W

(.W

(.07)

(-

.MI

(.W

(-1

(.MI

.20*

.2

5*

-.16 .31

-.33*

-.36*

-.20 .04

.06

-.22

-.13

-.21*

-t -.I

S*

.45

1.76

(1

10)

(.lo

)

(-.28

*)

~0

7)

(.32*

)

(-. 1

8*)

(-.@

I

(.la

*)

(.05

) (-.

Ol)

(.W

(-.

02)

(-.0

5)

(-1 (-

.08)

Coe

ffic

ient

sig

nific

ant a

t or

beyo

nd .0

5 le

vel (

F te

st)

t Too fe

w c

ases

for

ana

lysi

s 6 C

oeff

icie

nt le

ss th

an .0

1 **

Lis

twis

e de

letio

n of

mis

sing

dat

a

74 / Mike Burke, Harold D. Clarke and Lawrence LeDuc

Canada. Aside from this methodological focus, however, a number of substantive conclusions may be drawn from the foregoing analysis and discussion.

The scales constructed in this paper do not in- crease substantially the proportion of variance in participation rates which may be explained by conventional sociodemographic and attitudi- nal variables. In Canada, as well as elsewhere, improved explanations of patterns of participa- tion will likely be realized only by the utilization of alternative variables, not be refinements in the measurement of participation. Improve- ments in participation measures may, however, lead to some rethinking of our conventional understanding of Canada as a political culture characterized by very low rates of participation. The concept of Canada as a nation of political ‘spectators’ (Van Loon, 1970: 396-9) derives partly from the small proportion of the popula- tion found to engage in campaign activities coupled with a relatively high rate of voting participation. Successful measurement of some dimensions of intermediate activity, however, suggests a revision of this conventional under- standing. Employing the summary measure of federal participation discussed in this paper, no fewer than 60 per cent of the 1974 national sam- ple were found to have engaged in some form of participatory act beyond simply voting z 1 (see Table 2). This finding results from the relatively large number of persons who are found to have engaged in one or more intermediate types of participatory acts, that is, ‘discuss politics’, ‘convince friends’, ‘attend meetings’, and so forth. While such acts clearly are not the equi- valent of direct participation in politics, they represent more than a passive activity such as voting, and it is important to take note of these forms of participation in measures which may be employed in polities such as Canada.

In the foregoing analysis, a n attempt was made to distinguish clearly between federal and provincial participation by developing separate but comparable measures of participation at each level. The strong relationship between federal and provincial participation reported in this paper is a predictable but heretofore un- documented one.2z Also, analysis reveals that the range of variation among the provinces in

levels of federal participation is not great. However, the relatively greater variance in pro- vincial participation is a phenomenon worthy of further investigation within the context of the politics of particular provinces. The lower rates of provincial participation reported for Alberta and Ontario, for example, may be a function of the particular patterns of inter-party competi- tion in these provinces, a t least up through the time of the study. 2 1 The extensive variation in patterns of provincial politics makes this a topic of importance but also one beyond the scope of the present investigation. Similarly, the differ- ences in correlates of both federal and provin- cial participation are suggestive of arguments which have been advanced regarding the exis- tence of distinct regional and/or provincial political cultures in Canada (for example, Schwartz, 1974; Simeon and Elkins, 1974; Wilson, 1974). We have discovered, for exam- ple, that political efficacy is significantly related to federal activity in Quebec, Ontario, and the Prairies, but not in the Maritimes or in British Columbia. The impact of a number of other variables on levels of both provincial and federal participation was likewise found to vary significantly from region to region. Because of the importance of a variable such as efficacy to our understanding of political culture, these findings are also suggestive of directions for research in this area at the provincial or regional level. The present research clearly indicates that any investigation of political participation undertaken from a purely national perspective in Canada risks ignoring potentially significant provincial and/or regional differences in both the rates and correlates of particular aspects of political activity.

REFERENCES

Blake, Donald E. 1972 ‘The measurement of regionalism in

Canadian voting patterns.’ Canadian Journal of Political Science 555-81

Blishen, Bernard R., and Hugh A. McRoberts 1976 ‘A revised socioeconomic index for occu-

pations in Canada.’ Canadian Review of Sociology and Anthropology 12: 7 1-9

Chi, N.H. 21 Similarly, Verba and Nie (1972: 25-43) conclude that political participation is more widespread in the United States than had been previously suspected. 22 On the tendency for many party workers in Canada to be active in both federal and provincial election campaigns, see Jacek (1972: 194). Similar findings have been recently reported by Clarke et al. (1978). 23 On levels of inter-party competition in provincial politics, see LeDuc and White (1974: 81-9).

Political participation in Canada / 75

1972 ‘The regressional model of regionalism: a critique.’ Canadian Journal of Political Science 5 : 291-7

Clarke, Harold D., Allan Kornberg, and James Lee

1975 ‘Ontario student party activists: a note on differential participation in a voluntary organization.’ Canadian Review of Sociology and Anthropology 12: 21 3-20

Clarke, Harold D., Richard G. Price, Marianne Stewart, and Robert Krause

1978 ‘Motivational patterns and differential participation in a Canadian party: the Ontario Liberals.’ American Journal of Political Science 22: 130-5 1

Curtis, James E., and Ronald D. Lambert 1976 ‘Voting, election interest, and age: na-

tional findings of English and French Canadians.’ Canadian Journal of Political Science 9: 293-307

Elkins, David J., and Donald E. Blake 1975 ‘Voting research in Canada: problems

and prospects.’ Canadian Journal of Political Science 8: 313-25

Hilton, Gordon 1976 Intermediate Politometrics. New York:

Columbia University Press Irvine, William P. 1971 ‘Assessing regional effects in data

analysis.’ Canadian Journal of Political Science 4: 21-4

Canadian Journal of Political Science 5 : 1972 ‘A reply to Professor Chi’s critique.’

297-300 Jacek, Henry, John McDonough, Ronald

1972 ‘The congruence of federal-provincial Shimizu, and Patrick Smith

campaign activity in party organizations.’ Canadian Journal of Political Science 5 : 190-205

Kornberg, Allan, Joel Smith, Mary-Jane Clarke, and Harold D. Clarke

1973 ‘Participation in local party organizations in the United States and Canada.’ American Journal of Political Science 7: 23-47

Laponce, J.A. 1967 “on-voting and non-voters: a typology.’

Canadian Journal of Economics and Political Science 33: 75-87

Laskin, Richard, and Richard Baird 1970 ‘Factors in voter turnout and party pre-

ference in a Saskatchewan town.’ Cana- dian Journal of Political Science 3: 450-62

LeDuc, Lawrence, and Walter L. White

1974 ‘The role of opposition in a one-party dominant system: the case of Ontario.’ Canadian Journal of Political Science 7: 86- 100

LeDuc, Lawrence, Harold D. Clarke, Jane Jenson, and Jon Pammett

1974 ‘A national sample design.’ Canadian Journal of Political Science 4: 701-8

LeDuc, Lawrence 1976 ‘Measuring the sense of political efficacy

in Canada.’ Comparative Political Studies 8: 490-500

Milbrath, Lester 1965 Political Participation. Chicago: Rand

McNally Mokken, Robert 1971 A Theory and Procedure of Scale

Analysis. The Hague: Mouton Nie, Norman, and Sidney Verba 1975 ‘Political participation.’ In F.I. Green-

stein and N.W. Polsby (eds.), Handbook of Political Science, v.4. Readings, Mass.: Addison-Wesley

Nie, Norman, et al. 1975 Statistical Package for the Social

Sciences, 2nd ed. Toronto: McGraw-Hill Scarrow, Howard A. 1961 ‘Patterns of voter turnout in Canada.’

Midwest Journal of Political Science 5 : 351-64

Schwartz, Mildred 1974 Politics and Territory. Montreal: McGill-

Queen’s University Press Simeon, Richard, and David J. Elkins 1974 ‘Regional political cultures in Canada.’

Canadian Journal of Political Science 7: 397-437

Sproule-Jones, Mark, and Kenneth D. Hart 1973 ‘A public-choice model of political parti-

cipation.’ Canadian Journal of Political Science 6: 175-94

Van Loon, Rick 1970 ‘Political participation in Canada: the

1965 election.’ Canadian Journal of Political Science 3: 376-99

Verba, Sidney, and Norman Nie 1972 Participation in America. New York:

Welch, Susan 1975 ‘Dimensions of political participation in a

Harper and Row

Canadian sample.’ Canadian Journal of Political Science 8: 553-9

Wilson, John 1974 ‘The Canadian political cultures.’ Cana-

dian Journal of Political Science 7: 438-83


Recommended