Data suggests the NDP may win the Outremont by-election

A combination of polling data and monetary data suggests that the chances of the NDP winning the upcoming Outremont by-election and gaining a Quebec MP are good.  However, regardless of who wins, I would bet money on this being a close race and the data seems to validate this bet.

I’ll briefly explain the data and how I came to these conclusions.

I started off with the local results in the Outremont riding from the 2006 General Election.

	Lib	Bloc	NDP	Con	Green
2006 Outremont results:	35.18%	29.01%	17.20%	12.73%	4.82%

From that, we can take current provincial polling data, courtesy of the Paulitics Provincial Polling Resource and compare that with the provincial results in 2006 to get a ratio describing the relative increase or decrease of each party.  This ratio will later be multiplied through the 2006 Outremont results to get the first set of data.

	Lib	Bloc	NDP	Con	Green
Current polls @ prov. level	21.40%	35.40%	13.40%	22.00%	6.80%
	Lib	Bloc	NDP	Con	Green
2006 Quebec results:	20.70%	42.10%	7.50%	24.60%	4.00%
increase/decrease (ratio)	1.03	0.84	1.79	0.89	1.70

We’ll come back to that ratio later.

But for now, let’s move on to the monetary data portion of the analysis.

We know what the cash spending limit for this particular riding is from elections Canada and we know the financial statements of the candidates from the last election (note: the previous link was working earlier today, but seems to be down now.  There is an alternate, and less pretty source of the same data here).

From this, we get:

	Lib	Bloc	NDP	Con	Green
Cash spent: 2006 election	$69,816.11	$63,590.41	$26,625.29	$73,991.17	$572.33
cost per 1% of vote	$1,987.36	$2,207.23	$1,554.31	$5,762.55	$119.48
Spending limit:	$74,512.38	$74,512.38	$74,512.38	$74,512.38	$74,512.38
Raw vote potential	37.5%	33.8%	47.9%	12.9%	insuf. samp.
vote potential	28.8%	26.0%	36.9%	9.9%	insuf. samp.

I don’t think too many people would be willing to take just the polling data or just the financial data to come up with any sort of prediction.  So, I figured that the best way to come up with some sort of reliable prediction-worthy data, it would be suitable to take the pro-rated vote potential calculated from the financial data and then average that with the pro-rated vote potential calculated form the polling data (using the ratio calculated above).

I’ve run this calculation using 4 different scenarios so nobody can accuse me of bias (not that I’d vote for any of these parties if I had my first choice). 

In scenario #1, I haven’t weighted anything and I’ve assumed that the Green Party will not want to invest significant financial resources into this by-election and thus, I’ve listed their cash pro-rate as equal to their vote pro-rate.

Scenario #1	Lib	Bloc	NDP	Con	Green
vote pro-rate (raw)	36.3	24.2	30.6	11.5	8.1
vote pro-rate	32.8%	21.9%	27.6%	10.4%	7.4%
cash pro-rate (raw)	28.8	26.0	36.9	9.9	8.1
cash pro-rate	26.3%	23.7%	33.6%	9.1%	7.4%
predicted results:	29.5%	22.8%	30.6%	9.7%	7.4%
Elected:			x

The result is an NDP victory, although by the slightest of margins.

In scenario #2, I haven’t weighted anything but I’ve assumed that the Greens will throw a significant portion of financial resources at this by-election.  So I’ve listed their cash pro-rate as equal to the highest cash pro-rate of all of the other parties.

Scenario #2	Lib	Bloc	NDP	Con	Green
vote pro-rate (raw)	36.3	24.2	30.6	11.5	8.1
vote pro-rate	32.8%	21.9%	27.6%	10.4%	7.4%
cash pro-rate (raw)	28.8	26.0	36.9	9.9	36.9
cash pro-rate	20.8%	18.7%	26.6%	7.2%	26.6%
predicted results:	26.8%	20.3%	27.1%	8.8%	17.0%
Elected:			x

 The result is still an NDP victory, but by even smaller margins than before.

In scenario #3, I’ve assumed that the Greens will be middle of the road with their finances and won’t go as spartan as they did in the last election, but won’t go all out either.  I’ve also assumed for this scenario that polls matter more than cash on hand and have weighted to 2x its normal unweighted value.

Scenario #3	Lib	Bloc	NDP	Con	Green
vote pro-rate (raw)	36.3	24.2	30.6	11.5	8.1
vote pro-rate	32.8%	21.9%	27.6%	10.4%	7.4%
cash pro-rate (raw)	28.8	26.0	36.9	9.9	22.5
cash pro-rate	23.2%	20.9%	29.7%	8.0%	18.1%
predicted results:	29.0%	21.5%	28.5%	9.4%	11.7%
Elected:	x

The result is a bare Libearl victory.

Finally, in scenario #4, I’ve assumed the same thing about the Greens as in #3, but this time I’ve assumed that cash on hand for the candidates matters more than polls and have weighted it by 2x.

Scenario #4	Lib	Bloc	NDP	Con	Green
vote pro-rate (raw)	36.3	24.2	30.6	11.5	8.1
vote pro-rate	32.8%	21.9%	27.6%	10.4%	7.4%
cash pro-rate (raw)	28.8	26.0	36.9	9.9	29.7
cash pro-rate	22.0%	19.8%	28.1%	7.6%	22.6%
predicted results:	26.3%	20.6%	27.9%	8.7%	16.5%
Elected:			x

The result is an NDP win and by wider margins than before.

Now, keep in mind, there are any number of factors which can’t be calculated mathematically which will undoubtedly play a part in this by-election.

For starters, there’s no way of accounting for the fact that the NDP has a star candidate in this race.  This data assumes that the candidate is of little-to-no importance whatsoever.

Second, polls suggest that Dion may not be as much of an asset to the Liberal candidate in this election as Layton or Harper might be.

But, on the other hand, the Globe and Mail suggests that Mulcair’s leftist credentials are being questioned by NDP activists (I know, I was shocked too — who thought the NDP still had leftist credentials??) which could in turn cause a ‘get out the vote’ (GOTV) problem.

Either way, the only thing I’m willing to place money on right now is that it’s not going to be a blow out.  But I will say one thing:  The NDP’s chances are certainly good seeing as how they won 3 out of the 4 scenarios I ran.