Hostname: page-component-7b7567c8c-z4zkx Total loading time: 0 Render date: 2026-11-18T07:18:52.517Z Has data issue: false hasContentIssue false

Telephone versus Online Survey Modes for Election Studies: Comparing Canadian Public Opinion and Vote Choice in the 2015 Federal Election

Published online by Cambridge University Press:  15 August 2017

Charles Breton*
Affiliation:
University of British Columbia
Fred Cutler*
Affiliation:
University of British Columbia
Sarah Lachance*
Affiliation:
University of British Columbia
Alex Mierke-Zatwarnicki*
Affiliation:
Harvard University
*
Department of Political Science, University of British Columbia, 1866 Main Mall, Vancouver BC, V6T1Z1, email: cbreton@mail.ubc.ca (corresponding author)
Department of Political Science, 1866 Main Mall, Vancouver BC, V6T1Z1, email: fred.cutler@ubc.ca (corresponding author)
Department of Political Science, 1866 Main Mall, Vancouver BC, V6T1Z1, email: s.lachance@alumni.ubc.ca
Department of Government, Harvard University, 1737 Cambridge St., Cambridge MA 02138, email: amierkezatwarnicki@g.harvard.edu
Rights & Permissions [Opens in a new window]

Abstract

Election studies must optimize on sample size, cost and data quality. The 2015 Canadian Election Study was the first CES to employ a full mixed-mode design, aiming to take advantage of the opportunities of each mode while preserving enough commonality to compare them. This paper examines the phone interviews conducted by ISR-York and the online questionnaires from panellists purchased from a sample provider. We compare data quality and representativeness. We conduct a comprehensive comparison of the distributions of responses across modes and a comparative analysis of inferences about voting. We find that the cost/power advantages of the online mode will likely make it the mode of choice for subsequent election studies.

Résumé

Les études électorales doivent optimiser la taille des échantillons, leur coût et la qualité des données. L’Étude électorale canadienne de 2015 a été la première ÉÉC qui a adopté un plan à mode de collecte mixte, visant à tirer parti des possibilités de chacun des modes tout en préservant suffisamment d’éléments communs pour permettre la comparaison. Cet article examine les interviews téléphoniques menées par l'Institut de recherche sociale (IRS) de l'Université York et les questionnaires des répondants en ligne achetés auprès d'un fournisseur d’échantillons. Nous comparons la qualité des données et la représentativité. Nous effectuons une comparaison complète des distributions des réponses selon les modes et une analyse comparative des inférences au sujet du vote. Nous constatons que du point de vue des avantages coûts-efficacité, il est probable que le mode en ligne représentera le mode de choix des études électorales ultérieures.

Information

Type
Research Article
Copyright
Copyright © Canadian Political Science Association (l'Association canadienne de science politique) and/et la Société québécoise de science politique 2017 
Figure 0

Table 1Survey Details

Figure 1

Table 2Comparison of Data Quality

Figure 2

Figure 1Comparing Modes with Benchmarks: Demographic and Politically Relevant Variables

Figure 1 shows point estimates and 95 per cent confidence intervals for each mode after each variable has been dichotomized. These point estimates are shown as dots and the corresponding validated benchmarks as vertical lines.
Figure 3

Figure 2Comparing Modes with Benchmarks: Vote, Turnout, and Vote Forecast

Figure 2 shows point estimates and 95 per cent confidence intervals for voting variables. These point estimates are shown as dots and the corresponding validated benchmarks as vertical lines.
Figure 4

Table 3Differences between Modes and Benchmarks

Figure 5

Figure 3Vote Dynamics

Figure 6

Figure 4Leader Dynamics

Figure 7

Figure 5Political Interest and Attention

This figure shows the mean value for each mode on questions measuring respondents’ interest and attention to politics. Values for “Interest in the Federal Election” and “Politics in General” are scales that go from 0 to 10 where 0 is “No interest at all” and 10 is “A great deal of interest” and have been rescaled from 0 to 1. For level of attention, all measures have three response options that have been rescaled to range from 0 to 1 (No attention, 0; A little, .5; A lot, 1).
Figure 8

Table 4Minutes per day reported news consumption

Figure 9

Figure 6Feeling Thermometers: Political Parties

This figure shows the mean responses to feeling thermometer where answers can range from 0 (Really dislike) to 100 (Really like). Facets are ordered based on the difference between the two modes. Bloc Québécois only includes Francophone respondents in Quebec.
Figure 10

Figure 7Feeling Thermometers: Groups

This figure shows the mean responses to each feeling thermometer where answers can range from 0 (Really dislike) to 100 (Really like). Facets are ordered based on the difference between the two modes. Feeling thermometers for Anglophones and Francophones only include respondents who are not “members” of that group.
Figure 11

Figure 8Immigration and Minorities

Figure 8 shows point estimates for each mode on questions about immigration and minorities. All answers have been dichotomized where 0 can be interpreted as meaning “No” and 1 as meaning “Yes”. The only exception is “Ban Face Covering Veil” which was already dichotomous (Yes/No).
Figure 12

Table 5Sample Selection or Social Desirability

Figure 13

Table 6Vote Choice

Figure 14

Table 7Effects “Missed” by Mode

Figure 15

Figure 9Power and Cost

The figure displays post-hoc power analysis for two campaign effects among Franchophones in Quebec in the 2011 electoral campaign. The total cost is calculated by multiplying the sample size by the cost per complete for each mode in 2015 CES.
Supplementary material: File

Breton supplementary material

Figures A1-A4 and Tables A1-A2

Download Breton supplementary material(File)
File 879 KB