Melbourne and Werribee each returned about 77–78 formal Lower House votes per 100 enrolled electors in the 2022 main election. Their final shares look alike. How they got there does not.

Why this matters: A single formal-share figure can hide two different questions for election administrators, local reporters and voters: where fewer Lower House ballots entered the count, and where more counted ballots were informal. I chose Melbourne and Werribee after comparing their near-equal formal shares with the very different components underneath. The distinction changes which local question a reader would investigate first. District totals cannot identify any individual’s reason for not having a formal vote recorded.

How to read this report: hover or tap marks for exact values and use the chart toolbar to zoom or reset. From Figure 3 onward, choose a district in the search box or click a point to locate the same district in three different views. Figure 4 can isolate the opening pair; Figure 5 filters by candidate count. Each plot also has a written reading and the full district table is below.

1. Begin with two districts

The same formal share can conceal very different paths through the count. Melbourne and Werribee are an illustrative pair: their formal shares are close, while their informal rates differ markedly. Every segment below uses enrolled electors, not votes cast, as its denominator. “No Lower House ballot counted” is the arithmetic remainder; it does not identify a person’s reason for not having a ballot recorded.

Accessible reading: Melbourne recorded 19.3% of enrolment as no counted Lower House ballot and 2.7% as informal; Werribee recorded 14.4% and 8.3%, respectively. The two formal shares differ by only 0.7 percentage points.

Source: Victorian Electoral Commission (VEC, n.d.-a), 2022 state election results, individual district recheck tables; author calculations.

2. Put 2022 on a longer timeline

The share of enrolled electors whose Lower House vote entered the formal count was lower in 2022 than in 2018, even after accounting for Narracan. The distance between these two lines is informal ballots as a share of enrolment. It is deliberately not the informal rate among ballots cast.

Accessible reading: in 2018, the estimated formal share was 84.9%; in the 87 districts voting in 2022 it was 83.3%. The 2022 count comprises 3,617,000 formal ballots from 4,343,959 electors enrolled in those districts.

Denominator note: VEC’s published statewide 2022 Lower House turnout is 87.13%: 3,828,791 ballots divided by 4,394,465 enrolments, including Narracan’s 50,506 electors although Narracan voted later (VEC, 2023b). For the 2022 point here and all district comparisons, I use the sum of enrolments in the 87 districts that actually held a Lower House election at the main event (4,343,959), giving 88.14% turnout and 83.27% formal share. For 1999–2018, I use the published VEC participation and informal percentages. Rounding of those historical source percentages slightly limits their precision. This is an adjusted comparison, not the VEC’s published statewide 2022 rate.

Sources: VEC (n.d.-b), State election statistics (1999–2018 percentages); VEC (n.d.-a), 2022 state election results (87 district recheck tables and statewide totals); and VEC (2023b), Report to Parliament (Narracan enrolment and date). Author calculations.

3. Melbourne’s margin is unusually large

Melbourne has the second-highest share of enrolled electors without a counted Lower House ballot among the 87 districts. Only Broadmeadows is higher. Each dot ranks a district by this one component, which a ranking of formal votes would conceal. Choose or click a district to carry it through Figures 3–5. The dashed line is the pooled share across enrolled electors in the 87 districts.

Choose a district or click a point in Figures 3–5.

Accessible reading: Broadmeadows ranks first at 19.5%, Melbourne second at 19.3% (10,529 enrolled electors minus counted ballots), and Werribee ranks 13th at 14.4%. The pooled 87-district share is 11.9%. These are arithmetic differences, not counts of people proven to have abstained.

Source: Victorian Electoral Commission (VEC, n.d.-a), 2022 state election results, individual district recheck tables; author calculations.

4. See the two routes across 87 districts

Melbourne and Werribee arrive at nearly the same formal share by different routes. The three dashed curves show 80%, 85% and 90% formal share of enrolment (top to bottom at the same turnout wherever all three are visible). Nearby districts help test whether the pair is unusual: Melbourne, Prahran and Albert Park have low counted turnout but relatively low informal rates; Werribee, Melton and Point Cook all have among the highest informal rates. These named examples were identified after inspection, not predefined geographic groups. Use “Compare the pair” to isolate the opening contrast.

Accessible reading: Melbourne had 80.7% counted turnout and 3.4% of counted ballots informal; Werribee had 85.6% and 9.7%. Prahran and Albert Park each had counted turnout below 83.0% and informal rates below 3.5%. Melton, Point Cook and Werribee are three of the four highest-informal-rate districts. Broadmeadows is a counterexample with high values on both loss measures. The dashed guides trace equal formal shares rather than measured trends.

Source: Victorian Electoral Commission (VEC, n.d.-a), 2022 state election results, individual district recheck tables; author calculations.

5. A clue in the ballot, not a verdict on voters

Longer Lower House candidate lists are associated with more informal ballots, but they do not explain every district. The VEC reported this district-level association (VEC, 2023a); the correlation calculated from these 87 districts is 0.66. Use the buttons to compare the 71 districts with 6–9 candidates against the 16 with 10–15. The fitted line describes all 87 districts. Broadmeadows has high informality despite having only nine candidates.

Accessible reading: the median district informal rate is 4.7% among the 71 districts with 6–9 candidates, versus 7.7% among the 16 districts with 10–15. Melton recorded 11.1% with 14 candidates; Werribee 9.7% with 15; Broadmeadows 10.8% with only 9. The focus views hide the full-sample fit; these unweighted district medians and the overall association do not establish a cause for any individual ballot.

The VEC’s ballot-paper review classified about 56% of examined Lower House informal papers as showing a candidate preference, while about 41% appeared deliberate; the classification infers intent from marks and has its own review total, which is different from the rechecked count used in these plots (VEC, 2023a). Candidate number, ballot design, instruction, language access and local context deserve careful study together. Horizontal jitter in this chart only separates overlapping marks; hover shows the actual integer candidate count.

Sources: VEC (n.d.-a), 2022 state election results, district recheck tables; VEC (2023a), Review of informal voting: 2022 State election (ballot classification and interpretation). Author calculations.

The result: Melbourne and Werribee have formal shares only 0.7 percentage points apart, yet Melbourne ranks second of 87 for its uncounted share (19.3% of enrolment), while Werribee ranks 4th for informality among counted ballots (9.7%). Melton and Point Cook join Werribee in the top four informal rates. Candidate-list length is one useful association, not a diagnosis: Broadmeadows also has high informality with nine candidates and ranks first for the uncounted share. These are distinct district-level patterns that call for different follow-up questions, not evidence of any voter’s intent or the cause of an uncounted ballot.

Data, definitions and accessibility

  • Formal share of enrolment = formal Lower House votes ÷ district electors enrolled at close of rolls. Counted turnout = formal plus informal votes ÷ those electors. Informal rate = informal ÷ ballots counted. Do not compare these percentages without their denominators.
  • “No ballot counted” means enrolment minus counted ballots in the district Lower House result. It is not an individual-level abstention classification and is not evidence of intent, exclusion or wrongdoing.
  • The 2022 main-election analysis includes 87 districts. Narracan’s Lower House election was held separately in January 2023. The historic 2022 point is recomputed from those 87 districts, while 1999–2018 historical points use VEC’s published rounded percentages. Do not treat the series as a controlled causal comparison.
  • The two local CSV snapshots were created from VEC pages by scripts/fetch_vec.py (included with the submission). The script counts candidate rows in each rechecked first-preference table and checks that their votes sum to each district’s formal total. The report reads these CSVs and calculates the displayed rates in R.
  • All marks contain text tooltips. The linked district chooser works with a keyboard; clicking a point in Figures 3–5 also selects that district across the three charts, and the status line states its two relevant rates. A gold ring denotes the active selection when it belongs to the chart’s current filter. Shape, written readings, chart labels, accessible chart descriptions and the complete table provide alternatives to colour and hover. Use the toolbar’s autoscale or double-click to restore axes after zooming.

Design choices: The story moves from an apparently equal pair to a different ranking, then combines turnout and informal rate before examining a district ballot feature. Figure 3 ranks the uncounted component, rather than repeating Figure 1’s formal-share measure. The district chooser connects that ranking to the two multivariate plots; Figure 5’s filter lets readers inspect districts with shorter and longer candidate lists separately. This balances an author-guided sequence and reader-led exploration (Segel & Heer, 2010). Melbourne remains dark blue and Werribee purple in every district comparison, based on ColorBrewer’s Paired palette (Brewer & Harrower, n.d.); grey denotes other districts, gold denotes selection, and slate is used only for ballot-status components. Shape, labelled axes, written readings and the data table support readers who cannot use hover or distinguish the colours (Elavsky et al., 2022). The clean background keeps the points and text legible.

Read the complete district data as a table
District Enrolled Formal Informal Counted Candidates Turnout % Informal % of counted Formal % of enrolled
Albert Park 48788 39110 1396 40506 8 83.02 3.45 80.16
Ashwood 51209 44710 1565 46275 8 90.36 3.38 87.31
Bass 49984 41404 3135 44539 10 89.11 7.04 82.83
Bayswater 49707 42599 2512 45111 9 90.75 5.57 85.70
Bellarine 55022 47153 2156 49309 9 89.62 4.37 85.70
Benambra 50045 41135 2099 43234 8 86.39 4.85 82.20
Bendigo East 53134 45568 2184 47752 7 89.87 4.57 85.76
Bendigo West 47808 40248 2401 42649 9 89.21 5.63 84.19
Bentleigh 51415 45120 1408 46528 6 90.49 3.03 87.76
Berwick 51910 44289 1875 46164 6 88.93 4.06 85.32
Box Hill 50799 44356 1635 45991 9 90.54 3.56 87.32
Brighton 48260 41196 1786 42982 9 89.06 4.16 85.36
Broadmeadows 48414 34780 4198 38978 9 80.51 10.77 71.84
Brunswick 52263 43445 1733 45178 9 86.44 3.84 83.13
Bulleen 50496 42959 2326 45285 8 89.68 5.14 85.07
Bundoora 47319 40336 2209 42545 7 89.91 5.19 85.24
Carrum 51265 43300 2169 45469 7 88.69 4.77 84.46
Caulfield 48492 40683 1374 42057 7 86.73 3.27 83.90
Clarinda 48450 40564 2477 43041 8 88.84 5.75 83.72
Cranbourne 51173 40822 3234 44056 9 86.09 7.34 79.77
Croydon 52544 45639 2054 47693 6 90.77 4.31 86.86
Dandenong 50846 39507 3188 42695 7 83.97 7.47 77.70
Eildon 50090 42399 2343 44742 8 89.32 5.24 84.65
Eltham 48931 43844 1507 45351 6 92.68 3.32 89.60
Essendon 49525 41867 1761 43628 8 88.09 4.04 84.54
Eureka 54487 45247 3297 48544 10 89.09 6.79 83.04
Euroa 50663 42242 2431 44673 7 88.18 5.44 83.38
Evelyn 50210 43665 2254 45919 7 91.45 4.91 86.96
Footscray 48860 40340 1654 41994 7 85.95 3.94 82.56
Frankston 49160 39941 2323 42264 9 85.97 5.50 81.25
Geelong 52376 44762 1918 46680 7 89.12 4.11 85.46
Gippsland East 49698 42212 1992 44204 7 88.95 4.51 84.94
Gippsland South 49715 42363 2354 44717 7 89.95 5.26 85.21
Glen Waverley 50818 44667 1529 46196 7 90.90 3.31 87.90
Greenvale 49854 37686 3952 41638 8 83.52 9.49 75.59
Hastings 48674 41232 2369 43601 8 89.58 5.43 84.71
Hawthorn 50080 44310 1178 45488 8 90.83 2.59 88.48
Ivanhoe 46961 40599 1536 42135 6 89.72 3.65 86.45
Kalkallo 52205 39980 3494 43474 9 83.28 8.04 76.58
Kew 49918 43592 1367 44959 8 90.07 3.04 87.33
Kororoit 51060 40373 3683 44056 9 86.28 8.36 79.07
Lara 48920 39521 2668 42189 8 86.24 6.32 80.79
Laverton 48545 37137 3144 40281 10 82.98 7.81 76.50
Lowan 51182 43297 2612 45909 7 89.70 5.69 84.59
Macedon 48019 41589 1974 43563 8 90.72 4.53 86.61
Malvern 46863 40801 1136 41937 6 89.49 2.71 87.06
Melbourne 54533 42519 1485 44004 9 80.69 3.37 77.97
Melton 50151 37619 4716 42335 14 84.42 11.14 75.01
Mildura 47941 38140 3096 41236 11 86.01 7.51 79.56
Mill Park 45937 37798 2510 40308 6 87.75 6.23 82.28
Monbulk 48547 41797 2063 43860 8 90.35 4.70 86.10
Mordialloc 50643 44059 2035 46094 8 91.02 4.41 87.00
Mornington 48048 42062 1823 43885 8 91.34 4.15 87.54
Morwell 53782 43261 3526 46787 12 86.99 7.54 80.44
Mulgrave 47029 37971 3603 41574 14 88.40 8.67 80.74
Murray Plains 48622 40191 2635 42826 7 88.08 6.15 82.66
Narre Warren North 51900 41370 3602 44972 10 86.65 8.01 79.71
Narre Warren South 53430 42859 3349 46208 8 86.48 7.25 80.22
Nepean 48944 40766 2753 43519 11 88.92 6.33 83.29
Niddrie 50034 42242 2763 45005 9 89.95 6.14 84.43
Northcote 50369 42642 2445 45087 11 89.51 5.42 84.66
Oakleigh 51934 43910 2000 45910 8 88.40 4.36 84.55
Ovens Valley 48125 40390 2101 42491 7 88.29 4.94 83.93
Pakenham 48014 38867 3089 41956 11 87.38 7.36 80.95
Pascoe Vale 48584 40273 2249 42522 8 87.52 5.29 82.89
Point Cook 49788 39099 4448 43547 15 87.46 10.21 78.53
Polwarth 53064 46002 1949 47951 7 90.36 4.06 86.69
Prahran 48933 39243 1223 40466 6 82.70 3.02 80.20
Preston 50362 39857 3392 43249 10 85.88 7.84 79.14
Richmond 48305 39731 1381 41112 8 85.11 3.36 82.25
Ringwood 52040 45753 1676 47429 7 91.14 3.53 87.92
Ripon 50813 42341 3111 45452 9 89.45 6.84 83.33
Rowville 49738 43369 2070 45439 7 91.36 4.56 87.19
Sandringham 46308 40512 1699 42211 8 91.15 4.03 87.48
Shepparton 50994 41301 2663 43964 9 86.21 6.06 80.99
South Barwon 51546 45184 1637 46821 8 90.83 3.50 87.66
South-West Coast 53099 45496 2723 48219 10 90.81 5.65 85.68
St Albans 46448 35716 3638 39354 10 84.73 9.24 76.89
Sunbury 45582 37706 2620 40326 9 88.47 6.50 82.72
Sydenham 53100 43665 3591 47256 8 88.99 7.60 82.23
Tarneit 48792 38978 2946 41924 9 85.92 7.03 79.89
Thomastown 47657 37476 3104 40580 6 85.15 7.65 78.64
Warrandyte 50951 45075 1582 46657 6 91.57 3.39 88.47
Wendouree 50461 42175 2396 44571 7 88.33 5.38 83.58
Werribee 49962 38603 4156 42759 15 85.58 9.72 77.26
Williamstown 45965 38974 1878 40852 8 88.88 4.60 84.79
Yan Yean 47327 39419 2475 41894 9 88.52 5.91 83.29

References and AI acknowledgement

Brewer, C. A., & Harrower, M. (n.d.). ColorBrewer 2.0: Color advice for cartography. Pennsylvania State University. https://colorbrewer2.org/

Elavsky, F., Bennett, C. L., & Moritz, D. (2022). How accessible is my visualization? Evaluating visualization accessibility with Chartability. Computer Graphics Forum, 41(3), 57–70. https://doi.org/10.1111/cgf.14522

OpenAI. (2026, October 26). ChatGPT. https://chatgpt.com/

Segel, E., & Heer, J. (2010). Narrative visualization: Telling stories with data. IEEE Transactions on Visualization and Computer Graphics, 16(6), 1139–1148. https://doi.org/10.1109/TVCG.2010.179

Victorian Electoral Commission. (n.d.-a). 2022 state election results [Data set]. https://www.vec.vic.gov.au/results/state-election-results/2022-state-election-results

Victorian Electoral Commission. (n.d.-b). State election statistics [Data set]. https://www.vec.vic.gov.au/results/electoral-statistics/state-election-statistics

Victorian Electoral Commission. (2023a, August). Review of informal voting: 2022 State election [Report]. https://www.vec.vic.gov.au/-/media/dbb32440950847748761ea3481a397d4.pdf

Victorian Electoral Commission. (2023b, October). Report to Parliament: 2022 Victorian State election and 2023 Narracan District supplementary election [Report]. https://www.vec.vic.gov.au/-/media/2cf7d4413e9a4878858e6ed282f46c61.pdf

AI acknowledgement. I used ChatGPT (OpenAI, 2026) for topic suggestions, public-source discovery, R code troubleshooting, and wording. I reviewed and revised its output and checked the election data and claims against Victorian Electoral Commission sources.