UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interest group involvement in constituency election campaigns Sovka, Roseanne M. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_fall_sovka_roseanne.pdf [ 4.45MB ]
JSON: 831-1.0086304.json
JSON-LD: 831-1.0086304-ld.json
RDF/XML (Pretty): 831-1.0086304-rdf.xml
RDF/JSON: 831-1.0086304-rdf.json
Turtle: 831-1.0086304-turtle.txt
N-Triples: 831-1.0086304-rdf-ntriples.txt
Original Record: 831-1.0086304-source.json
Full Text

Full Text

INTEREST GROUP INVOLVEMENT INCONSTITUENCY ELECTION CAMPAIGNSbyROSEANNE MARGARET SOVKAB.A.(Hon), The University of Calgary, 1989A THESIS IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(Department of Political Science)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJuly 1993©Roseanne Margaret Sovka, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature) Department of  Political ScienceThe University of British ColumbiaVancouver, CanadaDate July 27, 1993DE-6 (2/88)AbstractThis study explores the range and variance of interest group activity in constituencycampaigns in the 1988 federal election as reported in the Constituency Party Association dataset created in 1991 for the Royal Commission on Electoral Reform and Party Financing.SPSSPC+ was used to analyze the relationships between variables in four main areas:political party affiliation, geographic variables, constituency association characteristics, andthe specific issues the interest groups were promoting or opposing.The most significant finding was that interest groups were actively involved in half ofthe riding association election campaigns, either supporting or opposing local candidates. Thecursory treatment of electoral involvement in the interest group literature provides aninadequate explanation for this widespread phenomena. This study provides an initial profileof interest group involvement in constituency campaigns. The exploration of the data revealedthat interest groups were more likely to be involved in the local campaigns of candidatesassociated with the governing party. They were less likely to be involved in Quebecconstituency campaigns, and more likely in wealthy competitive riding campaigns. The mostfrequently mentioned issues that motivated interest groups locally were abortion, followed byfree trade.iiTABLE OF CONTENTSAbstractTable of ContentsList of TablesList of FiguresAcknowledgementChapter One Possible Explanations for Interest Group ElectionCampaign InvolvementWhen is Election Involvement Efficient?Searching for LegitimacyCitizen ParticipationNational Forum for Policy DebateInternal Benefits for Interest Groupsvii13891012Chapter Two^The Constituency Party Association Study^14Questions Pertaining to Interest Groups 17Chapter Three^Calling the Government to Account^ 21Chapter Four^Geographic Variation^ 31Province and Region 32Urbanization^ 38Chapter FiveChapter SixConstituency Association CharacteristicsLocal CompetitivenessCampaign TypeFinancial IndicatorsAssociation MembershipMotivating IssuesFree TradeAbortion"Other" issues414146525964677684Chapter Seven^Conclusion^ 87iiiBibliography^ 92Appendix 1^Selected Questions from the Constituency PartyAssociation Survey^ 96Appendix 2^Derivation of the Variable "Margin2"^100ivList of TablesTable 4.1Interest Group Electoral Involvement by Province and Region^34Table 5.1Constituency Competitiveness and Interest Group Activity^ 44Table 5.2Campaign Control and Interest Group Activity^ 49Table 5.3Association Membership and Interest Group Activity^ 61vList of FiguresFigure 3.1Interest Group Activity by Party^ 24Figure 3.2Interest Group Activity by Party and Region^ 25Figure 5.1Election Spending and Interest Group Activity^ 57Figure 6.1Issues Supported or Opposed by Interest Groups^ 68Figure 6.2Interest Group Issues and Party^ 73Figure 6.3Interest Group Issue by Region^ 74viAcknowledgementMany people have assisted me in the completion of this thesis. My supervisor, KenCarty, provided the initial encouragement to undertake graduate studies, as well as timelyhelp at critical points in the process of researching, data analysis and writing. I appreciatedhis collegial approach to supervision and his willingness to focus on my questions, regardlessof his other projects and pressures. Don Blake also provided advice on statistical tests andother data related questions.I would also like to thank Penelope Love, Rita and Tim Schouls, Karen Janes, LindaMcAllister, Michelle Voros, Lisa Anthony, Joanne Goertz, Richard Somerset, JenniferBennett, Mervin and Margaret Reimer, Frank and Elena Sovka, Caroline and Vince Reimer,and Deborah Strangward for their encouragement and their interest in this project.My family has had to bear most of the consequences of a tired, stressed and crabbywife and mother. Thank you David, James and TBA Sovka for your patience andforgiveness. I am most indebted to my husband, David Sovka, who provided constantencouragement, excellent proofreading, and demonstrated complete commitment to me and tothis project from beginning to end. I know that I could never have completed this projectwithout your love and prayers.viiChapter One: Current Explanations of Interest Group ElectionCampaign InvolvementWith an election call required before the end of this year, and a new Canada ElectionsAct recently signed into law, questions concerning what exactly constitutes a "fair" electionare currently being raised in Canada. One of the most controversial issues centres on theappropriate role of interest groups in election campaigns.' Advertising by groups andindividuals other than political parties and election candidates during the 1988 federalelection sparked much debate about who the participants in an election should be, and whatrestrictions, if any, should apply. The Royal Commission on Electoral Reform and PartyFinancing, for example, devoted one volume of its research exclusively to the question ofinterest groups and elections. The Commission's Senior Research Coordinator, LeslieSeidle, referred directly to the free trade issue in outlining the importance of "independentexpenditures" in elections and the prominence of this issue in the Commission's hearings(Seidle 1991).The leading scholar on Canadian interest groups,' Paul Pross, has developed ananalytical framework for understanding interest groups which takes very little account ofinterest group involvement in elections (1992, 1986). Basing much of his analysis on thedegree of internal organization of interest groups, Pross constructed an organizationalcontinuum designed to explain most interest group activity. In addition, he generated a'For example, the cover story of the February 22, 1993 Western Report focused on the implications ofthe changes to the Canada Elections Act for individual and group participation in elections.2Pross ' framework for understanding interest groups is frequently cited and described as the mostcomprehensive model of Canadian interest groups.(Stanbury 1993:133, Tanguay and Kay 1991:83 Jackson andJackson 1990:543, Galipeau in Gagnon and Tanguay 1989:410, and Kernaghan and Siegel 1987:400).1policy development model based on the idea of a "policy community" which is an informalnetwork of policy actors (Pross 1986). Interest group election tactics, on the other hand, areonly briefly mentioned in the chapter which addressed atypical policy development, when"contentious issues" arise. In this context, Pross admitted that some interest group electioncampaign involvement occurred but he described it as a new tactic employed by interestgroups that were marginalized in the regular policy development process yet desired to injecttheir concerns into the election debate. Pross was not convinced that an election campaignwas the best environment for interest groups to influence public policy and he stressed theinappropriateness of riding campaign involvement by highlighting the many drawbacks ofsuch activity.In contrast to Pross' superficial treatment of interest group election involvement, theConstituency Party Association (CPA) study of local associations' in all federalconstituencies after the 1988 election showed that fifty percent of them had interest groupsinvolved in their local campaigns, either promoting or opposing the association's candidate.Clearly, the level of interest group activity reported by the survey does not justify Pross'passing treatment of election involvement by interest groups. Because of the widespreadacceptance of Pross' analytical framework concerning interest group activity in policycommunities, and the apparent contradiction with the results of the CPA study, the new dataprovided in the CPA study needs to be explored. Is an election campaign an important3The Constituency Party Association study surveyed the constituency associations of five politicalparties in the spring of 1991. As one of the last research projects commissioned by the Royal Commission onElectoral Reform and Party Financing it sought not only to explore the range and variance of local associationactivity but to canvass the reactions of riding associations to various proposals for the reform the Canada'selectoral law.2avenue for interest groups to influence public policy? If Pross' analysis of typical interestgroup activity is accurate, why would so many constituency associations report interest groupactivity?In the rest of this chapter, I explore several notions central to the question of interestgroup involvement in local constituencies, including the efficiency of campaign involvementfor influencing public policy; the search for legitimacy of interest group through participationin the political process; elections as a forum for the public discussion in which interestgroups would like to participate; and the internal benefits to interest groups fromconstituency campaign involvement.When is Election Involvement Efficient?While efficiency is a concept usually employed in economic analyses, it can behelpful to analyze the efficiency of interest groups in order to ask what the rational, self-interested behaviour of an interest group would be given its policy agenda (Stanbury 1993). 4It may be difficult to view all interest groups, especially the altruistic ones, as seeking solelyto maximize their resources. A list of resources available to interest groups includes theobvious factors such as money, staff time, and research capacity but it also includes moreambiguous "commodities" including the perceived legitimacy or status of the group's causewith the public, knowledge of a particular sector or client group, knowledge of the policyprocess, key contacts and members, and the personal efficacy of staff members in making a4For a full discussion of political market theory see Reisman (1990), Landry (1987), and Hayes (1981).3contribution to a policy area (Reisman 1990, Pross, 1992). It is less than straight forward toefficiently allocate such diverse and often unquantifiable resources in an ever changing policyenvironment.Pross (1986:3) described interest groups as, "adaptive instruments of politicalcommunication, equipped with sensitive antennae for locating power." He viewed theactivities of interest groups as indications of the relative policy-making power of the variouspolitical institutions and claimed that as institutions gain or lose power, interest groups adjusttheir relationship with the institution either by expanding their contacts or letting theircontacts die away. Thus, the task for an interest group is to find and influence the mostimportant decision-maker(s) on each policy question. This is not a new idea; some of theearliest analyses of interest group behaviour came to the same conclusion. With reference tothe western liberal democratic phenomena of interest groups, Ehrmann, in an early work oninterest groups, commented, "...interest groups will concentrate their best efforts on thoseorgans of government which are responsible for decisions affecting most directly theirclientele" (Ehrmann 1958:6). And Truman claimed that, "...political interest groups willseek access to the key points of decision within these [political] institutions" (Truman inLuttbeg 1968:129-130). Presthus made a similar point when he argued that interest groups,"...seek access where the power and authority to shape policy reside" (Presthus 1974:3). Ina more current and popular tone, Harrison concluded that,To be successful, lobbyists must be familiar with the key figures in each of thepower cliques and be able to make accurate predictions about futureconfigurations of power. It is useless, from a business point of view, to befast friends with a minister who doesn't have the power of decision in hisdepartment, or with a senior mandarin whose political sympathies are blatantly4at odds with the ruling government. On the other hand, everyone in a seniorposition in Ottawa is at least potentially powerful so cordial relations must bemaintained (Harrison 1988:21).Because of the importance of influencing key decision-makers, Pross (1992, 1986) arguedthat the regular interest groups activity does not take place in electoral politics. Forexample, a farmers' group seeks to influence the bureaucrats in the Department ofAgriculture who are responsible for a pertinent regulatory decisions. If the person holding aposition changes, the interest group will continue its lobbying efforts with the new office-holder.When this same reasoning is applied to elected Members of Parliament or CabinetMinisters, a representative is lobbied because he/she is the Minister responsible for a givenportfolio, and not because of the particular constituency that he/she represents. It is difficultto apply the same argument to an electoral contest, where a candidate with an unknownfuture is lobbied (government or opposition, frontbench or backbench). Instead, interestgroups seeking to influence those in decision-making positions should not include electioncontests because they would not have a preference for one candidate over another at thatstage in the formation of a government.Another argument, which supports an election strategy for interest groups suggeststhat election candidates are policy actors too important to ignore. This conclusion should notbe lightly dismissed because interest groups, it is assumed, do not waste time or money onthe uninfluential or the powerless. The local campaign involvement reported in the CPAstudy supports the idea of the adaptive nature of interest groups as they seek out the mostimportant policy actors on a given policy question.5One assumption underlying the efficient power locating abilities of interest groups isthat interest groups must make a judgement about who the influential and powerful decision-makers are. As Pross' (1992, 1986) organizational continuum suggests, some groups have amore sophisticated understanding of the political system than others, and thus would be betterable to make such an analysis. The fact that a significant level of interest groups activity isreported by a given set of policy actors indicates, at best, that interest groups perceive themas powerful policy decision-makers.Another assumption is that interest groups have access to any policy actor theyconsider powerful. It may be that an interest group spends its resources on the mostinfluential decision-makers that it can influence rather than the decision-makers it would liketo influence. An issue-oriented group, for example, may have limited access to key actorsand its strategic lobbying choice may be to maximize the group's influence within itsconstrained range of policy actors. It is difficult to imagine how a rational resourceallocation for many interest groups would include attempting to influence the campaigns ofthe multiple candidates in all two hundred and ninety-five federal constituencies. For interestgroups whose interests are closely associated with one line department or contained withinone policy community, it would seem inefficient to attempt to influence the many localcontests in an election. Perhaps interest groups with more diffuse interests, whose policyinterests do not fit conveniently into one or several policy communities, would be morelikely to use the more general forum of an election campaign for influencing public policy.It can be argued that most electors do not base their voting decision solely on policyconsiderations; many electors place party loyalty or party leadership ahead of policy6considerations. This makes an election campaign a much more difficult forum for policyfocused groups to advance their causes and to make an impact on the electoral outcome. Theatypical nature of the 1988 federal election, however, where one important policy questiondominated much of the debate, may have prompted perceptive interest groups to becomeinvolved in the election campaign because it suddenly became possible to affect this policyquestion through electoral participation, a non-policy community mechanism.During 1988, abortion groups witnessed a number of key events and decisions in theirpolicy area. In January of that year, the Supreme Court of Canada struck down the 1969section 251 of the Criminal Code that addressed abortion, declaring it of no force or effect.This decision understandably provoked a response from both pro-life and pro-choice interestgroups. The federal government promised to act quickly but it was not until July that anabortion resolution was before Parliament where eventually every proposal was defeated.Perhaps the absence of legislation and the frustration of both pro-life and pro-choice activistsprompted their 1988 election involvement. No further litigation was possible and Parliamenthad not resolved the issue.In addition, the abortion issue in Canada, particularly the pro-choice position, hasbeen associated with feminism and women's issues.' Collins (1985:25) described how sincethe 1970 'Caravan,' "...the abortion issue and feminism in this country have held to astrained but cheek-to-cheek course ever since." It is also unlikely for an interest groupcentral to this policy community to become involved in a local constituency campaign'There are also health (federal and provincial) and legal (criminal and civil) aspects to the issue whichmake the resolution of any abortion policy very complex.7because it would be taking a policy issue outside of the policy community and subjecting it toa much wider public debate. This policy community, which claims to incorporate theabortion issue, has failed to include a significant dissenting viewpoint, the pro-life interestgroups (Tanguay and Kay 1991). If pro-life interest groups are marginal in the legitimateforum for policy development, this may have encouraged them to take the issue outside orbeyond the policy community to the broader policy environment.Searching for LegitimacyIn 1967, John Meisel commented that interest groups',...numbers and means permit them to become rivals of political parties. Theserapidly proliferating groups and institutions can gain access to the decision-making process without having recourse to parties which thus become by-passed in the vital area of mediation between individual and group interestsand the state (Meisel in Pross 1986:1).The exclusion of the elected members in completely bureaucratic and interest groupgenerated policy lacks legitimacy in Canada. A process that does not involve any electedrepresentatives of Canadians discredits the policy output regardless of its contents. Membersof Parliament, and consequently all election candidates, may have an important role in thepolicy process solely to provide the needed impression of public input, public debate andpublic approval for policy decisions (Pross 1992). A lack of legitimacy in a democraticsociety is very problematic because the free election of a government signifies the consent ofthe people to be governed. Without the responsibility to the electorate that comes withelected office, questions of public protection also arise. Interest groups, no matter how largetheir membership, do not have the same legitimacy as elected officials to make decisions on8behalf of the electorate. The activities of interest groups begin to threaten the state whenthey undermine the role of elected members.One of the ways of achieving the "required" legitimacy would be to stimulate policydebate during an election campaign. When election candidates and political parties take aposition on an issue and then are subsequently elected by voters, this provides at least astarting point for a connection between further policy development and the citizenry.Citizen ParticipationHaving interest groups prompt candidates to take a distinct position on a policy issuealso helps voters to distinguish one candidate for another during the election campaign shouldthey wish to incorporate policy into their voting decision.An election contest usually involves a wider group of citizens than are normallyinvolved in local party affairs or political activities of any kind. A policy question thatarouses strong and divergent opinions could motivate an even greater fraction of theelectorate and involve otherwise non-political or apathetic citizens. While such an influx ofnew participants may upset the inter-election balance of party interests, it could hardly besuggested that the involvement of the greater portion of the electorate was somehowdetrimental to the electoral process.Given that both political parties and interest groups aggregate and articulate policypositions during an election, and that they can either work together or against each other, thedistinction between the activities of an interest group and a political party can be difficult tomake. An interest group can become a party, or it can operate within or outside of apolitical party in its attempts to influence the political party's policy position and alter the9electorate's response to the political party. The comparative cost for an interest group tolaunch a new political party may be very high, which might encourage an interest group toattempt to "takeover" a political party from within. A nomination contest, for example, is avulnerable time for many political parties because their inter-election organization is usuallyweak and many riding associations can be manipulated with a limited number of people(Canadian Study of Parliament Group 1989, Tanguay and Kay 1991). The influx of newmembers for a nomination contest or the election itself can be positive for a local associationbecause it encourages greater participation, it may produce a more representative caucus, andit encourages political parties to address the substantive issues important to interest groupswhich are presumably important to at least a significant portion of the electorate. Some ofthe disadvantages of constituency association infiltration is that Members of Parliament whoowe their election to one interest group and are keenly focused on one policy area may berigid and uncompromising in their views, proving themselves to be difficult colleagues eitherin government or opposition.National Forum for Policy DebateWhen an election is viewed as an opportunity for wide discussion of issues among allgroups in society, the partisan nature of an election is diminished and the political parties aremerely vehicles for the exchange of ideas. One interest group that held this view in 1988was the Canadian Peace Pledge Campaign. During the 1988 election campaign, a groupspokesman claimed that, "...the electoral process does not belong solely to the politicalparties. We have important issues to discuss too" (Clark 1988:19). They believed that theyhad a place along side of political parties, and indeed, almost parity with them. In a similar10statement, the president of the National Citizens Coalition, referring to the Alberta rulingwhich struck down a law prohibiting any group other than political parties from spendingmoney in an election said that, "...the parties have had their stranglehold broken on thepolitical process" (Clark 1988:19). These groups suggested that it may not be the politicalparties who should determine key election issues or even the parameters of an electiondebate.^Kernaghan and Siegel suggest that, "...the growth in the number and apparentinfluence of pressure groups has raised concern that they rival and even threaten to supplantpolitical parties as the primary channel of communication between citizens and theirgovernments" (Kernaghan and Siegel 1988:415). They argued that collectively, interestgroups decrease the role of political parties as intermediaries between the citizen and thestate. This shift away from parties and elected Members of Parliament can be partiallyexplained as the rational and self-interested action of citizens who will choose whicheverchannel of communication best enables them to influence public policy (Kernaghan andSiegel 1988, Archibald and Paquet in Gagnon and Tanguay 1989).Tom Kent, in a article which called for renewed citizen participation in all parties,claimed, "...in our representative democracy, the responsibility for deciding public policylies almost entirely with politicians; that is, with the leading members of political parties"(Kent 1988:3). While this view excludes the central role of the bureaucracy, the executive,the judiciary and other groups that influence the policy process, it does highlight the role ofthe people who join political parties and especially those who let their names stand fornomination.11Kent suggested that the parties have failed to keep the confidence of Canadiansbecause they have lost their policy making role and their platforms are hollow. The partiesare, "...putting presentation before performance, the packaging before the product" (Kent198:3).Hiebert (1991) claimed that the increased activity of interest groups in elections hasadded a new dynamic to election contests where a general election becomes one of the mainopportunities for the electorate to participate in the political system. Because interest groupsquestion the ability of parties to adequately represent all of the interests in society, they arebecoming increasingly involved in election contests. A more cynical view of interest groupinvolvement in election campaigns suggests that their participation is merely a cheaper wayto access the political agenda.Internal Benefits to Interest GroupsFrom the perspective of a local interest group, an election contest may be animportant public venue to demonstrate to its own membership that it is actively involved ontheir behalf (Stanbury 1993). In addition, participation in the electoral process may providegroup members with a sense of efficacy. Because of activities associated with the election,the local membership could be galvanized to recruit new members and thus increase theirprofile in the community. The resulting solidarity and increased local strength of the groupmay increase the group's resources providing further momentum. The larger and moreactive the local membership, the more legitimacy the group's representative would have inany other policy forum such as public hearings, or advisory boards.12The CPA data set provides an opportunity to study the Canadian federal ridingassociations on a system-wide basis and to explore what evidence, if any, exists for the ideasfound in the literature about interest groups and their electoral involvement. The nextchapter addresses the limitations and strengths of this data set and of the central surveyquestions about interest groups.The subsequent four chapters explore the occurrence of interest group involvement inconstituency association campaigns as reported in the CPA study data. Specifically, ChapterThree focuses on the political parties and asks whether or not the different parties had similarexperiences with interest groups in the 1988 federal election campaign. The principal ideainvestigated is the government challenge notion where election involvement is primarilyviewed as an occasion for interest groups to question government candidates on the policyrecord of their most recent term. Chapter Four searches for geographic variation in interestgroup election involvement employing primarily provincial and regional comparisons.Chapter Five examines various characteristics of the associations to test for significantrelationships with interest group involvement. These include the relative competitiveness ofthe association, the type of campaign staged in the 1988 federal election, the association'sfinancial resources, and the number of local association members. Chapter Six considersonly the associations where interest group activity was reported, concentrating on the issuethat motivated the interest group's involvement. The final chapter concludes by summarizingthe findings from the CPA study that relate to interest group activity and speculating aboutfutures levels of interest group riding campaign involvement.13Chapter Two: The Constituency Party Association StudyIn this chapter I will highlight the most important methodological issues that relate tothe Constituency Party Association (CPA) survey data which was the principal data set 6 forthis thesis. The data for the CPA survey was collected for the Royal Commission onElectoral Reform and Party Financing in the spring of 1991.The CPA data set was chosen because it was the first system-wide study of theconstituency associations of Canadian political parties. Numerous studies have beenconducted on individual voters (such as the National Election Studies undertaken for eachelection), political activists, and convention delegates (Perlin 1988, Archer 1991); the CPAstudy shifted the focus of analysis from the individual participants to the constituencyassociations. In his volume for the Royal Commission on Electoral Reform and PartyFinancing, Carty stated the threefold objective in the Constituency Party Association study asseeking a) to fill the gap in the political science literature regarding local associations, b) toassess the current rules and process of local associations, and c) to obtain grassroots reactionto a number of reforms to the electoral system suggested through the Royal Commissionprocess (Carty 1991). Currently only Carty's volume for the Royal Commission has madeextensive use of this data set, leaving many areas of local experience open for furtherexploration.6Two additional data sets, created from information from the Chief Electoral Officer, were joined tothe Constituency Party Association data set to expand the analysis. One set of data included the 1988 federalelection results for each constituency (number of votes for each candidate, party affiliation, and rank in the finalresults). The second data set included financial information for each riding by candidate (contributions andspending).14The type of questions contained in the CPA survey indicate that the unit of analysis isthe constituency association. Many of the survey's questions address issues that are notapplicable to individuals. For example, questions about membership numbers, associationincome, or association activities were included. The central questions to this thesis regardinginterest group involvement in local association campaigns would be categorized in this way.Because the riding associations are the vehicles through which many of the tasksprescribed by election law are carried out, the Royal Commission on Electoral Reform andParty Financing sought to understand in a comprehensive way the practical experience of thelocal associations. The experience of all of the constituency associations was requested toexplore the previously unknown range and variance of constituency association activities.The Royal Commission surveyed the entire local association population in Canadaduring the spring of 1991 which included the local associations of the three major parties (theProgressive Conservative Party, the Liberal Party, and the New Democratic Party) and twominor parties (the Reform Party and the Christian Heritage Party).Each of the parties provided its most comprehensive list of constituency associationsand their presidents. There were some constituencies where no address was available or thesurvey was returned as undeliverable but no attempt was made to select certain associationsor to choose a sample of the ridings. Rather, the effort was made to include the entirepopulation of constituency associations of five parties from as many of the two hundredninety-five constituencies as possible. The survey was sent in the mail with three follow-up15efforts, including two reminder cards sent to constituencies that had not responded, andfinally telephone follow-up used to allow constituency associations to request another copy ofthe survey.The final survey response rate was fifty-four percent of the constituency associationpopulation. To ensure that this subset of the population was representative of all localassociations, two key characteristics of the responding associations were noted. The firstimportant characteristic was political party affiliation. In order for the Royal Commission tomake the best decisions about the electoral process, all parties should be allowed a chance torespond and indeed, all five of the parties included did respond to the survey in sufficientnumbers' (Carty 1991). All of the political parties included in the survey were aware of themandate of the Royal Commission to make recommendations to Parliament regardingchanges to the electoral laws that govern their activities. There was, therefore, a strongincentive for each constituency association to ensure that its experience and particularly itsreactions to the "Options for Reform" questions were considered. The second importantcharacteristic was province. Quebec and New Brunswick had lower response rates than theother provinces but all provinces had an acceptable quantity of surveys returned.''The lowest response rate by party was from the Liberals at forty-seven percent. All of the other partieshad over fifty percent of their constituency associations respond to the survey. The NDP had fifty-three, theProgressive Conservatives had fifty-seven percent, the Reform Party had fifty-nine and the CHP had the highestresponse rate of seventy-nine percent. The response rate for the CHP was quite a bit higher than the otherparties, however, the total number of CHP cases is only forty-eight or 9% of the 522 case total which was notenough to significantly skew the results.'The response rates by province were: Newfoundland 63%, Nova Scotia, 61 %, Prince Edward Island54%, New Brunswick 43%, Quebec 39%, Ontario 60%, Manitoba, 60%, Saskatchewan 57%, Alberta 60%,British Columbia 53%, and the North 63%.16Questions Pertaining to Interest GroupsTwo of the questions in the CPA survey specifically addressed interest groupinvolvement in constituency campaigns. 9 Because much of the present study's analysisutilized these two questions, it is important to address their limitations. The first questionasked, "Did any single-issue interest group play an active role in the local campaign in anattempt to support or oppose your candidate?" This was followed by the second question, "Ifso, what issue were they promoting or opposing?"The survey asked about "single-issue" interest groups which may have influenced arespondent's answer depending on whether or not he/she viewed the interest groups that wereactive in their local campaign as having only one issue. This introduces some uncertainty.A group formed solely to fight the free trade issue, for example, would clearly be considereda single-issue group. However, would a business or labour group that focused all of itselection activities on the free trade issue be reported as a "single-issue" interest group? Howeach respondent would decide such a question is unknown and is likely unknowable. It maybe that the survey captures only the subset of local interest groups viewed as "single-issue"groups.It is also possible for respondents to interpret the activity of interest groups in avariety of ways. For instance, what does it mean for an interest group to be "activelyinvolved"? Does this include the participation of regular party members who were alsointerest group members? Does the interest group activity have to be official or institutional9There were a number of questions which deal with groups that have interest group expression in thepolitical process such as women, youth, or aboriginal peoples but these questions did not address interest groupsin general.17for the respondent to attribute the activity to the interest group? Because the participation ofindividuals may be ambiguous, it is impossible to know how each respondent defined"interest group activity."Another problem with the "single-issue interest group" question is that it is limited toeither the presence or absence of such groups; no attempt is made to explore the frequency,intensity, or scope of the interest groups activities. Any further work in the area of interestgroup involvement in constituency campaigns should consider investigating these factors.The implications are quite different if interest groups are making occasional contact with aconstituency associations during an election campaign, versus their direct involvement in theongoing work of the association such as policy development or the election of constituencyassociation executives. A further question might have been to ask what changes, if any, alocal association would undertake in future election campaigns as a result of the experienceof interest group activity in the 1988 election.Finally, it may be important to ask not only the issue that an interest group waspromoting or opposing, but the name of the specific group involved. For example, in thecase of the abortion issue, it might be important to determine whether pro-life or pro-choicegroups were more often using local constituency associations in their election strategy. Itmight be revealing to know which issues prompted which interest groups to undertake localcampaign involvement; it is possible that the policy issues pursued during an electioncampaign may or may not be the same as non-election issues pursued by the same group.18Any differences would be intriguing and would further nuance the information about interestgroups and their activities.'What the two questions in the CPA survey do provide is the first system-wide attemptto explore interest group involvement in local campaigns. The responses suggest a generallevel of interest group activity in local associations and provide some indication of whichissues prompted such action in 1988. Because the whole area of interest group involvementin elections is a relatively recent phenomena, there is a need for the basic patterns to beestablished before the understanding can be expanded.The questions regarding interest groups in the CPA survey provide a profile of thelocal constituencies across Canada at one particular time. The survey was administered inthe spring of 1991 and many of the questions refer to the local associations experience in the1988 federal election. The particular political dynamics of that time will be reflected in thesurvey results. For example, the Reform Party was experiencing incredible growth and theLiberal Party had recently chosen a new leader at its 1990 leadership convention in Calgary.Given that political fortunes and parties' experiences are in constant flux, a series of glimpsesover time is probably the best reflection of the changes that can be expected. The CPA dataprovides one such glimpse.Carty's analysis of the CPA data set highlights the most important overall findingsregarding constituency association experience. In the following chapters I will build upon hisanalysis by focusing exclusively on the local campaign activity of interest groups. Any'°Some of these questions could be answered by a detailed analysis of selected cases. The range ofparties, regions, constituency association characteristics, and motivating issues should be balanced whenselecting case studies.19significant relationship will be reported and discussed with reference to any predictions aboutinterest group campaign activity found in the literature.Specifically, I will explore interest group activity in two of the most enduring areas ofpolitical differences in Canadian politics, political party affiliation and geography." Howthe phenomena of single-issue interest groups occurs in these two contexts is importantbecause it builds on much of the existing understanding of Canadian politics. If interestgroup involvement in local campaigns occurs evenly in all geographic areas and in allparties,' then other factor(s) must be encouraging their involvement.I will also explore whether or not the characteristics of the local associations wereassociated with different levels of interest group involvement. Finally, I will examine theissues that motivated interest groups to become involved in riding association campaigns,searching for significant associations and trends.Whatever relationship are found between these four areas and interest group campaigninvolvement will provide the first national profile of the interaction between interest groupsand the local associations of Canadian political parties. At the present time, no comparisonscan be made but future studies of the constituency based activities of interest groups mightuse this analysis as a point of reference to indicate how local interest group activity haschanged since 1988."Courtney and Perlin claimed that, "...no set of cleavages has been more central to the political life ofCanada than those related to the country's regional diversity" (Courtney and Perlin in Perlin 1988:124).12This is the null hypothesis, that there is no relationship between interest group activity and geographicor political party characteristics.20Chapter Three: Calling the Government to AccountPolitical parties are the central actors when the electorate is required to choose whowill govern the nation, making the necessary public decisions on behalf of all citizens.Because of this, the attentions of those interested in the political process are focused on thepolitical parties during an election. In the context of the present study, the central questionconcerning the partisan affiliation of local associations is whether or not interest groups focustheir attention equally on all parties.One reason interest groups attempt to influence the political parties during an electionis that they view an election as an appropriate method for challenging the current governmentand putting their issue onto the public agenda. They believe that their issue is important, andinadequately or inappropriately addressed by the current government to that point in time.The intense media coverage of an election campaign, and the heightened interest of theelectorate in public policy, provides an opportunity for interest groups to gain a greateraudience for their policy concerns.Incumbents and other candidates associated with the governing party are attractivetargets for local campaign involvement because they are in the position of having to defendthe government's prior action or inaction on an issue. For example, Every Voice Counts, "awomen's guide to personal and political action" prepared by the Canadian Advisory Councilon the Status of Women, suggests a number of strategies for local women's groups to use ininfluencing public policy. Under the heading of "election time tactics," the CanadianAdvisory Council on the Status of Women (Kome 1989:44) advised local groups not to,"...let an election go by without raising women's issues during the campaign...This is the21time that people who have power or who want power are most accessible." The first actionsuggested is that local groups target the incumbent on the basis of an inadequate track recordon the group's central issue, followed by increasing the awareness of all candidates on theissue.If interest groups perceive an election as an occasion for easier access to governmentMembers of Parliament, I expect that in local campaigns, interest groups would be morelikely to be involved in Progressive Conservative party campaigns. As a corollary, I expectthat interest groups should then seek an audience with the other candidates, based on theirwillingness to join the interest group in challenging the candidate associated with thegovernment's previous record. The other two major parties, the Liberals and the NDPwould be logical choices because of their relative electoral strength compared to the minorparties. I expect that interest group campaign activity would be lower for the minor partiesbecause of their more limited electoral support, and therefore limited capacity to mount aserious challenge to either an incumbent or candidate associated with the governing party.A second reason for interest group involvement in local campaigns is that someinterest groups have pre-existing alliances with some parties over others. The labourmovement, for example, is believed to be a "natural" ally of the New Democratic Party.While Archer (1990) argued that the official link between the NDP and trade union locals isweaker and of less significance than is popularly believed, it could be that labour interestgroups are more likely to be involved in local NDP campaigns, presumably supporting thelocal candidate. The Progressive Conservative associations might also experience labourinterest group campaign involvement but opposing their candidates. The centrality of the22issue of free trade with the United States in the 1988 election campaign, where labour groupswere clearly opposed to the position taken by the Conservatives (Campbell and Pal 1991),also suggests that they might actively oppose Conservative candidates and support NDPcandidates.The CPA data does show a significant association between the political partyaffiliation of a constituency association and whether or not interest group activity wasreported during the association election campaign (V=.22 and p < .00.' 3 The five partiessurveyed showed as much as thirty-eight percentage points variation (Figure 3.1).The Progressive Conservative associations reported the highest level of interest groupcampaign involvement, with sixty-two percent of Conservative associations reporting suchactivity. This result provides some evidence for the first expectation, that interest groupswould be more likely to involve themselves in the local campaigns of candidates affiliatedwith the most recent government, using the election as an opportunity to challenge thegovernment's record.On a regional basis, the Progressive Conservative associations in Atlantic Canada hadthe highest level of interest groups involvement in their campaigns, greater than any otherregion (77%) (Figure 3.2). Conservative associations in Western Canada also had a highlevel of interest group campaign involvement (70%). Progressive Conservative associationsin Ontario reported an incidence of interest group activity similar to the overall level for the13Cramer's V, a chi-based measure, measures the strength of association between nominal variables. Itvaries between 0 and 1 where 0 indicates no association and 1 indicates compete or perfect association. To testif a result is significant, only cases where p< .05 or p < .01 will be reported and described as significant.Generally, an association of .00 to .19 will be considered weak, .20 to .29 will be considered moderate, .30 andgreater will be considered strong.23706050403020100Figure 3.1Interest Group Activity by Party% Associations with Interest Groups Active► a, r err,.•^.•^.•^.•^.•^.•.•^.•^.•^.•^.••e e^e0 .^/ / / /•e^r •.• e."e^• . •r•^...•efefle•r• •^/^/ /• •^•^•.•^. . . .  : ;f ' ii% -; ir.• .•.•• . .• .•.^.e^e^e^•.^.^.^.^.^..•^.•^.•^.•^.•^.•^.•^•..• . • .• .• .ers..e.e.e.e.e.e•■••^." r•^ / / ••••^e.•^.^.•—10e^ f.°^.*^.•.^.^.^.^.^.^.^.e.e.e.e.e.e.e.e..•.• e^. ••.e e •^.•^.•^.•eee//eVe ef eff / / /  / /.•^.•^.•^.•^.•^.•^.•i f ' f'i f'i / ' ir•0..1 / I I I . I I .^e / ^e• /.• ^.•^.••.•• .•ede.e.e.e.e.ese^.• ^.*/..•^.•^.^.•^.•^.• .•^.•^.•^.•^.•^.•^.•^.• .^.•^.•^." .^•^. ^•^ erre...e'er ^ - ^ 0.000000/00•••.•.•^•••^••.••e ere/ere, •.•^.•• .. .•.• ./ . / e* e•All PC Liberal NDP Reform CHP•f "^.^.•.•. . .^..•.•.^.•.• . •/ 'e e 'er, r eeee. ee •ee.•e•• • • • •^^•(Cramer's V=.22 and p<.01)Interest Group Activity by Party and Region80 % Associations with Interest Groups Active706050403020100Figure 3.2PC Liberal NDP West E3 Ontario E2Quebec IMAtlanticparty (63%). Quebec was much lower than any other region, with only thirty-five percent ofthe Progressive Conservative constituency associations experiencing single-issue interestgroup involvement which reflects a pattern very similar to the overall regional trend ofinterest group involvement."The Liberal constituency associations reported an overall level of interest groupactivity of forty-five percent, slightly lower than for all parties. On a regional basis,involvement in local Liberal campaigns remained fairly close to the overall Liberal levelexcept in Quebec, where significantly fewer constituency associations (23%) reported interestgroup activity.The NDP reported the second highest level of interest group activity in its localcampaigns. Fifty-four percent of the responding NDP constituency associations had single-issue interest groups active. On a regional basis, the NDP displayed a statistically significanteast-west trend in interest group involvement, where it was increasingly more likely forinterest groups to be involved in local NDP campaigns, progressing westward from Atlanticto Western Canada (tau-c=-.26 and T=-2.72)." This pattern parallels the strength of theNDP in the different regions of Canada. Historically, the party has been much moresuccessful in Ontario and in Western Canada than Quebec or Atlantic Canada.14A more in-depth discussion of the geographic basis for interest group involvement is found in the nextchapter.15Tau-c will be used to capture not only the strength of the association but also the direction. Kendall's tauvaries between -1 and +1 where an absolute value of .00 to .19 was considered weak, .20 to .30 wasconsidered moderate and values greater than .30 was considered strong. Usually party and region will beanalyzed as nominal variables unless some reasonable ranking such as an east-west or left-right scale can bejustified.26Unlike the Conservatives and Liberals, NDP associations in Atlantic Canada were notlikely to have interest groups active in their local campaigns. This result provides furthersupport for the notion that interest groups choose to influence the campaigns of candidatesassociated with the government and perhaps the opposition but parties with little electoralsupport, such as the NDP in Atlantic Canada, are less likely to receive interest groupattention.The second reason for interest group partisanship, that certain interest groups havenatural connections to some parties, (and in particular the question of a labour-NDPrelationship) is a more complicated issue. In a cross-tabulation of political party affiliation,region, and labour interest group activity, most of the labour group activity was reported bynon-NDP associations. In accordance with the suggestion that interest groups challengecandidates associated with the most recent government, fifty-five percent of all labour interestgroup activity occurred in Progressive Conservative campaigns. The remaining labour groupactivity was equally split between the Liberal and NDP local campaigns. This furthersuggests that labour groups act in a manner similar to other interest groups by choosinggovernment candidates first, and then focusing remaining resources on opposition partycandidates. No labour groups were reported to have been involved in either Reform or CHPriding campaigns.A regional analysis supports this conclusion. Among the regions, labour interestgroup activity was evenly distributed, but there was a strong partisan element. In AtlanticCanada, labour interest group involvement was almost exclusively associated withProgressive Conservative campaigns (80%) and none was reported in NDP riding27associations, again supporting the "governmental challenge" theory of interest groupinvolvement. In Quebec labour groups were reported in all three major parties with the NDPthe most likely party for labour interest group activity. In Ontario, Progressive Conservativeriding associations were again the target of most of the labour interest group activity (67%).New Democratic campaigns in Ontario received only seventeen percent of the labour interestgroup campaign activity in Ontario. In Western Canada, labour interest groups were lesslikely in NDP campaigns than in the Liberal or Conservative riding association campaigns.This focused analysis of the labour-NDP relationship provides no direct evidence that labourgroups are more likely to support NDP candidates in local campaigns but contributesempirical evidence for the "governmental challenge" idea.An alternate explanation of the low level of labour interest group campaigninvolvement reported in NDP constituency campaigns is that NDP associations do notperceive labour groups as "single-issue interest groups." It could be that NDP associationsview labour groups as providing a broad and multi-faceted critique of the Canadian economyand society. As result, they would not classify labour groups as "single-issue"interestgroups. In a similar way, the involvement of labour group members in NDP associationsmight be so integrated with regular association activity that they are not viewed as external tothe association or promoting or opposing an external policy agenda.While a direct measure of incumbency was not available, interest group activity wasno more likely in the riding campaigns of winning or losing candidates.' While the16For the Progressive Conservative riding associations V=.15 and p> .05, for the Liberal ridingassociations V=.06 and p> .05, and for the NDP associations V=.08 and p> .05. No statistics are availablefor either the Reform or CHP riding associations because none of their candidates won their local contest.28different parties experienced different levels of interest group activity, individual candidateswithin each party were not supported or opposed at significantly different levels. This resultalso supports the government challenge idea where interest group involvement is more likelyin the campaigns of candidates affiliated with the governing party.The CHP had fewer interest groups either promoting or opposing its candidates (41 %)than the three major parties, as did the Reform Party constituency associations, whichreported that only twenty-four percent (24%) of their local associations had interest groupsactive in their 1988 federal election campaigns. The Reform Party, which did not run anycandidates in Quebec, had a rate of interest group involvement even lower than Quebec's(24%). Because there was no Quebec component to this low result, it is in a sense "doubly"low.Neither the CHP or the Reform Party ran a full slate of two hundred ninety-fivecandidates and so a regional analysis is not revealing. The CHP candidates were scatteredacross all the regions, excluding Quebec, with an equal level of interest group activity in allregions. The Reform candidates ran exclusively in Western Canada. The generally lowerlevel of interest group campaign involvement in both the CHP and Reform Party is consistentwith the notion that interest groups are more likely to choose local campaigns that areaffiliated with the most recent government, and less likely to chose the local campaigns ofparties that are electorally marginal.In summary, the involvement of interest groups in constituency campaigns wassignificantly related to the political party affiliation of the local association. The ProgressiveConservatives consistently had a higher level of interest groups involvement in their local29campaigns, even across regions, which supports the idea that interest groups might beinvolved in local campaign contests to call the government, or its representative candidates,to account for their actions during the last term. The idea of an NDP-labour link is notsupported by the CPA data because no evidence was found to suggest that labour interestgroups are more likely to be active in NDP riding campaigns, however, this may be partiallya result of the wording of the survey question and the type of relationship the NDP has withlabour groups. The two minor parties, particularly the Reform Party, experienced much lessinterest group involvement in their local campaigns in the 1988 federal election than did themajor parties.30Chapter Four: Geographic VariationThe physical, social and political geography of Canada are factors that must be takeninto account by any interest group that is attempting to make a national impact via electoralpolitics because the current system of representation in Canada is based on geographicconstituencies. One of the most enduring problems for political parties has been thechallenge of electing a federal government which reflects a national mandate or nationalconsensus. Confederation itself can be viewed not as a move toward unity but as a system ofcompensation for existing regional diversity (Muir 1975). When interest groups becomeinvolved in electoral contests they also encounter many of the same challenges in organizingacross diverse economies, cultures, religions, languages and levels of urbanization, factorswith which Canadian political parties have always struggled (Thorburn 1985, Curtis andTepperman 1988).It may be that interest group involvement in local constituency campaigns also reflectssome of the same geographic variables that have defined so much of Canadian politics.Pross argued that interest groups are organized and activated on a basis other than geographyand constructed much of his analysis of interest groups on the distinction between spatial andsectoral interests (Pross 1992). Spatial interests are geographically based interests such asthe relative importance of a natural resource to a local economy. Sectoral interests which aregenerally related to economic variables, are not necessarily the same as spatial interests. Forexample, all of the grocery store owners across Canada, regardless of location, have asectoral interest in legislation that affects retailers regardless of their geographic location.31Thorburn (1985) also raised the question of the incompatibility of territorial andsectoral interests, however, he believed that geographic variables, such as provincialboundaries, have had a significant impact on interest groups. Pross (1992, 1986) held acontrary position and argued that geographic variables, while historically important, havebeen superseded by sectoral variables. Pross argued that the rise of interest groups is areflection of the public realization that the most efficacious way to influence public policy isnot through geographic representatives, such as Members of Parliament, but through interestgroups organized along sectoral lines.The CPA data set includes a number of geographic variables such as province,constituency, and level of urbanization and, therefore, it is possible to explore whether or notinterest group local campaign involvement is related to, or independent of, these geographicvariables. Because there is no clear consensus in the interest group literature regarding therole of geographic factors on interest groups, the results of this analysis provide somedirection to the understanding of the geographic dimension of interest group involvement inconstituency electoral politics.Province and RegionThorburn (1985) centred his analysis of interest groups on the dynamics ofgovernment-interest group relations in a federal state. Throughout his study, Thorburnfocused on the tension between national and provincial jurisdictions in Canada and how thisinfluences interest groups. He believed that the structure and function of interest groups areshaped by the federal nature of the Canadian system. The expectation for province andregion, according to Thorburn, would be little variation between provinces because all32interest groups are subject to the same conditions and limitations of federalism and anyvariation would be explained as due to differences in provincial political culture. Forexample, Thorburn (1985) suggests that Atlantic Canada should have a low rate of interestgroup involvement because of the strength of political patronage in the region and the abilityof the two traditional parties to include most interests.'As mentioned in Chapter One, the rate of interest group campaign activity for allconstituency associations was fifty percent. There was a similar level of single-issue interestgroup involvement in most of the provinces and territories except Quebec (Table 4.1)2 8Considering each province individually, Newfoundland was exactly on the average, NovaScotia and New Brunswick were slightly lower, and Prince Edward Island was slightlyhigher. Each of the Atlantic provinces had less than twenty cases and because the provincialresults were so similar, these four provinces were grouped together as the "Atlantic" region.As a region, the Atlantic provinces reported that forty-nine percent of the local constituencyassociations had interest groups involved in their campaigns.17Thorburn is also concerned with the systemic imperatives that federalism places on interest groups whichhe believes that unduly impede smaller, recently formed or less wealthy interest groups. He argued that, "...byimposing one form of collective action (based on territory) upon another (based on functional, pluralistcompetition), federalism increases the external and internal tensions that would be much less for interest groupsin a unitary system" (Thorbum 1985:56). In addition, "...since groups must be prepared to respond to eitherlevel of government, they must maintain well-staffed offices at two levels or risk losing influence" (Thorbum1985:58). While posing organizational barriers to interest groups, Thorburn does not suggest that federalism isnot a good system for Canada, rather he views it as the necessary price of union.18There is an unavoidable dilemma in isolating groups of respondents based on province because of thewidely uneven sizes of the populations designated as provinces and territories. Separating Prince Edward Islandand the two northern territories into discrete groups does not provide an adequate number of cases for statisticalanalysis. Much of the present study relies on statistical indicators and when so few cases exist, even with ahigh response rate, the conventional tests are not reliable.33Table 4.1Interest Group Electoral Involvement by Province and RegionPercent of Constituency AssociationsReporting Interest Group Campaign ActivityAll 50%Atlantic Canada 49Newfoundland 50Nova Scotia 47Prince Edward Island 57New Brunswick 46Quebec 33Ontario 57Western Canada 51Manitoba 41Saskatchewan 54Alberta 54British Columbia 53Yukon/Northwest 50Territories(Cramer's V=.16 and p < .01)Quebec was the one province that differed significantly from the other regions in therate of interest group involvement. Quebec constituency associations reported interest groupactivity of only thirty-three percent, which was seventeen percentage points lower than thenational rate.Ontario was not dramatically different from the national results and over half (57%)of its constituency associations reported interest groups actively supporting or opposing localcandidates. I grouped the four western provinces and the two territories together as the"Western Canadian" region because the overall level of interest group involvement was verysimilar between these provinces. Manitoba reported a result nine percent lower than thenational rate at forty-one percent but Saskatchewan, Alberta, British Columbia, the Yukonand Northwest Territories had rates of interest group involvement within four percentagepoints of each other.Regional differences in the occurrence of interest group involvement in constituencycampaigns clearly did exist in the 1988 federal election. Quebec stands out from all of theother provinces and regions because it had a significantly lower level of interest groupinvolvement (V =.16 and p < .01).A number of explanations in the interest group literature suggest why Quebec's rateof interest group electoral involvement would be lower than that in the other regions. Themost frequently cited explanation is that the Quebec electorate has a fundamentally dissimilarand negative view of interest groups (Tanguay and Kay 1991, Boivin 1984). The 1977Quebec election finance law, La Loi sur le financement de partis politiques, is viewed as theprovincial expression of this disapproving attitude. Boivin (1984), for example, pointed to35this legislation as an indication of Quebec's distinctiveness in its view of lobbying andinterest group activity in general. The legislation allows only individual electors to makedonations to political parties or election candidates and excludes groups of any kind.Tanguay and Kay described Quebec's attitude as a, "...suspicion that any form of third-partyspending is vaguely illicit" (Tanguay and Kay 1991:97), and a fear of even the possibility ofcorruption and influence peddling. In their study of interest group activity in the 1988federal election, which included a survey of the interest groups, Tanguay and Kay found that,"...groups in Quebec were overwhelmingly opposed to third-party spending during elections,with almost 90 percent supporting a ban on this activity" (Tanguay and Kay 1991:96). Thesesuggestions point to a unique provincial attitude toward election financing which may have a"spill-over" effect, causing a lower rate of interest group campaign involvement in thefederal election context.Another explanation of the significantly lower level of interest group involvement inQuebec relates to the specific issues that interest groups were promoting or opposing in the1988 federal election. Most of the interest group activity in the other regions of Canada wasmotivated by two issues, abortion and free trade with the United States. Quebec isdramatically different from the other regions because abortion groups were insignificant inQuebec local contests. The Quebec electors' views on abortion' are different from theother regions because support for abortion on demand is higher in Quebec, but so is supportfor a complete ban on abortions under any circumstances. Of the Canadian regions, Quebec19A more thorough discussion of the issues that motivated interest groups to become involved in localelectoral contests is in a later section of this study.36had the most polarized views on the issue and the lowest proportion of respondents choosingthe middle category of abortion under certain circumstances (knish 1989:16).Because Quebec is the only region that had a notably different level of interest groupactivity in constituency campaigns, the question of language must also be addressed. Atissue is whether or not the level of interest group activity reported in Bilingual Quebecridings (generally those with an important anglophone component) was different from that inthe French ridings." If interest group activity in constituency campaigns was related tolanguage, the Bilingual Quebec ridings' would be expected to have a higher level ofinterest group involvement than the French ridings because all of the other provinces andregions which are primarily anglophone, had a higher rate of interest group involvement.The data shows that there was no significant difference in the reporting of interest groupcampaign involvement associated with language. The Bilingual ridings reported a rate threepercent lower and the French ridings reported four percent higher than the rate for all ofQuebec (V = .08 and p > .05).A similar result was found when considering the language of the survey therespondent chose to return. If Quebec's lower level of interest group activity was related tolanguage and the anglophones in Quebec more closely resembled their anglophonecounterparts in the rest of Canada than did the francophones in Quebec, those respondentswho chose to respond in English (21% of the Quebec group) would have reported a different"The Chief Electoral Officer classifies each federal constituency as either English, French or Bilingual.All of the ridings in Quebec were categorized as either French or Bilingual and all of the Bilingual ridings werein Montreal (Every constituency association in Quebec received a French and an English copy of the survey).21Among the Bilingual constituencies across Canada, ninety-seven percent of all of the French responsescame from Quebec (1 from New Brunswick, 1 from Ontario and 58 from Quebec).37and higher level of interest group involvement than those who responded in French. Again,the data show that this was not the case; the French and English respondents had equalproportions of interest group involvement in local constituency campaigns (V=.02 andp> .05). Quebec had a lower level of interest group activity which was unrelated tolanguage. The factor(s) which produced the difference in Quebec affected all constituencyassociations regardless of language.Region was an important geographic variable with regard to interest group campaigninvolvement because Quebec had a rate of such activity significantly lower than any otherregion. The evidence provided by the CPA data suggest that Pross' analysis of the relativeunimportance of geographic variables and interest group activity be reconsidered with respectto interest group electoral participation.UrbanizationDespite Canada's vast land area, the country's population is concentrated in a fewurban areas. According to Curtis and Tepperman (1988), Canada is seventy-five percenturban and sixty percent of the total population lives within the Quebec City - Windsorcorridor. The median urbanization rate for the responding constituency associations in theCPA study was quite similar to the level found by Curtis and Tepperman (72%).Urbanization is a geographic variable to consider with regard to interest groupinvolvement in electoral politics because rural constituency associations may have had adifferent experience with interest groups than urban associations. The simple reason beingthat the proximity of urban ridings may make the organization of electoral participationeasier in urban areas.38The number of urban and rural electors in each constituency as defined by the ChiefElectoral Officer was used to create a measure of the degree of urbanization which expressedurbanization as urban electors as a proportion of all electors. For example, if a riding iscompletely rural, none of its electors are in urban polls, and it would be described as zeropercent urban.For all cases, a higher level of urbanization was related to interest group activity inlocal campaigns but the association was not significant (tau-c=-.08 and T=-1.67). This wastrue for all of the regions and all of the political parties. While an acceptable significancelevel was not obtained, this relationship was stronger in Atlantic Canada where constituencyassociations with a higher level of urbanization also were more likely to report interest groupactivity (tau-c=-.19 and T=-1.51). Western Canada also had a similar association betweenhigher levels of urbanization and interest group election involvement (tau-c=-.14 and T=-1.84). 22 There was no difference in interest group campaign involvement associated withthe level of urbanization in Ontario or Quebec.Although the direction of the relationship for all parties was in the expecteddirection, the association between level of urbanization and interest group campaigninvolvement was not significantly associated with the different political parties. TheProgressive Conservative, New Democrat, Reform and CHP associations had little differencein the occurrence of interest group campaign activity that was related to level of22While the significance level was insufficient for either the Atlantic or Western regions alone, when thetwo regions are combined, the association between level of urbanization and interest group campaigninvolvement is significant and of approximately the same strength which seems to indicate that the significancelevel was insufficient due to a the small number of cases (tau-c=-.15 and T=-2.24).39urbanization. The Liberals, however, had interest group activity associated with more urbanridings but, again, the significance level was insufficient (tau-c=-.11 and T=-1.22).In summary, interest group campaign activity was not clearly associated with a higherlevel of urbanization. While the results indicated a tendency toward more interest groupinvolvement in urban constituencies, the relationship was not strong or significant. Theexpected direction was found but a higher level of urbanization was not unambiguouslyassociated more interest group campaign involvement.40Chapter Five: Constituency Association CharacteristicsCarty's (1991) volume for the Royal Commission on Electoral Reform and PartyFinancing explored the range of variation among constituency associations on a number ofattributes. He found that the membership size, the type of organization, and the financialresources of constituency associations varied greatly. However, the basis upon whichinterest groups target specific of ridings when they choose to become involved in electoralcontests is largely an unknown in electoral politics.An important contribution of the present study to the understanding of interest groupelection activity centres on the question of whether or not any of the characteristics of localassociations themselves are identified with interest group campaign involvement. The nullhypothesis is that there is no relationship between local associations' characteristics and theinvolvement of interest groups, or that each interest group's decision to be involved in a localcampaign is based on factors other than the characteristics of the association. Theconstituency association characteristics which were considered to be potentially associatedwith different levels of interest group involvement across all parties and regions includedassociation competitiveness, campaign type, and financial indicators.Local CompetitivenessThe competitiveness of one association in relation to the others in a riding is aplausible association characteristic that might disproportionately attract interest groups. Onereason for interest group campaign involvement might be to influence the outcome of theelectoral contest. In this case, interest groups might employ a strategy of supporting thecandidate with the best chance of winning, hoping that when issues related to the interest41group's area(s) of interest are on the political agenda, the local representative will be mindfulof the interest group's role in their election. In a very competitive riding, where the supportof an interest group alters the vote distribution enough to change the order of the first andsecond place candidates, interest group involvement could be very significant. 23 Thisapproach assumes that interest groups perceive the outcome of an election as its centralpurpose in campaign involvement. In addition, an interest group would have to identifywhich associations were competitive to ensure that they maximized their impact byconcentrating their support or opposition in constituencies that would make a difference tothe electoral outcome. If interest groups meet these two conditions (seeking to influence theoutcome of a local contest identified as very competitive), then it would be expected thattheir involvement would be disproportionately found in the most competitive contests.To explore this question, I created a variable which orders respondents according tothe margin between the winning candidate and the responding association's candidate. Thedifference between these two proportions of the total vote was used as an indication of thecompetitiveness of the local contest for the respondent's association in 1988. In the case ofrespondents from winning constituency associations, the second place candidate's proportionof the vote was subtracted from the winner's to give the competitiveness margin. A lowresult reflected a small difference or a close race and, therefore, a more competitive riding(see Appendix 2 for the derivation of the variable "margin2").23Another interesting but intestable strategy would be for an interest group to oppose a front-runningcandidate in a close race by bolstering a third place or worse campaign, drawing away enough support from onecandidate to ensure the election of the other less offensive candidate.42Respondents were then divided into the following four categories: "very high"competition (10 percentage points difference or less), "high" competition (between 11 and 25percentage points difference), "moderate" competition (between 26 and 40 percentage pointsdifference), and "low" competition (over 40 percentage points difference between the winnerand the respondent's candidate). These categories were chosen so that approximately onequarter of the respondents would be in each category.' The data reveals that there was anassociation between the competitiveness of the local association and the occurrence of interestgroup campaign activity. A higher level of interest group campaign activity was associatedwith more competitive contests (Table 5.1).Earlier in this chapter, the geographic variables of province and region were discussedin the context of interest group campaign activity. The Atlantic provinces were groupedtogether and the Western provinces and territories were grouped together. With respect tocompetitiveness, a regional analysis was less appropriate because there was a significantprovincial variation within the regions. For example, unlike other Atlantic provinces, NovaScotia had over half of its responding constituencies (53%) in the most competitive categoryof ten or fewer percentage points difference between the winner and the respondent. InWestern Canada, Alberta stood out from the other western provinces and territories becausevery few of its local contests were determined to be "highly" or "very highly" competitive.In fact, only nine percent of the local contests had the winner and respondent within tenpercentage points of each other. The other provinces in the Western Canadian region were241 tested the competitiveness level of the group of constituency associations that responded to the CPAsurvey to ensure that is was similar to the whole population of constituency associations. Each of the first,second, third, fourth and fifth place margins of competitiveness were checked and an almost identicaldistribution resulted in each case.43Table 5.1Constituency Competitiveness and Interest Group Activity(horizontal percent)Percentage PointsDifference betweenWinner and RespondentInterest Groups activein local campaignNo Interest Groupcampaign activity reported10 or less 32 2211 to 25 26 2626 to 40 28 29Over 41 15 24(tau-c=.15 and T=3.11)44much more evenly distributed among the levels of competitiveness. There was no consistentdirection to the provincial variations (tau-c=.02 and T=.62), but if each province wasconsidered an unordered nominal variable, there was a significant relationship between theprovinces and the competitiveness of their constituency associations (V=.24 and p< .01).There were differences among the five parties considered in the CPA study in theirinterest group campaign involvement. Recall that the Conservative and the New Democraticassociations had a higher level of interest group involvement, followed by the Liberals, CHPand the Reform Party. In the competitive environment of an election, the parties were, ofcourse, unequally distributed among the different categories of competitiveness. Thequestion for this study is whether or not interest group activity occurred evenly across thesedifferent levels of competitiveness. Among Conservative and Liberal campaigns, interestgroup activity occurred at equal rates among all campaigns whether or not they were "veryhighly" competition or "low" competition (Conservative riding associations: tau-c =.07 andT=.78; Liberal riding associations: tau-c =.04 and T=.42). In NDP riding associations, onthe other hand, interest groups were concentrated in the most competitive campaigns but therelationship was not significant (tau-c = .16 and T =1.63).Neither the Reform Party nor the CHP had riding associations that were in the mostcompetitive category of less than ten percentage points from the winner. The CHP had all ofits respondents in the two lowest categories so it was not meaningful to make anycomparisons between the competitiveness of these two parties and interest groupinvolvement.45A less precise measure of local association competitiveness involves consideringwinning and losing candidates as distinct groups. Because of the number of candidatescontesting most of the seats in the 1988 federal election, a majority of the candidates lost.The election results from all constituencies in conjunction with the CPA study data revealedthat seventy-three percent of respondents were from constituency associations where thecandidate had lost, and twenty-seven percent were from constituency associations where thecandidate won, a result very similar to the entire population of election candidates. Interestgroups promoting or opposing an issue in local campaigns were not associated with eitherwinners or losers (V =.06 and p > .05).In conclusion, there was a positive association between interest group campaigninvolvement and more competitive local campaigns. The CPA data set does provideevidence to support the theory that interest groups are concerned with the outcome ofelectoral contests and do seek the more competitive ridings where their influence on theoutcome would be maximized. The relationship is weak, however, which may be attributedto the difficulty in the task of predicting which local contests will be the most competitive.In addition, it is likely that the competitiveness of the riding is only one of many factorsconsidered by interest groups when they make the decision to become involved in electoralpolitics.Campaign TypeThe CPA study provides a basic point of comparison for the different types ofcampaigns that local riding associations staged in the 1988 federal election. Because all localassociations have many of the same tasks to perform during an election, the similarities46among all association campaigns are strong regardless of party or region. Campaigncharacteristics warrant consideration with regard to interest group involvement because it isat this time in the electoral cycle that the constituency associations are reporting interestgroup activity. The CPA study provides a measure of the involvement of single-issueinterest groups at this particular time, during an election campaign. Therefore, the variablesthat distinguish campaign types may provide clues to the rationale for interest groupinvolvement.Carty (1991) found campaign patterns related to the timing of the campaign team'sestablishment, the control of the campaign, the campaign manager's appointment, andwhether or not the campaign manager was a local party member. The differences he foundwere associated with the party affiliation of the constituency association and wereincorporated into his cadre/mass party thesis. He suggested that while many of theresponsibilities are similar across all local associations, the Liberals and Conservatives stagea different style of campaign which he described as candidate-centred. The NDP, on theother hand, is viewed as having a more bureaucratic campaign style where the ongoingorganization of the party plays a larger role in many aspects of the electoral cycle,particularly election campaigns.It could be argued that an interest group's choice to become involved in a particularlocal contests was in part based on the type of campaign that the local riding associationstaged. In such a case, it would be expected that a candidate-centred campaign wouldfacilitate interest group influence because a more autonomous candidate would be more ableto accommodate or incorporate the policy demands of a local interest group during an47election. This reasoning suggests that interest groups would be found disproportionately inthe candidate-centred campaigns. The more bureaucratic campaigns would have less policyflexibility and less autonomy to respond to interest group challenges based on policy issues.If interest group campaign activities were not welcomed by the riding association, the morebureaucratic campaign would be more difficult for interest groups to penetrate and less likelyto give policy concessions.Respondents were asked who had control of the direction of the election campaign,whether the riding association executive maintained control, the control was delegated to thecampaign team, or a balance was sought between the two groups. Very few associations(8%) left complete control of the campaign with the association executive and manyassociations indicated that a balance of control was the best description of the campaign'sorganization (40%). About half of the riding associations (52%) completely delegated thedirection of the local campaign to the election campaign team. Complete delegation ofcampaign control would indicate a more candidate-centred campaign and the retention ofcontrol by the local executive would indicate a less candidate-centred campaign, therefore, itwould be expected that interest groups would be more likely to be active in the campaignswhere control was delegated to the campaign team and less likely where the executiveretained control.The CPA data did show that there was a relationship between interest group campaigninvolvement and campaign control (see Table 5.2). Where control of the campaign wasdelegated completely to the campaign team, interest groups were more likely to be involvedin the campaign. Interest groups were slightly less likely when a balance of control between48Table 5.2Campaign Control and Interest Group Activity(horizontal percent)Interest GroupActivityNo InterestGroup ActivityCampaign Teamin complete control 57 47Balance 38 43Association Executiveretains control5 11(V =.13 and p< .05)the campaign team and the association executive was pursued. In the most bureaucraticcampaigns, where control of the campaign was retained by the association executive, interestgroup activity was less likely.Carty (1991) found that Quebec had more associations that maintained executivecontrol of the local campaign (15%), almost twice the proportion of the total sample (8%).In keeping with the expectation that interest groups would be less likely to be involved inthese types of campaigns, ninety percent of these more bureaucratic associations did notreport interest group involvement. The scenarios of complete campaign team controldelegation showed the same pattern as the other regions of more interest group campaigninvolvement in candidate-centred campaigns (V =.22 and p= .17). Of all of the parties, onlythe Conservative associations had a significant relationship between campaign control andinterest group involvement which showed a similar but more intense pattern as the overallresults (V = .22 and p < .05).Some associations organized the local campaign teams before the association hasnominated a candidate while others waited until the candidate was chosen before launchingthe election organization. Carty (1991) found that in keeping with the cadre/mass partyanalysis, the Liberals and Conservatives were more likely to wait to establish their campaignteams, presumably to allow candidates more autonomy in choosing the campaign teammembers. Whereas the NDP associations, viewed as having many mass party characteristics,were more likely to establish their campaign teams prior to the nomination of the localcandidate because of the stronger internal organizational dynamics which characterize thebureaucratic style and which mute the autonomy of individual candidates. It would be more50likely for interest groups to be involved in candidate-centred campaigns which allowindividual candidates to consider their cause.The data, however, do not show any association between interest group campaignactivity and the timing of the campaign team organization. Interest groups were equallylikely to be active in campaigns where the campaign team was set up before or after thenomination of the candidate (V =.01 and p > .05). 25Another indicator that Carty found to support the cadre/mass party idea concerned theappointment of the campaign manager. The candidate-centred campaigns of the cadre partieswere more likely to have the candidate appoint the campaign manager which Cartyinterpreted as a reflection of the autonomy of the candidate. With regard to interest groups,however, there was no relationship between interest group involvement and whether thecandidate, constituency association executive or any other group made the campaignmanager's appointment (V = .10 and p > .05). The most probable explanation for thisoutcome is that the campaign manager's appointment reflects an internal association decisionwhich would not be generally known or viewed as significant by interest groups.Carty suggests that non-local campaign managers are brought in to manageconstituency campaigns when the local race is very close and the expertise of a professionalcampaign manager may make the difference between winning and losing, or when theassociation is so desperate for help that there are no local party members willing to accept25There was no relationship between having a paid campaign manager in the 1988 federal election (V=.07and p > .05). This did not change for party or region. Carty found that NDP had more paid managers but thiswas not related to interest group involvement (V = .01 and p > .05). The number of paid staff was also not afactor.51the job. For all respondents, there was no relationship between interest group involvementand whether or not the campaign manager was a local party member.' To ensure that thetwo scenarios for bringing in a non-local campaign managers were not cancelling each other,I controlled for winners and losers. Interest group involvement was not at all related towhether the campaign manager was a local party member in losing associations (V =.00 andp > .05). Although the significance level was inadequate, winning campaigns with non-localcampaign managers were more likely than other winning campaigns to have interest groupcampaign involvement (V = .13 and p = .13).Of the factors which Carty found to be important in distinguishing one type ofcampaign from another, only campaign control seemed to be significantly related to interestgroup involvement. Alone this result provides only minimal support for the idea that thecampaign type was an important consideration for interest groups in their decision to becomeinvolved in local constituency campaigns.Financial IndicatorsThe CPA data set indicated a clear disparity in financial resources between Canadianpolitical parties. For example, the Reform Party and the CHP constituency associations didnot have the same level of resources available to them in the 1988 election campaigns as thethree older parties did. In addition, regional variation in financial resources within partieswas also apparent, particularly in the NDP.26In the Progressive Conservative riding associations, interest groups were more likely to be involved in thecampaigns run by local members than those run by outsiders (V = .17 and p< .05). None of the other partieshad any significant association. Ontario had a weak association between interest group involvement and localcampaign managers where interest group activity was also more likely if the local campaign was run by a localparty member (V= .16 and p< .05). No other regions had any significant association.52With regard to interest group campaign involvement in the 1988 federal election, aninteresting question is whether or not interest groups were disproportionately involved inwealthy local campaigns. One could hypothesize that interest groups contribute to the wealthof a local campaign by providing a significant portion of election financing, as in Americanelections. Campaign funding might be viewed by interest groups as an important avenue ofinfluence. In such a case, I expect that there would be a positive association betweenadequate, or even surplus, campaign funds and interest group local campaign involvement,and a negative relationship with campaign deficits.Constituency association presidents were asked in the CPA study whether or not theirassociation had a campaign surplus in the 1988 federal election. The expected relationshipwas found: there was a positive association between a campaign surplus and interest groupactivity in the local campaign (tau-b = .13 and T=2.78).Another way I tested the expected relationship was to explore the official electionspending reported by each local association to the Chief Electoral Officer.' Thisconstituency-specific information was merged with the CPA survey data to further investigatecampaign financing in relation to interest group involvement.In order to make a comparison between the riding association's ability to finance itslocal campaign and interest group activity, I added together the two sources of funds reportedto the Chief Electoral Officer, local contributions and the candidate's personal expenses.Official election expenses were then subtracted, and this new total considered the campaign27This data set included the spending limit for the constituency, the number of contributors to the localcampaign, the total amount of contributions, election expenses, the percent of the limit spent, the candidate'spersonal expenses, and the amount of any reimbursed election expenses.53surplus. While this assessment does not take into account the funds available to a localconstituencies from other sources, such as fundraising prior to the campaign, it does indicatewhether or not the amount of money raised during the campaign from contributions wasadequate to cover campaign costs.On this basis, more than one third (38%) of the respondents had a campaign deficit.Fifty-three percent of this deficit group reported interest group involvement which was verysimilar to the overall rate of fifty-one percent. The constituency associations that had acampaign surplus were further divided into two groups, those with a "modest" surplus($3500 or less), and those with a "large" surplus (over $3500). The group with the"modest" campaign surplus was slightly less likely than the deficit group to have interestgroups active, but this was not a significant difference. The group with a "large" campaignsurplus was as likely as the deficit group to have interest groups active (54%). Based on thisindicator of campaign financing, there is no evidence to suggest that interest groups aredisproportionately associated with wealthy local campaigns (tau-c=.00 and T=.02).Instead of grouping the data into deficit and surplus contributions in relation toelection expenses, the "raw" calculated campaign surplus can also be used to show a similarresult. The mean campaign surplus is almost identical with or without interest groupinvolvement ($3397 with interest groups active, and $3315 with no interest group activityreported). It is clear that a surplus of campaign contributions over election expenses isindependent of interest group involvement. Interest group are not disproportionatelychoosing to involve themselves in local campaigns on the basis of campaign-generatedmoney.54This does not provide a straightforward answer to the question of whether or notinterest groups are using campaign contributions as a significant method of influencingpolitical parties. I suggest that direct campaign funding of election candidates by interestgroups is generally less likely in Canada because election spending by political parties isregulated and limited by the Canada Elections Act. While fundraising is unlimited, spendingis limited and based on the number of electors in each riding. It is important to be mindfulthat a deficit campaign may resemble a surplus campaign if both have spent an equal amount,as a percent of the allowable limit. During a campaign, voters do not know the source oflocal financing (campaign contributions, pre-campaign fundraising, or loans from the nationalparty), but it is possible for them to discern which candidates and campaigns are spendingsignificantly more or less than others.The amount of the allowable limit spent by a local constituency association is a goodindicator of the profile of the local campaign because the constituency-specific limit is thesame for all parties in the riding. While spending close to the limit does not guaranteesuccess in the campaign, the limit does set the parameters for the local competition byproviding a general spending boundary.' Spending a very small portion of the allowablelimit is associated with receiving fewer votes on election day. The most competitivecampaigns, where the winner and his/her opponent(s) are less than ten percentage pointsapart, are strongly associated with campaign spending very close to the legal limit (tau-c = .47 and T=16.03).28If all of the competitors have access to financing that equals the constituency limit, the influence ofmoney is substantially decreased and competition is based on other factors.55Assuming that interest groups are employing a strategy of influencing the outcome oflocal election contests in pursuing campaign involvement, I expect that campaign spendingclose to the allowable limit would be disproportionately associated with interest groupinvolvement while low levels of campaign spending would have the opposite association.The expected association between campaign spending and interest group involvementwas found: interest groups were positively associated with local campaigns where theassociation spent a higher percent of the allowable limit. For all associations the meanpercent of limit spent was sixty-four percent. Associations which reported interest groupactivity spent an average of sixty-eight percent of the limit but those associations which didnot report interest group activity spent only sixty percent (Correlation=.12). I collapsed theraw percentages into four classes: 40% or less of the allowable limit, 41% to 70%, 71% to90%, and over 90% of the allowable limit. In the cross-tabulation table of the above classesand interest group campaign involvement, associations spending a smaller percent of the limitwere associated with less interest group involvement, and associations spending a greaterpercent of the limit were associated with more interest group involvement (tau-c=-.13 andT=-2.62).To further explore this relationship, the spending limit was tested by region, whichindicated some regional association (Correlation=.18). Because spending was uneven acrossthe regions, it was important to ensure that the relationship between the percent of the limitspent and interest group involvement could not be explained as a regional effect (Figure 5.1).In Atlantic Canada, the relationship between the percent of the limit spent and interestgroup activity was much stronger than the overall relationship (tau-c=-.38 and T=-2.61).56Figure 5.1Election Spending Limit and Interest Group ActivityInterest Group ActivityE3Low tte Moderate EN High 1111 Very HighOnly twenty-seven percent of the local campaigns in the lowest class (40% or less of theallowable limit) had interest groups involved in their campaigns. Almost three quarters(73%) of the campaign associations that spent as much as the limit allowed (over ninetypercent of the limit) reported interest group activity. In Atlantic Canada, the percent of thelimit a constituency association spent was strongly associated with interest group localinvolvement. Riding associations which reported interest group activity in their campaignsspent, on average, seventy-six percent of the allowable limit. The average spending wasalmost twenty percentage points lower in constituency associations that did not report interestgroup activity (57%).By contrast, there was no association between the amount of constituency associationspending and interest group activity in Quebec (tau-c =.02 and T=.16) or Ontario (tau-c =-.07 and T=-.81). The mean percent of the spending limit for both Ontario and Quebec didnot vary significantly on the basis of interest group involvement.In Western Canada there was a positive relationship between the percent of the limit aconstituency association spent and interest group campaign activity (tau-c=-.24 and T=-3.18). The constituency associations spending forty percent or less of the allowable limit hadinterest groups involved in forty-one percent of their campaigns; the campaigns closest to thelimit, those spending over ninety percent of the allowable limit, had interest groups involvedin seventy percent of their campaigns. The mean of the ungrouped data shows a similarrelationship where interest group activity and a higher percent of the allowable spending limitwere positively and significantly related.58The percent of the constituency spending limit used has a strong correlation with thepolitical party affiliation of the riding association (Correlation =.63). It is clear from thedata that the spending of the parties corresponds to the number of seats that each won in theelection. The Progressive Conservative associations spent an average of eighty-six percent ofthe constituency spending limit. Liberal and New Democratic associations had a mean ofsixty-nine and fifty-six percent of the spending limit. The two newer parties spentsignificantly less; the Reform Party and the CHP spent an average of twenty-eight and thirty-one percent, respectively, of the constituency spending limit.The differences between the parties are much more significant than any differenceswithin parties related to interest group involvement in the local campaigns. There was nosignificant variation in any of the parties spending associated with interest groupinvolvement.'Association MembershipOne suggested understanding of interest group - political party interaction is thatinterest groups use political parties, and local associations in particular, as vehicles to furthertheir policy objectives. Instead of, or in addition to, independent electoral action, an interestgroup could influence the local association(s) of existing parties either through regularlobbying techniques focused on the riding association, or through aggressive participation inthe internal activity of the riding association. The second approach, sometimes described asinfiltration or "take-over," would be manifest in greater local party membership where29For the five parties the associations were: PC associations (tau-c=-.02 and T=-.19), Liberal associations(tau-c=.02 and T=.19), NDP associations (tau-c=-.13 and T=-1.31), and Reform associations (tau-c= .08 andT=.70), and CHP associations (tau-c= .13 and T= .95).59interest groups were involved in the local campaign than the association where they were notactive.The 1988 association membership figures reveal that the mean membership for alllocal associations was seven hundred sixty individuals (Table 5.3). Local associations thathad interest groups active had a higher mean membership (848 individuals) than those thatdid not (664 individuals). 30Examining each party individually shows that the Conservative and NDP ridingassociation had higher memberships identified with interest group involvement, while theLiberal, Reform and CHP membership numbers were not significantly related to interestgroup campaign involvement. Conservative associations with interest group involvement intheir local campaign had a mean membership of one thousand forty-two individuals,compared to nine hundred sixty-four individuals in associations which reported no interestgroup campaign involvement. The mean membership in NDP associations was sevenhundred thirty-four individuals when interest group activity was reported, compared to amean of five hundred sixty-one individuals when interest groups were not involved in theriding campaign.30 An often made argument regarding "instant' members is that they are drawn disproportionately fromone ethnic group or another in a particular riding. To test this notion I considered the percentage of anassociation's membership described as members of a visible minority, however, the median percent was fivepercent with or without interest group campaign involvement. Nor was there any relationship betweenassociations that had made special efforts to involve members of visible minorities and interest group campaignactivity. A weak relationship between having a significant proportion of electors in the riding who weremembers of a visible minorities and interest group involvement did exist (V= .11 and p < .05). Yet when askedwhether the visible minorities in the constituency had become a factor in the association's politics, there wasagain no relationship at all with interest group activity (V=.01 and p > .01). These results do not support theidea that the recruitment of visible minorities to swell local party membership is a widespread tactic or even adiscernible trend associated with interest group campaign involvement.60Table 5.3Association Membership and Interest Group Activity(Number of individuals, mean)Interest Groups activein local campaignNo Interest Groupcampaign activity reportedAll 848 664PartyPC 1086 964Liberal 935 921NDP 734 560Reform 256 265CHP 218 181RegionAtlantic 1195 621Quebec 832 781Ontario 705 631West 957 634The regional breakdown of constituency associations reveals an even strongerassociation between membership and interest group involvement. The Atlantic region had amuch higher mean membership associated with interest groups involvement where the meanmembership was almost twice as high as the mean for associations that reported no interestgroup activity. The other regions had similar but more moderate relationships betweenhigher association memberships and interest group activity.The CPA study data provides evidence for the idea that interest group involvement inlocal campaigns is associated with higher membership figures. This connection indirectlysupports the idea that interest groups use the local associations of existing parties to furthertheir policy objectives. The Progressive Conservative and New Democratic associationswere most affected in this way, while the Liberal, Reform and CHP were much less affected.All of the regions, and Atlantic Canada in particular, showed the same positive relationshipbetween interest group involvement and local membership.In general, the characteristics of an association's campaign were not convincinglyrelated to interest group campaign involvement. The relative competitiveness of theassociation and the amount of campaign spending, on the other hand, were consistentlyrelated to interest group involvement. There were a significant relationships with the riding'scompetitiveness, a delegated campaign style, and non-local campaign managers in winningridings but the three other campaign characteristics (the timing of the campaign team'sorganization, and the campaign manager's appointment) were not related to interest groupinvolvement. The strongest relationship found with the financial indicators tested was thepercent of the allowable limit an association had spent, followed by whether or not the local62campaign reported a surplus. The number of members in a local association was higherwhen interest group activity was reported, suggesting some validity to the notion of interestgroup members becoming party members to influence party policy.63Chapter Six: Motivating IssuesTo this point I have considered all interest groups together in a search for any generalpatterns to their local campaign involvement. This chapter concerns only the constituencyassociations where interest groups were reported to be active in order to focus on the issuesthat motivated interest groups to become involved in local campaigns. Its central purpose isto determine whether or not groups motivated by different issues have different patterns oflocal campaign involvement.To many observers of the 1988 federal election, the Free Trade Agreement betweenCanada and the United States was the dominant theme of the election campaign (Campbelland Pal 1991, Jackson and Jackson 1990, Stone and Woodside 1991); the question of Canadapursuing even closer political, economic and cultural ties with the United States sparkedintense debate. Individuals and groups who felt compelled to participate in the electiondebate were not limited to the political parties, candidates, and political activists. Business,labour, and cultural groups also actively supported one side or the other of the issue, andindividual citizens and citizen groups arose, hoping to persuade fellow voters of their position(Campbell and Pal 1991).The exceptional policy focus of the 1988 federal election leads to the expectation thatmost of the interest groups actively supporting or opposing local candidates would have beenmotivated by the free trade issue.' Given the issue's prominence in the national campaign,31Local associations might also experience campaign involvement from business and labour interest groupsrelated to the free trade issue, but these groups were not "single-issue" groups. For example, the BusinessCouncil of British Columbia was heavily involved in the free trade debate but this issue is only one of manyongoing business related concerns of the Business Council of British Columbia.64it would not be surprising if local campaigns constituted a miniature version of the nationaldebate. The three major parties were clearly on one side or the other of the issue. As thegoverning party which was pursuing the free trade policy, the Progressive Conservative Partysupported the deal. The Liberals and the New Democrats were both opposed to the deal andMr. Turner, in particular, who was the leader of the official opposition at the time, foughtvigorously against free trade.In terms of interest group involvement, I expect that those motivated by the free tradeissue would participate in local contests at uneven rates among the parties with more interestgroups active in Conservative association campaigns. Because the Conservatives were theonly major party that supported free trade, they would have benefited from the support of thepro-free trade interest groups. In addition, the interest groups that were active in the electioncampaign did not have equal resources. The business groups which supported free trade, andindirectly the Conservative party, contributed a disproportionate amount of funds to thedebate. Hiebert (1991) estimated that 4.7 million dollars were spent by interest groupsduring the 1988 federal election campaign and 4.6 million of the total can be attributed topro- and anti-free trade groups. Hiebert pointed out the extremely lopsided nature of thespending such that four times as much was spent promoting free trade as opposing it (Hiebert1991). The major contributors to the total were pro-free trade business groups, followed byanti-free trade labour groups and then individuals and smaller citizen groups.Because the opposition to free trade was split between the Liberals and the NewDemocrats, even if the pro- and anti-free trade interest groups had the same persuasive effecton voting decisions, the anti-free trade campaign would have a weaker electoral effect65because of the division of votes between two parties. In areas where one of the oppositionparties was not competitive, I would expect the most viable opposition party to benefit fromthe anti-free trade interest groups' involvement in their campaigns.The few interest groups involved in Reform Party campaigns are expected to havebeen motivated by the free trade issue, because the party is particularly concerned abouteconomic issues. The CHP, on the other hand, would be less likely to have free tradeinterest groups either supporting or opposing its candidates, because its focus was moral andnot primarily economic.Responses to the CPA survey question concerning motivating issues do not delineatethe position of free trade groups, merely that they were motivated by the issue. With bothpro- and anti-free trade interest groups categorized together, I would not expect muchregional variation in the free trade interest group involvement because the issue wassignificant to the entire electorate.Another issue that I expect to have been prominent in local associations in the 1988federal election is abortion, because of abortion policy development immediately prior to theelection. The most important decisions affecting abortion policy during 1988 were theSupreme Court's decision to strike down the existing legislation in January, and theunresolved debate that had occurred in Parliament during the summer months. Of the majorparties, the Liberal Party seems to have been the party in which most of the debate on thisissue has taken place. This may be merely a media portrayal which typically focuses on the66Liberal encounters with pro-life groups.' The CHP candidates would also be expected tohave the support of pro-life interest groups and the opposition of pro-choice interest groups,however, it may be that even though this party held a pro-life position, interest groups wouldhave chosen to become involved in other parties more likely to win the constituency.On a regional basis, I would not expect significant variation in the abortion interestgroup campaign involvement because polling information on abortion opinions does notindicate a strong regional component (knish 1990).Free Trade with the United StatesFree Trade was clearly an important issue in the two hundred and ninety-five ridingsas well as in the national campaign. Unexpectedly, however, free trade was not the mostfrequently mentioned issue motivating interest groups involved in local campaigns (Figure6.1). Twenty-six percent of all the issues mentioned by constituency associations wererelated to free trade, about half as often as abortion. No other issue was mentioned by asignificant percent of constituency associations.This surprising result might be related to the role of individual Members ofParliament. The Canadian variant of parliamentary government includes strong partydiscipline on most matters before a legislature. The freedom of any individual Member ofParliament to vote according to constituency preferences is quite limited in most cases. In avote on a major international economic agreement, such as the free trade bill, party disciplinewas enforced. In fact, such an important economic bill would most certainly be considered a32For example, see Virginia Byfield's article, "Baby Paul and the Liberals," Western Report, October 21,1988 (3:41), 38.67Figure 6.1Issues Promoted or Opposed by Interest Groupsconfidence motion where failure to pass the legislation would result in the dissolution ofParliament and the election of a new government. The implication of party discipline is thatthe views of constituents and individual Members of Parliament are muted. Interest groupsmotivated by free trade would have an incentive to intervene in the electoral process, toensure that as many constituencies as possible were represented by Members of Parliamentfrom the party promoting their position.When considering the emphasis placed on the free trade issue in the nationalcampaign, and the fact that one quarter of the local campaigns experienced interest groupinvolvement related to the issue, free trade was unquestionably a significant issue in the 1988election. Stone and Woodside, in an article comparing the Canadian and American electionsof 1988, outlined the unusual set of circumstance in which an issue is likely to have animpact on voting behaviour (Stone and Woodside 1991, Kay g a. 1991).'The first condition is that the electorate must not be ambivalent about the issue butmust have some view or feeling about an issue. Stone and Woodside (1991) argued that thisneeds to be coupled with a significant subset of voters who hold their views intensely. Theissue of closer ties with the United States engaged most Canadians in some way, and asignificant proportion of voters and other interests such as business, labour, and culturalgroups were intensely motivated by this issue (Campbell and Pal 1991).The second condition is that there should be an imbalance in the distribution of viewssuch that one party benefits more than the other(s). As Hiebert (1991) noted, there was a33Gibbins, for example, claims that, "...policies or specific plans have never been central to winningelections" (Chase 1993a:6).69significant imbalance in the campaign advertising by interest groups, an indication of animbalance in interest group or sectoral views which would clearly benefit the ProgressiveConservatives.The third condition is that the electorate must be able to associate the various politicalparties with specific positions on the issue. In the 1988 election, the positions of the threemajor parties on free trade were clear to the electorate. The televised debate had beenparticularly effective, not only in making free trade the central issue of the campaign, but inclarifying the positions of the three leaders and their parties to the electorate. Stone andWoodside claimed, "...the [television] debates turned the 1988 election into a single-issuecampaign, and that issue was crucial in determining the result" (Stone and Woodside1991:217).The 1988 federal election was unusual because the interest groups involved in thecampaign were also clearly divided along sectoral lines, with business groups generallysupporting the deal and labour groups generally in opposition to it. Typically, interestgroups do not associate themselves with a particular party, but prefer the freedom of a non-partisan approach to policy (Tanguay and Kay 1991). Even though there can be some policyinfluence advantage in partisan identification, "...being identified with one political party is atwo-edged sword that many groups prefer to avoid. Fearing that such an attachment willclose their routes of access to other parties, many groups declare non-partisanship" (Jacksonand Jackson 1990:563; see also Boivin 1984).Free trade became the pivotal election issue in 1988. According to Stone andWoodside (1991), this was because of the occurrence of a specific set of unusual70circumstances. Indeed, the three conditions they outlined were met in the case of the freetrade issue. The notable level of interest group activity in local campaigns reported in theCPA study provides empirical evidence of the local importance of this issue as well.Another explanation of the interest group activity related to the free trade issue isfound in Pross' (1992, 1986) explanation of interest group activity. Pross argued that theregular activity of interest groups does not include electoral politics, and is based primarilyon influencing the people who have decision-making power because of the position they hold;as a result, most policy questions are resolved within the policy community, an explanatoryframework that he developed. This understanding of lobbying results in a very "Ottawacentred" view of interest group involvement in the policy process.It is difficult to transfer Pross' reasoning to an electoral contest where electioncandidates would be the targets of lobbying efforts. Pross did admit that there wereuncommon cases where an issue was resolved outside of these normal channels, which hedescribed as "beyond" the policy community. He identified three situations when a policyquestion is likely to be resolved outside of the policy community: when a policy is tooimportant to the nation; when at least two policy communities are in conflict over the issue;and when there is a serious and open disagreement within a policy community (Pross 1986).To employ Pross' analytical framework of interest group activity, I would classify the extra-policy community interest group activity associated with the Free Trade Agreement as one ofthe rare occasions when an issue is so important to the nation that the electorate exercises its"super-ordinary" authority to decide the issue. In addition, because of the comprehensivesectoral nature of free trade, many different policy communities would be affected by the71deal and were not in accordance with one another. With either Stone and Woodside's (1991)or Pross' (1992, 1986) explanation of the interest group activity related to free trade in the1988 election, the emphasis is on the unusual nature of the issue and the circumstances.There are some interesting differences between the political parties and the occurrenceof free trade interest groups. Figure 6.2 shows that free trade interest groups were active atapproximately the same level in the Conservative, New Democratic and Reform Partyconstituency associations (19%, 16% and 14% respectively). The Liberal Party, despite itsknown opposition to the deal had relatively fewer interest groups motivated by free tradeactively involved in its local associations (8%). This suggests a less than even division ofthe anti-free trade interest groups between the Liberal and NDP associations to the NDP'sbenefit. As expected, the CHP received almost no attention from free trade interest groups(2%). Only the Reform Party had free trade as the dominant issue promoted or opposed byinterest groups in their local campaigns (60%).Free trade was not the dominant issue in any of the regions and the issue motivatedinterest groups to become involved in local campaigns at fairly even levels across regions(Figure 6.3). Western Canada had the most interest group activity related to free trade,where eighteen percent of all local associations experienced the involvement of interestgroups motivated by free trade. Ontario and Atlantic Canada had slightly lower levels offree trade interest group activity (11 % and 13% respectively). Quebec had the lowest levelof free trade groups of all of the regions (8%)'34Over the three issue categories, the level of urbanization was very similar.72Figure 6.2Interest Group Issues and PartyE3 Other GroupsE23Free Trade Groupsrza Abortion GroupsAM No Interest Group Activity(Cramer's V=.28 and p<.01)100%80%60%40%20%0%PC Liberal NDP Reform CHP E3 Other GroupsFree Trade GroupsElAbortion GroupsNo Interest Group Activity(Cramer's V=.30 and p<.01)Atlantic Quebec Ontario^WestFigure 6.3Interest Group Issues and Region100%80%60%40%20%0%The competitiveness of the local association was not a factor related to interest groupsmotivated by the free trade issue. There were no significant variations in the frequency offree trade interest groups associated with the competitiveness of the riding association. 35The percentage of the spending limit used by a local constituency during the course of theelection campaign was not related to the free trade motivated interest groups except inAtlantic Canada and Quebec, where having spent a greater percent of the allowableconstituency limit was associated with free trade groups.The free trade issue motivated a significant number of interest groups to involvethemselves in local campaigns, however, the level (26%) was lower than expected given theissue's dominance in the national campaign. For any single issue to have such prominence isunusual and both Stone and Woodside's and Pross' analysis concur in the atypical centralityof one policy question in an election campaign.The free trade question was lower in both the Liberal and CHP riding associations.The Liberal result is worth noting because of the party's dynamic campaign based on thisissue. As a party with a more limited policy agenda, it was not surprising that the CHP hadalmost no free trade groups active in its riding association campaigns. Constituencyassociation characteristics did not provide any further nuances to the campaign involvementof free trade interest groups.35 Among Liberal ridings, there were slightly more free trade groups in the more competitive Liberalridings, but not the difference was not significant. Quebec had more free trade groups in uncompetitive ridingassociations, but again this was not significant.75AbortionAnother surprising result from the CPA survey, concerning interest groupinvolvement in local campaigns, was the dominance of the abortion issue at the local level(Figure 6.1). Over half (52%) of all of the issues mentioned related to abortion.'Tanguay and Kay, in their study of interest groups in the 1988 federal election, also foundthat, "...the most widespread interest group campaign during the 1988 federal election at theconstituency level," was the Campaign Life Coalition motivated by the abortion issue(Tanguay and Kay 1991:97). This suggests that much of the abortion interest group activityreported in the CPA survey was a result of Campaign Life Coalition activities. It is alsopossible that this pro-life activity incited groups on the other side of the debate to participatein local campaigns. In addition, there may have been interest groups concerned withabortion but which were not primarily abortion groups, such as women's group or labourgroups.The role of a Member of Parliament in an abortion bill is quite different from that ona free trade bill. Abortion resolutions were treated as a "free votes," in the summer of 1988where Members of Parliament were not bound by party discipline and were be "free" to voteaccording to the will of their constituents, their own conscience or other factors unrelated toparty discipline. In such a situation, it is expected that a local interest group would make its361 used only the first response, even though up to three responses were coded, because the inclusion of allthree responses exponentially complicated the analysis, and their inclusion did not change the overall ranking ofthe issues (abortion remained first, followed by free trade). When only the first responses are considered,abortion constituted fifty-two percent and free trade constituted twenty-six percent of the issues mentioned. Thischanged to forty-four percent for abortion and twenty-seven percent for free trade when all three responses werecompiled. The pro-abortion, anti-abortion and unspecified abortion positions were coded together under thegeneral category of "abortion" groups.76position on the issue known to all of the candidates, and particularly to the local winner,regardless of party affiliation. Pross agreed stating that, "... the independent views ofmembers can be truly influential only during 'free votes,' when the member is permitted tovote according to his or her conscience. On these occasions, the average backbencher issubmitted to a degree of persuasion comparable to that normally endured by CabinetMinisters" (1992:171)."Because of the importance of the views of the each election candidate, abortiongroups, and Campaign Life Coalition in particular, crossed party lines and both endorsed andopposed candidates from all three major parties (Kay et J. 1991). Such a strategy requiredthat the abortion interest group make known its preferred candidate in each of the twohundred ninety-five constituencies. Kay et al. (1991) suggested that an issue that crossedparty lines is made much more difficult in Canada by election campaigns that usuallyemphasize party leaders and the national campaign. The 1988 election campaign and thefocus on the Free Trade Agreement with the United States was a case in point. In addition,Kay et al. observed that many electors could not name the candidates in their local riding andin urban areas, "...voters may not necessarily even know the name of their constituency"(Kay et al. 1991:146). Against these information obstacles, it is not surprising that Kay etal. (1991) found that there was no significant impact on the electoral results of candidatesthat had their views on abortion made public in the context of the election campaign.Perhaps the goal of Campaign Life Coalition and other interest groups motivated byabortion was to make all candidates aware of their presence in the constituency, and to37Marsh (1984) found a Parliamentary focus among British abortion groups.77initiate the ground work for future influence when a vote on the issue would be called (Pross1986). Winning candidates, whether they were opposed or supported by abortion interestgroups, would then be aware of the possibility and extent of electoral retribution such groupscould initiate. Pross believed that the essence of interest group influence is, "...that pressuregroups speak for a significant part of the public, a part that can be mobilized into politicalaction should its interest not be reasonably accommodated in public policy" (Pross 1986:8).In making their claims to elected representatives, interest groups often use the implicit ordirect threat of electoral retaliation. Posing as spokespeople for many citizens, interestgroups claim that they and their adherents are ready to use the political process for thefurtherance of their cause, especially if they are not given what they believe to be anadequate hearing. There are many interest group tactics which would stress the importanceof a given issue to an elected representative but the most final and significant would be thevote (Pross, 1992).Kay pi pi. take a contrary position, suggesting that "part of the lore of participatorydemocracy" attributes the capacity for "electoral retribution" (Kay gt. al. 1991:142) to interestgroups as the basic threat behind their demands for policy influence. Instead, Kay et B1.suggested that the idea of electoral retaliation was accepted by both interest groups andelected representatives but was not clearly proven in electoral experience.When the pro-life and pro-choice groups gave their supporters the "call to action,"the result was minimal impact on the electoral outcome. Kay el al. determined that, "...theevidence presented here suggests that Members of Parliament will face little electoral falloutfrom their parliamentary votes on the abortion issue" (Kay et al. 1991:152). While this was78true in 1988, it seems premature to completely discount the potential electoral clout ofabortion groups. In a future election campaign which was not centred on such a significantpolicy issue as free trade, the impact of these groups might be more significant.Stone and Woodside's (1991) analysis of the role of issues in election campaigns canbe applied to the abortion issue to explain its failure to become a prominent issue in thenational campaign in 1988. Stone and Woodside's first condition, that electors not beambivalent about the issue and that a significant portion hold their views intensely, ispartially fulfilled by the abortion issue. Kay tt d. (1991) cite a series of opinion polls whichshow that a majority of Canadians do not hold strong views on the question of abortion.However, they did identify two significant and polarized sub-groups of the electorate who dohold their views on abortion intensely. The pro-life and pro-choice advocates are believed toconstitute thirteen percent and twenty-five percent of the electorate, respectively. Excludingthese two groups, sixty-two percent of Canadians neither support nor oppose abortion withoutreservation (Kay gl mil. 1991). Stone and Woodside's first condition is not met by theabortion issue because even while adequate activists are motivated by the issue, most electorsdo not hold a clearly defined position.The second condition, an opinion imbalance in favour of one party, is not present.The NDP might benefit if the imbalance of opinion favoured the pro-choice side, or at leastsplit the pro-life voters evenly between the Liberals and Conservatives. The third condition,that the electorate be able to associate the political parties with a particular position, is notmet either. While the NDP does have a clear and well known pro-choice position, the othertwo major parties are not as plain in their position on this issue.79Any electoral benefit that one party might gain from a particular abortion positionwould be weakened by the expectation that any future abortion bill would be introduced as avote of conscience. Because of this, unless the NDP were to form the government, there islittle incentive for voters to cast their ballot on the basis of this issue.The Campaign Life Coalition strategy to cross party lines and to endorse and opposecandidates within parties captures the complexity and individual nature of the abortionquestion. Such cross-party endorsements, however, do not make the voter's choiceimmediately clear, creating an information obstacle for voters in the polarized opinionsubgroups who would be inclined to take this issue into consideration in their votingdecision. In addition, the majority of electors who do not hold strong views on this issuewould be unlikely to seek out the necessary additional information on the abortion position ofcandidates when there is another important, clearly defined national election issue on whichthey do hold an opinion. Because of these barriers, it was unlikely for the abortion issue tohave been the pivotal issue in an elector's voting decision in 1988.This analysis does not suggest that abortion is unique in this regard; it is Free Tradewhich is unusual in that it does meet all of the conditions outlined by Stone and Woodside(1991). There are very few policy questions that would meet their three conditions, whichsuggests that in most elections, policy is unlikely to become the focus of the campaign.The political party affiliation of a constituency association was related to the abortionissue as a motivator for locally active interest groups (Figure 6.2). The clearest differencebetween the local associations of the different parties was the low level of abortion interestgroup activity in Reform Party associations. Only five percent of all Reform constituency80associations reported abortion groups either supporting or opposing their candidates. Theoverall percentage of Conservative, Liberal, NDP and CHP associations was quite similar(29%, 31 %, 25% and 29%, respectively), however when only the associations thatexperienced interest group involvement are considered, the CHP and the Liberal Party hadmore of the interest groups active in their local campaigns motivated by abortion. Seventy-one percent of interest groups activity in CHP campaigns and sixty-six percent of Liberalcampaigns was attributed to groups either promoting or opposing abortion.Quebec stands out from the other regions because only three percent of Quebec'sconstituency associations reported that abortion was the issue that prompted interest groupactivity in their local campaigns (Figure 6.3). This is coupled with the fact that of all theregions Quebec had the lowest overall campaign involvement by interest groups. Clearly,unlike the rest of Canada, abortion was not an issue that sparked much local campaigncontroversy in Quebec. All of the other regions had abortion as the most frequentlymentioned issue and at similar levels (Atlantic 26%, Ontario 37%, and Western Canada25% 38).The competitiveness variable indicated that the abortion groups were not associatedwith the more competitive ridings in any of the parties except the Liberal associations whereabortion interest groups were more likely to be involved in the most competitive ridings andless involved in the least competitive associations.38TH the province of Saskatchewan, forty-six percent of all its constituency associations reported interestgroup activity related to abortion. This constituted eighty percent of all interest group activity in the province.81Tanguay and Kay (1991) studied the interest groups directly and one of their findingsmay have some explanatory value for this present study. From their interviews they foundthat the wealthier interest groups were the most satisfied with the political system, and thatthey tended to be more discreet in their relations with Members of Parliament and thepolitical system in general than the less well-off groups. Conversely, interest groups withlimited resources tended to contact their Members of Parliament more frequently and werealso more likely to employ confrontational or unconventional political strategies. If Hiebert's(1991) estimates of the amount spent in the 1988 election campaign by interest groups reflectthe relative wealth of free trade and abortion groups, perhaps the tactics of the two types ofgroups parallel their findings. Abortion groups (the less well-off interest group) were moreconfrontational in their use of tactics such as confronting individual candidates on their policyissue whereas free trade groups (the wealthier groups) opted for a more general advertisingcampaign not directed at any particular candidate.The notion that the Liberal party has a more cantankerous relationship with abortiongroups, and particularly pro-life groups, was not supported by the CPA data. Consideringonly the Liberal constituency associations, there was no relationship between the abortiongroups and greater competition for local nominations, or leadership convention delegates.Nor was an association's experience with "instant members" more negative when abortiongroups were involved.In Atlantic Canada and Ontario, abortion groups more active in the most competitiveriding associations. This was not the case in Quebec and in Western Canadian ridingassociations where there was no association between the competitiveness of the local riding82association and abortion groups. There was no relationship between abortion groups and theamount of constituency spending limit.Because so much of the interest group campaign involvement in the 1988 federalelection was due to local abortion groups, some discussion of gender seemed worthwhile.There was, however, no difference in the reported participation of women in a constituencyassociation and interest group involvement. The median percent of women members wasreported to be fifty percent overall, and fifty percent with or without interest group activitywhich may reflect an unwillingness of the parties to report gender imbalances in theirmembership. Two other indicators of women' participation were whether or not theconstituency association had made any special efforts to involve women in the central affairsof the local party or to recruit women candidates. Again, there was no association witheither variable and the incidence of interest group involvement (involvement of women: tau-b = .06 and T=1.23, and recruiting women candidates: tau-b = .05 and T=1.20). Nor washaving nominated a women candidate associated with interest group activity (V= .04 andp> .05)." In general, the decision for interest groups to involve themselves in localelection campaigns did not have a gender dimension.Abortion was the interest group issue with which riding associations most frequentlyhad to contend in the 1988 federal election. One interpretation for the local dominance ofthe abortion issue is its nature as a "conscience" issue and its treatment in Parliament as afree vote. The local action by interest groups, however, had no significant impact on the39There were no further highlights when controlling for party or for region except in Western Canadawhere having a female candidate was associated with greater interest group involvement (V= .21 and p < .05).83electoral outcome. Stone and Woodside's conditions provide the most probable explanationof the failure of abortion to become a national issue. As expected the Liberals and CAP hada greater proportion of their interest group campaign involvement from abortion groups. Thevery low rate of abortion group involvement in Quebec local associations was not expectedgiven the high levels of abortion group involvement in the other regions."Other" IssuesAll of the "other" issues mentioned by constituency associations as motivating localinterest group activity are scattered across many policy areas (Figure 6.1). In total twenty-four "other" issues were mentioned by respondents. Given that the context of their activitieswas the 1988 federal election, it would be expected that the issues would be federal ornational responsibilities. While many of the issues mentioned have some overlap ofjurisdiction, only eleven of the issues mentioned are either predominantly or solely federalresponsibilities, two issues could be either federal or provincial issues, and three wereprovincial issues. In addition to the issues that are associated with some level ofgovernment, six of the remaining issues mentioned had broad over-arching policyimplications and two were not based on policy but were a consequence of the election contest"Other" issues were more frequently mentioned in Conservative ridings (14%). Themost probable explanation for this is the "government challenge" idea presented earlier which401 categorized abortion, free trade, languages, agriculture, women's issues, defense, capital punishment,native issues, the Goods and Services Tax, privatization, and scientific research as federal issues; transportationand the Meech Lake Accord as federal or provincial; and labour, literacy, and social welfare as provincialissues. The broad policy issues were the environment, business, Christian principles, seniors' issues, pro-smoking, and ethnic issues. The election related issues were groups promoting or opposing the candidatepersonally and another political party.84suggests that many groups use an election campaign as a forum for questioning governmentcandidates about their past record. NDP ridings had a similar level of other issues (15%)which may reflect the NDP's reputation for accommodating special interest groups such asworkers, women, homosexuals, and environmentalists. The data show that when theseissues' were mentioned it was only by NDP associations. "Other" issues comprised lessthan ten percent of the issues that motivated interest groups in Liberal, Reform and CHPriding associations."Other" issues were not important in any region or province except Quebec wheretwenty-two percent of all Quebec constituencies had "other" interest groups active. This wastwice as high as any other region and constituted two thirds of all interest groups activity inQuebec. However, like the other regions, no one issue dominated the "other" category.Within the parties, the competitiveness of the local association was not associated withinterest group activity but by region there were two slight nuances. In Atlantic Canada andOntario, "other" issues were pursued in the more competitive ridings, while in Quebec,interest group activity motivated by "other" issues was more frequent in the least competitiveridings. "Other" interest groups occurred evenly across different levels of competition inWestern Canada. There was no association between the amount of the election spendinglimit used and "other" issues.Because more than three quarters (78%) of the interest group campaign involvementin the 1988 federal election can be attributed to interest groups motivated by either abortiona ' The  typically" NDP issues mentioned were: women's issues, social welfare, defense issues, andopposition to the privatization of crown corporations.85or free trade, the nature of each of these two issues is central to the analysis of this chapter.Free trade was a national partisan issue which also affected local campaigns. Abortion, onthe other hand, was a non-partisan and almost individual issue which failed to be included inthe national campaign to any significant degree even though there was enormous localcampaign activity on this issue. "Other" issues came from many policy areas andgovernment jurisdictions and constituted the remaining twenty-two percent of local interestgroup activity. While more frequent in Quebec, no issue or type of issue dominated the"other" interest group activity.86Chapter Seven: ConclusionThe starting point of the present study was the apparent contradiction between thehigh level of interest group election involvement experienced by the ridings associations inthe 1988 federal election and the low level of interest group election involvement predictedby the interest group literature. While searching for trends in interest group electioninvolvement in the light of the many possible explanations, the two positions have becomemuch closer to reconciliation.The experience of the five political parties with interest group involvement wasmarkedly distinct. The Progressive Conservative and New Democratic Party localassociations had more interest groups promoting or opposing their candidates. The Liberalsand the CHP had less interest group involvement overall, but their interest group activity wasassociated primarily with the abortion issue. The Reform Party had the lowest level ofinterest group campaign involvement, most of it related to the free trade issue.The analysis of interest group election activity and political party provides significantsupport for the government challenge theory in which an election is viewed as an opportunityto question the most recent government's policy record. An election campaign furnishesinterest groups with an environment of easy access to elected representatives, a situationwhich not all interest groups enjoy on a regular basis.No support was found for an alliance between the NDP and labour interest groups andthis was most clear in the regions where the NDP electoral support was very low. This issurprising given the predominance of the free trade issue in the election campaign and labourinterest groups' clear anti-free trade position. The most probable explanation for this low87result is that NDP associations did not view labour groups as external to their associations oras "single-issue" interest groups.The most important geographic variable in interest group local campaign involvementwas region, where Quebec had not only a much lower level of interest group involvement butthe two dominant issues in the 1988 federal election, abortion and free trade, were issues ofmuch less significance in Quebec. Abortion interest group involvement was almostnonexistent and free trade interest group involvement in Quebec was the lowest of anyregion. Yet language was not a factor, and interest group local campaign involvement wasnot clearly associated with a higher level of urbanization.The characteristics of a local campaign, particularly the features Carty (1991) foundto be informative, were generally weakly associated with interest group involvement, if atall. A significant relationship was found between interest group campaign involvement andthe local association's competitiveness where interest groups more often promoted or opposedthe most competitive candidates. Of all of the campaign organization characteristics, onlythe control of the campaign was associated with interest groups. This result suggests that thecadre/mass party distinction elaborated by Carty (1991) is not central to whether or notinterest groups choose to become involved in riding campaigns.While there was a significant relationship between constituency associations thatreported a campaign surplus and interest group involvement, the strongest financial indicatorwas the percentage of the allowable spending limit that a constituency association used whichwas positively associated with interest group involvement. A similar effect was found withparty membership figures where interest group activity was associated with higher mean88membership levels. While some significant relationships did exist between interest groupcampaign involvement and constituency association characteristics, they were all secondaryor incidental associations because many closely related characteristics were not associatedwith interest group involvement. It is most probable that interest groups are not primarilychoosing local associations on the basis of the characteristics of their campaign.The issue that prompted most local campaign involvement by interest groups wasabortion. One important conclusions about the motivating issue were that Quebecassociations experienced a very different level of abortion related interest group activity thanthe other regions. From the standpoint of the political parties, the Reform Party associationswere different from other associations because they had an almost complete absence ofabortion as an issue in their local campaigns.Stone and Woodside's conditions for a successful election policy issue were not metby the abortion issue in 1988. Should abortion groups attempt to make the abortion issuecentral in a future election, I would predict that such a campaign would again fail to haveany significant electoral outcome because neither the distribution of opinion on abortion northe number of abortion activists is likely to change sufficiently to meet their first condition.As well, the Progressive Conservative and Liberal Parties have little incentive to present aclear position on the abortion issue because of the non-partisan way the issue would behandled if it were again debated in Parliament. Although abortion is unlikely to become thecentral issue in a national election campaign it may continue to have significance toconstituency campaigns.89Free trade, on the other hand was a significant national issue and the local campaigninvolvement across the nation reflected a "spill-over" from the national campaign. It isunlikely for the free trade issue to resurface as a central policy question in a future electionbecause the issue has been resolved. The relative lack of controversy concerning the NorthAmerican Free Trade Agreement supports this idea.The phenomena of scattered "other" issues will reappear in future elections becausethey represent marginal issues that are of intense significance to a local riding or to aninterest group within a riding. The fact that less than half of the "other" issues wereexclusively federal in a federal election supports the idea that such groups are seeking ageneral forum to present their ideas to the electorate. Quebec was the only region where"other" issues were important, and even there, no greater frequency of one issue overanother occurred.In conclusion, patterns to the interest group election involvement in the 1988 federalelection campaign were evident. The results of the present study suggest that interest groupswill continue to be involved in future election contests. It is likely that the governmentchallenge theory will provide a solid basis for the interpretation of future interest groupactivity, where the party in power will receive the most interest group campaign activity.Quebec's lower overall rate of interest group activity, perhaps due to its different andnegative view of interest group activity, will endure for the short-term. However, becauseinterest groups are central to most policy development, this attitude will most likely erodeover the long-term. Interest groups do not choose to involve themselves in local campaignsprimarily on the basis of the associations characteristics and it unlikely for this to change90significantly in the future. While the issues that motivate interest groups will change, it islikely that abortion will continue to be a local, albeit unsuccessful, issue.Elections are important forums for interest groups because it is in this context thatthey can seek increased legitimacy, find easier access to the political agenda, and gaineffective political participation for their local members. The desire to influence the outcomeof an election is present; however, it does not provide the only motivation, and perhaps noteven the primary one. The superficial treatment of the election involvement of interestgroups in the literature should be reexamined, and a perspective broader than an interestgroup's influence on public policy outcome adopted.91BibliographyArcher, Keith A. Political Choices and Electoral Consequences: A Study of OrganizedLabour and the New Democratic Party. Montreal and Kingston: McGill-Queen'sUniversity Press, 1990.Archer, Keith A. "Leadership Selection in the New Democratic Party" in Herman Bakvis,Editor, Canadian Political Parties: Leaders:, Candidates and Organization, Vol. 13,Royal Commission on Electoral Reform and Party Financing. Toronto: Dundurn andMinister of Supply and Services Canada, 1991.Bloom, David. "The dangerous narrow vision of single-interest protesters," CanadianSpeeches, May 1991 (5:3) 59-62.Boivin, Dominique. Le Lobbying ou le pouvoir des groupes de pression. Montreal:Editions du Meridien, 1984.Byfield, Virginia (a). "Assessing the Abortion Factor: Prolifers pursue major and selectivecandidate campaigns", Western Report, November 14, 1988 (3:43) 42-43.Byfield, Virginia (b). "Baby Paul and the Liberals", Western Report, October 31, 1988(3:41) 38.Campbell, Robert M. and Leslie A. Pal. The Real Worlds of Canadian Politics: Cases inProcess and Policy. Peterborough, Ontario: Broadview Press, 1991, pp. 1-70, 187-266.Carty, R.K. Canadian Political Parties in the Constituencies. Vol. 23, Royal Commission onElectoral Reform and Party Financing. Toronto: Dundurn and Minister of Supplyand Services Canada, 1991.Canadian Study of Parliament Group. Interest Groups and Parliament. Ottawa: CanadianStudy of Parliament Group, 1989.Chase, Stephen (a). "A paucity of policy and surfeit of support", Western Report, January25, 1993, (8:2) 6-7.Clark, Marc. "Outside agitators: Special-interest groups join the battle", MaClean's,November 7, 1988 (101:46), 19.Coleman, William D. "Interest groups and democracy in Canada", Canadian PublicAdministration, Winter 1987 (30:4), 609-622.92Coleman, William D. and Wyn P. Grant. "Regional Differentiation of Business InterestAssociations: A comparison of Canada and the United Kingdom", Canadian Journalof Political Science, March 1985 (18:1), 1-29.Collins, Anne. The Big Evasion: Abortion, the Issue That Won't Go Away. Toronto:Lester & Orpen Dennys Limited, 1985.Curtis, James and Lorne Tepperman. Understanding Canadian Society. Toronto: McGraw-Hill Ryerson Limited, 1988.Ehrmann, Henry W., ed. Interest Groups on Four Continents. Pittsburgh: University ofPittsburgh Press, 1958.Gagnon, Alain G. and A. Brian Tanguay, eds. Canadian Parties in Transition: Discourse,Organization and Representation. Scarborough: Nelson Canada, 1989.Ginsberg, Benjamin and Alan Stone, Editors. Do Elections Matter? Armonk, New York:M.E. Sharpe, Inc. 1991.Harrison, Robert, as told to Laird Stevens. And Justice for Some: Power and Patronage inOttawa. Montreal: Eden Press, 1988.Hayes, Michael T. Lobbyist and Legislators: A Theory of Political Markets. NewBrunswick, New Jersey: Rutgers University Press, 1981.Hiebert, Janet. "Interest Groups and Canadian Federal Elections", in F. Leslie Seidle, ed.Interest Groups and Elections in Canada. Vol. 2, Royal Commission on ElectoralReform and Party Financing. Toronto: Dundurn and Minister of Supply andServices Canada, 1991, 3-76.Jackson, Robert and Doreen Jackson. Politics in Canada: Culture, Institutions, Behaviourand Public Policy. Scarborough, Ontario: Prentice-Hall Canada Inc., 1990.Jenish, D'Arcy. "Abortion on Trial", MaClean's, July 31, 1989 (102:31), 14-19.Jenish, D'Arcy. "Support for women's rights", MaClean's, January 1, 1990 (103:1), 44-45.Kaihla, Paul. "Single-issue Liberals", MaClean's, April 23, 1990 (103:17), 15.Kay, Barry J., Lambert, Ronald D., Brown, Steven D. and James E. Curtis. "Single-IssueInterest Groups and the Canadian Electorate: The Case of Abortion in 1988", Journalof Canadian Studies, Summer 1991 (26:2) 142-155.93Kent, Tom. "Now is the time for all good persons to come to the aid of the parties", PolicyOptions Politiques, October 1988 (9:8), 3-5.Kernaghan and Siegel. Canadian Public Administration: A Text. Toronto: Methuen, 1987.Kome, Penny. Every Voice Counts: a guide to personal and political action. Ottawa:Canadian Council on the Status of Women, 1989.Landry, Mean. "Elements pour une theorie de la demande d'interventionsgouvernementales par les groupes d'interet," Quebec: University Laval, Departementde science politique, pp. 1-25.Laver, Ross. "The Debate about Life", MaClean's, July 31, 1989 (102:31), 20.Likely, Fraser. "The expanded role of interest groups in issues management," CanadianSpeeches, May 1988, 27-29.Luttbeg, Norman, ed. Public Opinion and Public Policy: Models of Political Linkage.Homewood, Illinois: The Dorsey Press, 1968 (119-205).Marsh, David. Pressure Politics: Interest Groups in Britain. London: Junction Books,1983.Muir, Richard. Modern Political Geography. London: the MacMillan Press Limited, 1975,pp. 106-110, 203-211.Perlin, George, ed. Party Democracy in Canada: The Politics of National PartyConventions. Scarborough, Ontario: Prentice-Hall Canada Inc., 1988.Presthus, Robert. Elites in the Policy Process. London: Cambridge University Press,1974.Pross, A. Paul. Group Politics and Public Policy, second edition. Toronto: OxfordUniversity Press, 1992.Pross, A. Paul. Group Politics and Public Policy. Toronto: Oxford University Press,1986.Reisman, David. Theories of Collective Action: Downs, Olson and Hirsch. London:MacMillan, 1990.Roscoe, Elizabeth. "Getting to government: how to lobby in Ottawa," Canadian Speeches,October 1990 (4:6) 74-76.94Sancton, Andrew. "Canada as a highly urbanized nation: new implications forgovernment", Canadian Public Administration, Autumn 1992 (35:3), 281-298.Seidle, F. Leslie, ed. Interest Groups and Elections in Canada. Vol. 2, Royal Commissionon Electoral Reform and Party Financing. Toronto: Dundurn and Minister of Supplyand Services Canada, 1991.Stanbury, William T. Business Government Relations in Canada. Scarborough: NelsonCanada, 1993.Stone, Alan and Kenneth Woodside. "Where Do Elections Matter More - Canada or theUnited States?", in Do Elections Matter? Benjamin Ginsberg and Alan Stone,Editors. Armonk, New York: M.E. Sharpe, Inc. 1991, pp. 207-221.Tanguay, A. Brian and Barry Kay. "Political Activity of Local Interest Groups" in F. LeslieSeidle, ed. Interest Groups and Elections in Canada. Vol. 2, Royal Commission onElectoral Reform and Party Financing. Toronto: Dundurn and Minister of Supplyand Services Canada, 1991, 77-115.Tedesco, Theresa. "A division on abortion", MaClean's, August 8, 1988 (101:33) 10.Thorburn, Hugh G. Interest Groups in the Canadian Federal System. Vol. 69, RoyalCommission on the Economic Union and Development Prospects for Canada.Toronto: University of Toronto Press, 1985.Waldman, Steven (a). "The Real Price of Buying Influence", Newsweek, January 13, 1992(119:2) 46-47.Wallace, Bruce (b). "Back to Square 1: the Senate rejects new abortion legislation",MaClean's, February 11, 1991(104:6), 15.95Appendix 1Selected Questions from the Constituency Party Association Survey1. Did any single-issue interest group play an active role in the local campaign in anattempt to promote or oppose your candidate?[ ] Yes [ ] NoIf so, what issue were they promoting or opposing?2. How would you describe the relationship between the local association executive andthe local campaign team?Association executive maintains controlControl completely delegated to the campaign teamA balance between these two positions3. Was your campaign organization:[ 1 In place before the candidate was nominated[ 1 Only set up after the candidate was nominated4. Did your local campaign team have nay individuals come in to help you form outsidethe constituency in 1988?[ ] Yes^[ ] Noif Yes,a) How many? ^b) Were these outsiders (please check as many as appropriate):[l Volunteer workers[ 1 Strategists/managers[l Paid organizers[ 1 Other ^ (please specify)965.^Who appointed the local campaign manager?Local constituency party executiveThe candidateProvincial or national party officialOther^ (please specify)6. Was the campaign manager a regular member of your local constituency association?[ ] Yes^[ ] No7. How many volunteers do you require to run an effective local campaign in yourconstituency?^8. How many volunteers did you have in 1988? ^9. Of the volunteers you had in your campaign team in 1988, what proportion would yousay:Are regular party membersAre regular financial contributors to your associationWorked in the 1984 general electionWere women10. If you had a different candidate would you expect to have:] Almost completely different group of volunteers] Somewhat different group of volunteers] Largely the same group of volunteers] The very same group of volunteers11. How would you rank in importance the following factors in attracting volunteers toyour campaign? (please rank 1 to 5)Local candidateOpposition candidateParty leaderParty policiesTraditional party loyalty12.^Did your candidate's campaign have a financial surplus at the end of the 1988 electioncampaign?[ ] Yes^] No9713. Could you tell us what the size of your constituency party membership was in thefollowing years? 1990^ 19881989 198714. Of your current membership what proportion are:Women%^Visible minorities15.^Has your constituency association made a special effort to involve any of thefollowing groups in an attempt to increase their participation in the central affairs ofthe local party? (please check as many as appropriate).WomenYouthVisible minoritiesAboriginal peoplesOther ^ (please specify)No special efforts made16. Do groups of visible minorities exist in significant numbers in your constituency?[ ] Yes^[ ] NoIf yes, have they become a factor in your constituency's politics?[ ] Yes^[ ] No17. What has been the experience with recruiting "instant members" in your associationfor nominations?Positive, it increases our membershipNegative, it is very disruptive of the associationIt has not happened in our constituency association9818. Please indicate the gender of those holding the following positions in yourconstituency association and whether they are a member of a visible minority or not.Male Female^VisibleMinority(please check if yes)Constituency Association President [ ] [ ] [ 1Constituency Association Treasurer [ ] [ ] [ ]Constituency Association Secretary [ ] [ ] [ 11988 election candidate [ ] [ 1 [ ]1988 election campaign manager [ 1 [ ] [ 11988 election candidate's agent [ 1 [ 1 [ 119. Have any special efforts been made to recruit women candidates by your association?[ ] Yes^[ ] No99Appendix 2Derivation of the Variable "Margin2"Definition of variables used to create margin2:win="dummy" variable which counts the winners.var2=Rank of PC candidate in 1988 election resultsvar3 =Rank of Liberal candidate in 1988 election resultsvar4=Rank of NDP candidate in 1988 election resultsvary =Rank of Reform candidate in 1988 election resultsvar6=Rank of CHP candidate in 1988 election resultsvar7=Number of votes received by PC candidatevar8=Number of votes received by Liberal candidatevar9=Number of votes received by NDP candidatevarl0=Number of votes received by Reform candidatevarl 1 =Number of votes received by CHP candidatevarl2=Total Valid vote in constituency (All votes - spoiled ballots)second="dummy" variable which counts the second place candidates.portion="dummy" variable which calculates the proportionv307=political party affiliation (1=PC, 2=Lib, 3=NDP, 4=Reform, 5 =CHP)."Margin2"compute win =0.if (var2 eq 1) win = var7/var12.if (var3 eq 1) win = var8/varl2.if (var4 eq 1) win = var9/varl2.if (var5 eq 1) win = var10/var12.recode win (0.00 =sysmis).compute second =0.if (var2 eq 2) second = var7/var12.if (var3 eq 2) second = var8/var12.if (var4 eq 2) second = var9/var12.if (var5 eq 2) second = varl0/var12.recode second (0=sysmis).compute portion=15.if (v307 eq 1) portion =var7/var12.if (v307 eq 2) portion =var8/var12.if (v307 eq 3) portion =var9/var12.if (v307 eq 4) portion =varlO/var12.if (v307 eq 5) portion = var11/12.recode portion (15.00 = sysmis).compute margin = win - portioncompute margin2 =margin.if (margin eq 0.0000) margin2=win - second.100Frequency of Margin2Value (Percentage Points Difference)^Frequency Percent^Valid%^Cum% 1. 10 or less 133 26 26 262. 11 to 25 135 26 27 533. 26 to 40 139 27 28 814. 41 and over 98 19 19 100Missing 17 1 Missing522 100 100101


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items