Just how accurate was http://www.boards.ie/vote?

Darragh

This post is long overdue but we've run the figures on how accurately those who voted on http://www.boards.ie/vote predicted the 2011 General Election

We'd like to thank everyone who took the time to vote and who encouraged others to vote. Final results are over here.

Here it is by constituency, showing that

• In 8 of the 43, Boards.ie was 100% correct
• In 23 of the 43, the results were one seat out
• In only 3 constituencies were we more than 2 seats out - Cork North West, Galway West and Kerry South. Galway West only predicted 1 out of 5.

Constituency|Boards Votes|Seats|Predicted on Boards.ie|% Accuracy
Carlow-Kilkenny|626|5|4|80
Cavan-Monaghan|382|5|3|60
Clare|530|4|3|75
Cork East|410|4|3|75
Cork North Central|406|4|3|75
Cork North West|295|3|1|33.33
Cork South Central|663|5|3|60
Cork South West|269|3|3|100
Donegal North East|258|3|2|66.67
Donegal South West|349|3|3|100
Dublin Central|636|4|3|75
Dublin Mid West|484|4|3|75
Dublin North|780|4|2|50
Dublin North Central|513|3|3|100
Dublin North East|542|3|2|66.67
Dublin North West|358|3|2|66.67
Dublin South|1,051|5|4|80
Dublin South Central|590|5|4|80
Dublin South East|806|4|2|50
Dublin South West|596|4|4|100
Dublin West|707|4|3|75
Dun Laoghaire|726|4|3|75
Galway East|533|4|2|50
Galway West|913|5|1|20
Kerry North-West Limerick|315|3|3|100
Kerry South|257|3|1|33.33
Kildare North|712|4|3|75
Kildare South|444|3|2|66.67
Laois Offaly|521|5|3|60
Limerick|318|3|2|66.67
Limerick City|539|4|3|75
Longford Westmeath|483|4|2|50
Louth|643|5|4|80
Mayo|590|5|3|60
Meath East|434|3|3|100
Meath West|372|3|2|66.67
Roscommon South Leitrim|338|3|3|100
Sligo North Leitrim|404|3|2|66.67
Tipperary North|275|3|3|100
Tipperary South|315|3|2|66.67
Waterford|654|4|3|75
Wexford|654|5|3|60
Wicklow|817|5|4|80
Total|22,508|166|117| |

Your comments are welcome and if you have any questions that I can help you with, please post below.

Once again, thanks for being part of an interesting experiment for us.

Darragh

SkepticOne

That is not too bad, imo, given the demographics of boards.ie.

voodoodoll1654

I'd like to see accuracy based on first preference votes too, if possible? As in boards %pref for each nominee against actual %pref.

TheCosmicFrog

Yeah, Galway West was way off. I don't think Eamon Ó Cuív even got a seat on the Boards.ie election. He missed the first seat by a very narrow margin in the real thing.

mr_angry

I'm not sure if the above table really represents the 'accuracy' of the Boards poll. For example, in my constituency of Dublin South East, the Boards poll had Dylan Haskins on top, when in reality, he was a long way from being elected. By that measure, the poll was some way from being '50% correct'.

zig

Darragh wrote: »

• In 23 of the 43, the results were one seat out
• In only 3 constituencies were we more than 2 seats out - Cork North West, Galway West and Kerry South. Galway West only predicted 1 out of 5.

It was a great idea and well done, but it only really served the purpose of showing how boards people feel as opposed to the Irish population.
The bit I put in bold is huge imo, it was the difference between FF either crashing to nothing or being the main opposition party.
While it wasnt too bad it wasnt a patch on a the real polls unfortunately.

kmick

Off topic but something struck me about Dylan Haskins at the time.

Dylan Haskins=D.H.=Doogie Howser

Coincidence or spooky?

DeVore

I think if you look very very closely at the fine detail of the results you will find inaccuracies, for example I doubt we got a single candidates' % of votes-in-constituency correct but when you begin to zoom out and look at the overall picture two phenomena begin to take effect. 1. the large of big numbers and 2. the wisdom of crowds. (quite similar effects to be fair).

So, this leads to things like being waaay out on the Cork constituency but begin relatively accurate on things like the rise of Independents and the collapse of FF.

No one else get constituency level predictions, did they?? I would love to have seen how accurate RedC or MRBI would have been with their 1000 sample set giving them just 25 votes per constituency. Did any of them try that??

This was an interesting experiment and we're delighted people responded and trusted us as much as they did.

DeV.

Permabear

This post has been deleted.

To Be fair,it was a nice experiment and a worthwile game.
However:
13 seats under estimate for FG
17 seats under estimate for FF
10 seat over estimate for labour
14 seat over estimate for sinn féin

On those figures,we could have had a ULA,Labour SF left leaning independents alternative government and thats in no way an accurate reflection of how people voted.

It's all very well to say you got Stephen Donnelly correct but in Wicklow you had him first in and Brady of SF comfortable too with Billy timmins losing his seat.
The haskins thing in Dublin SE was another example.
There are loads more.

The only way round those symptoms of a skewed mindset on boards is to widen the poll audience to include 1000's country wide of non boardsies.

The most accurate poll I saw,prior to the result/tallies was on just 1500 people and that was RTE's exit poll.

dilallio

TheCosmicFrog wrote: »

Yeah, Galway West was way off. I don't think Eamon Ó Cuív even got a seat on the Boards.ie election. He missed the first seat by a very narrow margin in the real thing.

A number of posters noted at the time in Galway West, that in the space of 5 hours one day, approx 30 people voted for Green candidate Niall O'Brolchain, pushing him into the top 5. Nobody in any of the threads gave him any chance of being elected so it was surprising to see him get that many votes, and equally surprising that most of these votes weren't removed after boards checked.

The Sinn Fein candidate also had a lot of votes appear in a short period of time, but this wasn't that surprising as a lot of posters expected Sinn Fein to do a lot better than they actually did in GW.

zig

Deleted User wrote: »

The most accurate poll I saw,prior to the result/tallies was on just 1500 people and that was RTE's exit poll.

Exit poll is always going to be bang on though, and its fairly pointless because its right before the count starts.

But overall your right, the other polls used small samples and were faaarr more accurate than this one.

[Jackass]

Or, in short, the accuracy of predicting elected officials was 71%.

Not bad.

Especially considering how difficult it was to cast the extraordinary first time ever results that did occur.

It would be interesting to see how accurate it is in a more settled political environment. More so I'd imagine.

Kudos Boards. Nice experiment.

K-9

Deleted User wrote: »

To Be fair,it was a nice experiment and a worthwile game.
However:
13 seats under estimate for FG
17 seats under estimate for FF
10 seat over estimate for labour
14 seat over estimate for sinn féin

On those figures,we could have had a ULA,Labour SF left leaning independents alternative government and thats in no way an accurate reflection of how people voted.

It's all very well to say you got Stephen Donnelly correct but in Wicklow you had him first in and Brady of SF comfortable too with Billy timmins losing his seat.
The haskins thing in Dublin SE was another example.
There are loads more.

The only way round those symptoms of a skewed mindset on boards is to widen the poll audience to include 1000's country wide of non boardsies.

The most accurate poll I saw,prior to the result/tallies was on just 1500 people and that was RTE's exit poll.

Strangely enough Labour and SF weren't that far out percentage wise about 2% over.

The percentage bias was more towards Independents and the Green vote over estimated too.

SF getting 25 seats was odd with only 11.5% of the vote

[Jackass]

Where the failure was, was that the Boards poll was overly harsh on FF (possibly based on a large Dublin base of users?) and didn't predict FG picking up the pieces, instead, more or less Sinn Fein and Labour doing better than expected, again, possibly due to disproportionate Dublin based voters, and lastly, the age profile meant that FF, mostly an older generation of voters, would not be as prominent on an internet forum.

Demographics failed the poll, but within the specified sample population, I think it was quite accurate, if not at predicting the overall Government.

Victor

[Jackass] wrote: »

Where the failure was, was that the Boards poll was overly harsh on FF (possibly based on a large Dublin base of users?) and didn't predict FG picking up the pieces, instead, more or less Sinn Fein and Labour doing better than expected, again, possibly due to disproportionate Dublin based voters

Assuming people voted in ttheir own constituencies, this shouldn't have mattered.