Secondary menu

Interest of Canadians in Internet Voting (2004, 2006, 2008 and 2011) – Research Note

Perception of Risk Associated With Internet Voting

Elections Canada added a question to its survey following the 2004 general election to examine elector perception of the risk associated with Internet voting. The question asked electors for their opinions on the potential for fraud or error associated with Internet voting. For the 2011 election, in order to check whether measuring only the risk aspect might introduce a negative bias, we asked respondents to indicate which of two statements came closest to their own views: "Voting on the Internet is risky" or "Voting on the Internet is safe." For this reason, the 2011 results cannot be compared with earlier results.

That said, Graph 5 shows that, from 2004 to 2011, about one out of every three respondents indicated that Internet voting was safe, regardless of the wording of the question.

Graph 5: Perception of Risk Associated With Internet Voting (2004, 2008 and 2011)*

Graph 5: Perception of Risk Associated With Internet Voting (2004, 2008 and 2011)
Text version of "Graph 5: Perception of Risk Associated With Internet Voting (2004, 2008 and 2011)"

*The question on risk associated with Internet voting was not asked in 2006.

Otherwise, we note that, in 2004 and 2008,Footnote 12 university graduates tend to disagree the most with the idea that Internet voting is too risky (46% and 44% respectively), while respondents with a high school diploma or less education tend to disagree the least (32% and 27% respectively). That said, as observed in Graph 6, the gap between the different levels of education has increased slightly in 2008.

The association between education and perception of risk is statistically significant in 2004 and 2008. Furthermore, the strength of this positive association increased slightly from 2004 to 2008. As such, the higher an elector's education level is, the higher the chance that this elector will disagree with the idea that Internet voting is too risky.

Graph 6: Education and Perception of Risk Associated With Internet Voting (2004 and 2008)
(% of respondents disagreeing with the statement that Internet voting is too risky)**

Graph 6: Education and Perception of Risk Associated With Internet Voting (2004 and 2008)
Text version of "Graph 6: Education and Perception of Risk Associated With Internet Voting (2004 and 2008)"

2004 (EC) 2008 (EC)
Χ2 = 18.13 (p < 0.001)
γ = 0.20
Χ2 = 26.86 (p < 0.001)
γ = 0.26

**Response categories were recoded using a dichotomous nominal scale (Agree/Disagree). In addition, the "Neither agree nor disagree", "Don't know" and "Refused" categories were excluded from the analysis.

For 2011, we notice that electors with a university degree are the most inclined to think that Internet voting was safe (43%). In contrast, we note that electors with a high school diploma or less education are the most inclined to believe that voting on the Internet is risky (72%), as observed in Graph 7.

These observations are confirmed by the presence of a statistically significant relationship between the education variable and perception of risk. However, we note that this positive association is weak. That said, in the last general election survey, the higher the individual's education level, the lower the chance that this individual will be likely to agree that voting on the Internet is risky.

Graph 7: Education and Perception of Risk Associated With Internet Voting (2011)Footnote 13
(% of respondents agreeing with one of the following statements: Voting on the Internet is safe / Voting on the Internet is risky)

Graph 7: Education and Perception of Risk Associated With Internet Voting (2011)
Text version of "Graph 7: Education and Perception of Risk Associated With Internet Voting (2011)"

2011 (CES)
Χ2 = 13.14 (p = 0.001)
γ = 0.21

Graph 8 shows that the gap in perception of risk among age groups is quite small in 2004, but not in 2008. In fact, the percentage obtained in 2008 reveals significant differences between the 18-to-34 (58%), 35-to-64 (45%) and 65+ (27%) age groups.

Confirming these data, we notice that the relationship between age group and perception of risk is not statistically significant (p = 0.327) for 2004. This result contrasts with the result for 2008, where a significant association of moderate strength and negative direction is observed. In other words, only for 2008 do we note that, the higher an individual's age group, the smaller the chances this individual is likely to disagree with the idea that Internet voting is too risky.

Graph 8: Perception of Risk by Age Group (2004 and 2008)
(% of respondents disagreeing with the statement that Internet voting is too risky)

Graph 8: Perception of Risk by Age Group (2004 and 2008)
Text version of "Graph 8: Perception of Risk by Age Group (2004 and 2008)"

2004 (EC) 2008 (CES)
Χ2 = 2.23 (p = 0.327)
γ = 0.02
Χ2 = 37.02 (p < 0.001)
γ = – 0.33

As indicated previously, the question on perception of risk was reworded in 2011. Graph 9 shows that there is a similarity in the level of agreement of respondents in the 18-to-34 age group (40%) and the 35-to-64 age group (41%) with the statement that voting on the Internet is safe. In contrast, we note that in the 65+ age group, nearly three out of every four persons consider Internet voting risky.

The association between age group and perception of risk is statistically significant for 2011. However, the strength of the association is relatively weak and negative.

Graph 9: Perception of Risk by Age Group (2011)
(% of respondents agreeing with one of the following statements: Voting on the Internet is safe / Voting on the Internet is risky)

Graph 9: Perception of Risk by Age Group (2011)
Text version of "Graph 9: Perception of Risk by Age Group (2011)"

2011 (CES)
Χ2 = 14.63 (p = 0.001)
γ = – 0.22

The results related to the 2008 general election show perception of risk by elector employment status. Graph 10 shows that nearly half of those in the labour market do not consider Internet voting too risky, compared with a little less than one third of retired respondents.

The association between employment status and perception of risk is statistically significant. Furthermore, the strength of that association is moderate.

Graph 10: Perception of Risk by Employment Status (2008)
(% of respondents disagreeing with the statement that Internet voting is risky)

Graph 10: Perception of Risk by Employment Status (2008)
Text version of "Graph 10: Perception of Risk by Employment Status (2008)"

2008 (CES)
Χ2 = 25.23 (p < 0.001)
Cramer's V = 0.15

For 2011, Graph 11 sets out the results for the reworded question on perception of risk by elector employment status. At first glance, we note that retired people (32%) are slightly less likely to consider Internet voting safe when compared to individuals who are employed (40%) or unemployed (38%).

However, the association between employment status and perception of risk is not statistically significant (p = 0.10). That is, the results indicate that perception of risk associated with Internet voting was not influenced by the employment status of Canadian electors at the time of the 2011 general election.

Graph 11: Perception of Risk by Employment Status (2011)
(% of respondents agreeing with one of the following statements: Voting on the Internet is safe / Voting on the Internet is risky)

Graph 11: Perception of Risk by Employment Status (2011)
Text version of "Graph 11: Perception of Risk by Employment Status (2011)"

2011 (CES)
Χ2 = 4.60 (p = 0.10)
Cramer's V = 0.07

Footnote 12 Even though the subject matter was very similar, we are of the view that the questions in 2004 and 2008 are not directly comparable with the question used for the 2011 federal general election.

Footnote 13 The response categories were recoded using a dichotomous nominal scale (Voting on the Internet is safe / Voting on the Internet is risky). In addition, the "Not sure" and "Refused" categories were excluded from the analysis.