Election 2020 – The One With All The Tables…

Acknowledging I am by no means an expert on statistics or Excel formulas, this post is essentially providing the raw data and working out of trends that underpin my previous two posts. I am putting it here so that others can fact-check my work and identify if – and where – I may have made an error.

All the raw data is culled from the Parliamentary Registrar’s website; I simply took some time to do the data entry then developed what formulas made sense to me for the questions I wanted to ask.

Also, in reviewing my sums I realised I made some mistakes in my initial analysis, and the corrected data is represented below. I messed up calculating the % change between the two elections. This post identifies these and corrects them.

NB – I disregarded all the spoiled/rejected ballots in my analysis. However, they can prove a proxy for FDM or OBA votes in constituencies lacking such candidates, or general voter dismay.

2017 results
Constituency PLPOBAIndependentNonvoter
15903650318
26903600360
37372280453
46524290372
58081920336
663528160292
74654860347
827472915323
94055400329
104544970310
115134190302
121527480341
136163370321
1456838541265
15733730344
16820580386
175402830322
186393260286
193795020350
201766500364
2153822834422
223977040376
233176450413
247233340376
254284930335
267241920358
276613380284
285305180331
297261450371
304635160392
314575060361
3264526319351
336533390411
347301700340
356442610382
365772970380
Totals200591383716912604
Table 1 – The 2017 Election Results
2020 results
Constituency PLPOBAFDMIndependentNonvoter
160819500502
2576177650521
366312500638
465522300533
56411051140520
663416200503
7422326890471
831854700473
943240300471
1032450300441
1156728200411
121725319612542
13496197630453
1445422200543
156034400538
1600 00
1700 00
18626124370425
19368307390509
2012448500549
2143201250626
2236547200553
2328244700567
24564240013529
2541940500487
2600 00
275321924839440
2849832900521
296307800520
3048640000510
3144138800486
325511311000481
3353202020580
34611631310486
3553897903529
364311141850520
Totals15995831413846716878
Table 2 – The 2020 Election Results
Constituency PLPOBAIndependent/FDMNonvoter
2017200591383716912604
2020159958314145116878
2020 % to 2017 %79.7460.09858.58133.91
% Change20.2639.91-758.58-33.91
Table 3 – This is the % change between the 2017 and 2018 elections – the key line is the bottome one “% Change” – and the numbers should be read in the inverse. The PLP vote declined by -20.3%; the OBA vote declined by -39.9%; the non-voter bloc increased by +33.9%.
2017 results
Constituency PLPOBAIndependentNonvoter
15903650318
26903600360
37372280453
46524290372
58081920336
663528160292
74654860347
827472915323
94055400329
104544970310
115134190302
121527480341
136163370321
1456838541265
15733730344
186393260286
193795020350
201766500364
2153822834422
223977040376
233176450413
247233340376
254284930335
276613380284
285305180331
297261450371
304635160392
314575060361
3264526319351
336533390411
347301700340
356442610382
365772970380
Totals179751330416911538
Table 4 – 2017 Results with 3 uncontested seats removed (C16, C17 and C26).
2020 results
Constituency PLPOBAFDMIndependentNonvoter
160819500502
2576177650521
366312500638
465522300533
56411051140520
663416200503
7422326890471
831854700473
943240300471
1032450300441
1156728200411
121725319612542
13496197630453
1445422200543
156034400538
18626124370425
19368307390509
2012448500549
2143201250626
2236547200553
2328244700567
24564240013529
2541940500487
275321924839440
2849832900521
296307800520
3048640000510
3144138800486
325511311000481
3353202020580
34611631310486
3553897903529
364311141850520
Totals15995831413846716878
Table 5 – 2020 Results with 3 uncontested constituencies removed (C16, C17 and C20).
Constituency PLPOBAIndependent/FDMNonvoter
2017179751330416911538
2020159958314145116878
2020 % to 2017 %88.9862.49858.58146.28
% Change11.0237.51-758.58-46.28
Table 6 – Demonstrates the % change in the vote by category, accounting for the 3 uncontested seats. The key line is the bottom “% Change” which should be read in the inverse. The PLP vote decline by -11%; the OBA vote declined by -37.5%; the non-voter bloc increased by +46.3%.

Basically what we can see here is that the PLP vote did not decline anywhere near to the degree that the OBA vote did. The combined FDM/Indpendent column is really rather meaningless at this stage – 2020 provides a baseline for the 2020 going forward only. However, the data – at another level – will help us try to understand where the PLP/OBA lost voters went and where the FDM vote came from. This to be looked at later; first I was curious what the picture looked like at each constituency.

% Change 2017-2020
Constituency PLPOBANonvoter
1-3.0546.58-57.86
216.5250.83-44.72
310.0445.18-40.84
4-0.4648.02-43.28
520.6745.31-54.76
60.1642.35-72.26
79.2532.92-35.73
8-16.0624.97-46.44
9-6.6725.37-43.16
1028.63-1.21-42.26
11-10.5332.70-36.09
12-13.1629.01-58.94
1319.4841.54-41.12
1420.0742.34-104.91
1517.7439.73-56.40
182.0361.96-48.60
192.9038.84-45.43
2029.5525.38-50.82
2119.70100.00-48.34
228.0632.95-47.07
2311.0430.70-37.29
2421.9928.14-40.69
252.1017.85-45.37
2719.5243.20-54.93
286.0436.49-57.40
2913.2246.21-40.16
30-4.9722.48-30.10
313.5023.32-34.63
3214.5750.19-37.04
3318.53100.00-41.12
3416.3062.94-42.94
3516.4662.84-38.48
3625.3061.62-36.84
Mean9.6542.14-47.15
Table 7 – % Change in vote for PLP, OBA and non-voters between 2017 and 2020. To be read in the inverse (positives to be read as negatives).

Now we get to see the change in each constituency. I am rusty with Excel, but all of these should be read in the inverse. That which is positive is actually a decline and vice versa. I’ve also not included the FDM here – they are looked at elsewhere.

Also, one should note that at least two of these constituencies lacked an OBA candidate, and so saw a -100% decline, which somewhat warps the data. As such, the key takeaways are:

  1. The PLP declined on average by -9.7% in each constituency.
  2. The OBA vote declined on average by -42.1% in each constituency. Adjusting for the 2 constituencies lacking OBA candidates, their actual average decline was -38.4% in each constituency.
  3. The non-voter bloc increased on averaged by +47.2% in each constituency.
  4. The PLP vote increased in some areas and decreased in some.
  5. The largest PLP vote declines were in C20 (-29.6%); C10 (-28.6%); C36 (-25.3%); C24 (-21.99%); C5 (-20.7%); C14 (-20.1%); C21 (-19.7%); C27 (-19.5%); and C13 (-19.5%).
  6. The largest PLP vote increases were in just 7 constituencies (29 saw declines): C8 (+16.1%); C12 (+13.2%); C11 (+10.5%); C9 (+6.7%); C30 (+5%); C1 (+3.1%); and C4 (+0.5%).
  7. The largest PLP vote decline, in C20, may largely be attributed to a boundary change that removed a significant number of traditionally PLP voters, moving them to the previously marginal C17. C17 was uncontested in this election, however it is likely this boundary change renders C17 less of a marginal and more of a PLP safe seat going forward.
  8. The causes for the changes in the PLP votes reflected above need to be investigated further (boundary changes, candidate ‘star’ power, canvassing/campaign investment are all possible, but harder to quantify).
  9. The OBA vote saw an increase in only one constituency, C10, with a +1.2% increase.
  10. The five largest OBA vote declines were in C34 (-63%); C35 (-62.8%); C18 (62%); C36 (61.6%); and C2 (50.8%).
  11. The median OBA decline was 39.7%; The median PLP decline was 11%.
  12. It is worth noting that three of the OBA’s largest declines were in seats that saw a specific concentration of FDM activity (more later).
  13. It is also worth noting that one of the OBA’s largest declines was the one held by their Leader and considered one of the OBA’s safest seats – indicative of a general vote of no confidence in the OBA Leader by its supporters?

Now, as this post is getting long and is already chockful of data, I am going to end it here, and devote another post specifically to trying to answer the following questions:

  1. Where did the missing PLP/OBA voters go?
  2. Where did the FDM vote come from?

One thought on “Election 2020 – The One With All The Tables…

  1. Excellent analysis.

    I am hopeful that an internal analysis by the PLP will look closely at the 11% median reduction.
    I hope it isn’t lost since the end result (30 – 6) was so large.

    The OBA, at a much higher median reduction, has more than a lot to do.

    Thanks for taking time to provide the analysis.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s