A reader sends in his analysis:

I have recently performed some statistical analyses of results from the 2008 and 2011 elections, in order to test a theory about voter behaviour in 2011.

You are probably aware of what this analysis shows already. I hope so.  But given

a) the importance of the results and

b) the widespread assumption of contrary conclusions among the commentariat (eg that the low turnout hurt Labour);

I thought I should pass this on to you, just in case.

When the 2011 election proved unexpectedly close on a reduced turnout, I theorised that, misled by the polls, many National-sympathetic voters had simply stayed home or indulged themselves with a vote for NZ First or the Conservative Party in order to make a political point.

It recently occurred to me that my theories were testable from the 2008 and 2011 election statistics by testing for statistical correlations. The results of this were:

 Variables Correlation co-efficient* National Party vote in 2008 vs Conservative Party vote 2011 (general electorates) 0.601 Change in National Party vote between 2008 and 2011 vs change in NZ First Party vote 2008-2011 (general electorates) -0.379 Change in Labour Party vote between 2008 and 2011 vs change in NZ First Party vote 2008-2011 (general electorates) -0.004 National Party vote in 2008 vs change in overall turnout 2008-2011 (general electorates) 0.271 National Party vote in 2008 vs change in overall turnout 2008-2011, (general electorates, excluding Christchurch) 0.379 Labour Party vote in 2008 vs change in overall turnout 2008-2011, (general electorates, excluding Christchurch) -0.278

* A coefficient of 0 means the variables are unrelated. A coefficient of ±1 means there is a perfect relationship.

What this means

1)      Conservative Party votes in 2011 came overwhelmingly from National (no surprise).

2)      There is an inverse relationship between the increases in NZ First vote in 2011 and the relative performance of National. In other words, the increased NZ First vote was mainly at National’s expense. There was almost zero correlation between the NZ First and Labour vote.

3)      The higher the National vote in 2008, the bigger the decline in turnout in 2011. In other words, it was National voters (more than Labour voters) who stayed home.

I sought corroboration of this finding by listing those general electorates where the turnout decline between the two elections was greater than 6% (on average, the turnout decline was 5.35%).

The electorates were: Botany, Chiristchurch Central, Chch East, Clutha Southland, Dunedin North, ECB, Hamilton East, Helensville, Hunua, Invercargill, Manurewa, North Shore, Northcote, Pakuranga, Whangarei and Wigram.

Of these 16 electorates, 11 had a 2008 National Party party vote greater than the national average for general electorates of 47.3%. With the Christchurch electorates excluded (there being special reasons for turnout decline there) this becomes 11 out of 13 electorates.

Conclusion

Contrary to “received wisdom” it was National that suffered from the reduced turnout in 2011. Additionally, the NZ First vote was boosted primarily by defections from National. Uncontroversially, it is confirmed that Conservative votes came overwhelmingly at National’s expense.

My theory that the above phenomena were a result of complacency in the face of the widespread expectation of National waltzing home with a win remains only a theory. But it is one that fits the facts quite well.

However, it seems to me that if true, the greatest danger for National in 2014 is, again, complacency and a failure of potential supporters to vote for the party (whether by staying home or by risking a vote for other parties that may not meet the threshold criteria or may not support National after the election).

This is a fascinating analysis, backed by hard data, not guesswork.

Labour’s entire strategy seems to be to have abandoned moderate centrist voters on the assumption the non voters last time were primarily Labour voters who just need a reason to turn out.

I agree with the conclusion that National supporters can not be complacent. If National supporters do not turn out in 2014, then Mr Dotcom may be picking the Government.