Analysis of informal voting (House of Representatives 2016 Election)

Updated: 16 January 2019

Executive summary

Following each general election for the House of Representatives, the Australian Electoral Commission (AEC) undertakes an Informal Ballot Paper Study (IBPS) to analyse the levels and types of informal voting. Research based on the IBPS is fundamental to the AEC’s role in supporting electoral integrity by:

  • Informing education and information strategies to reduce informal voting, including through the provision of robust information at the polling place level.
  • Providing an evidence base for reforms to the electoral system, for example, by enabling analysis of:
    • the impact of Optional Preferential Voting, and
    • aligning savings provisions between the House of Representatives and the Senate.

The national informality rate (informal votes as a percentage of all votes cast) decreased from 5.91 per cent of all votes cast at the 2013 House of Representatives elections to 5.05 per cent at the 2016 House of Representatives elections. At the state and territory level, the highest informality rates for the 2016 House of Representatives elections were in the Northern Territory, New South Wales and Victoria, while the lowest informality rates were in the Australian Capital Territory, Tasmania and Western Australia.

Figure 1. Informality rates by state/territory, 2013–2016 House of Representatives elections

Title: Figure 1. Informality rates by state/territory, 2013–2016 House of Representatives elections - Description: Shows House of Representatives informality rates by state/territory for the 2013 and 2016 elections. The highest informality rates were in the NT, NSW and Victoria, while the lowest informality rates were in the ACT, WA and Tasmania.

(Australian Electoral Commission, 2013; 2016f)

Compared with the 2013 House of Representatives elections, informality rates decreased in every state and territory other than the Northern Territory.

The largest decreases were in New South Wales, Western Australia and the Australian Capital Territory.

For the first time since 2001, the ten Commonwealth electoral divisions with the highest levels of House of Representatives informality were not all located in Sydney. While eight of these divisions were in Sydney (Lindsay, Blaxland, Watson, Fowler, McMahon, Parramatta, Werriwa and Barton), the two other divisions were located in Victoria (Murray) and Queensland (Longman).

Major process improvements for the 2016 IBPS (including data based on scanned ballot paper images) have improved data quality and the range of analyses that can be conducted, but mean that caution should be used when comparing 2016 results against previous years.

For the first time since 2001, more than half of all informal ballots cast at the 2016 House of Representatives elections were assumed to be intentionally informal. However, among the ten divisions with the highest rates of informal voting, more than half of all informal ballots were assumed to be unintentionally informal, and assumed unintentional informality was a highly significant predictor of the total informality rate.

More than a quarter of all informal votes cast in 2016 had incomplete numbering, with more than half of these showing a number ‘1’ only. A further quarter of all informal ballots cast were totally blank, while about one in five were informal due to scribbles, slogans or other protest vote marks and one in six showed non-sequential numbering.

Table 1 below summarises the number of informal votes in each category, as well as the proportion each category represents of the total number of informal votes and of the total of all votes cast.

Table 1. Informal ballot papers by category, 2016 House of Representatives elections
Category Clear first preference
no.
No clear first preference
no.
Total Informality rate
%
no. %
Totally blank .. 179,243 179,243 24.9 1.26
Incomplete numbering 183,183 .. 183,183 25.4 1.28
Number ‘1’ only 105,093 .. 105,093 14.6 0.74
Other incomplete numbering 78,090 .. 78,090 10.8 0.55
Ticks and crosses 48,444 6,677 55,121 7.6 0.39
Other symbols 9,458 1,678 11,136 1.5 0.08
Non-sequential numbering 84,960 26,055 111,015 15.4 0.78
Scribbles, slogans and other protest vote marks .. 142,933 142,933 19.8 1.00
Illegible numbering 12,083 7,652 19,735 2.7 0.14
Voter identified 117 .. 117 0.0 0.00
Other 5,085 13,347 18,432 2.6 0.13
Total 343,330 377,585 720,915 100.0 5.05

(Australian Electoral Commission, 2018a).

There are many factors that appear to affect informal voting at the House of Representatives. Over the past few elections the Australian Electoral Commission (AEC) has investigated a variety of factors using both AEC and Australian Bureau of Statistics (ABS) data.

While the AEC does not usually conduct studies of by-election informality, the 45th Parliament has had an unusually high number of by-elections, providing the opportunity to examine some specific factors affecting informality. In order to examine the effect of confusion over the Senate voting system and the effect of not having a candidate from a major party, the AEC conducted studies on the Longman and Fremantle by-elections respectively.

These analyses have found that:

  • Higher levels of informality are likely to be associated with higher levels of social exclusion or relative disadvantage.
  • The number of candidates on the ballot paper was a significant predictor for both incomplete numbering (other than a number ‘1’ only) and non-sequential numbering where a clear first preference was evident. In other words, if there were more candidates on the ballot paper, voters were more likely to make a mistake or simply stop numbering before they had assigned preferences to all candidates
  • It appears that many voters may have been confused about the difference in voting requirements between the House of Representatives and Senate at the 2016 federal election, recording preferences ‘1’ to ‘6’ only (in line with the ‘above the line’ instructions for the Senate) on their House of Representatives ballot paper. These ballot papers are included in incomplete numbering (other than a number ‘1’ only) in Table 1 above.
    • This proposition is strongly supported by the outcome of the 2018 Longman by-election. The number of candidates was unchanged from the 2016 federal election, but in the absence of the Senate vote (and the accompanying AEC messaging and media attention), ‘1’ to ‘6’ votes fell substantially at the by-election.
  • Voter confusion about the differences between state and federal electoral systems may be contributing to some categories of informal ballots (particularly for House of Representatives ballots with incomplete numbering or where ticks and crosses have been used as the first preference).
  • As some informal votes are cast intentionally rather than representing an error on the part of the voter, voters’ attitudes to and opinions of the electoral system or politics in general will also likely influence informality.
    • This was supported by the 2018 Fremantle by-election, in which the most notable change from 2016 was the higher prevalence of ballot papers marked with scribbles, slogans, or other protest marks. Many of these ballot papers referenced the lack of a Liberal candidate or appeared to be an attempt to vote for a Liberal candidate.

Divisional Summaries