Smoothing out the peaks and troughs of the Democrat vote

For a politician, Andrew Murray never was very good at numbers. As one of the architects of the Democrats’ GST deal he massively overestimated the amount the Howard Government would spend on the environment in return for his support. When he ran against Meg Lees for the leadership of the Democrats in 1999 he misread the mood of the members completely when he attracted only single digit support. But his recent attempt at electoral analysis (see his On Line Opinion article) is by far his worst performance to date.

For a start, Senator Murray relies on a rather unsophisticated statistical trick to conceal both the depths that the polls sank to under Senator Lees and the peaks they reached under Senator Stott Despoja and the then Senator Kernot. While he states in his article that his graph “stands alone”, most readers most likely missed the implications of Senator Murray’s decision to present a “moving average” of the poll results rather than the actual poll results themselves.

A moving average is a way of smoothing data to conceal peaks and troughs. That is, rather than stating what the polls said about voter attitudes on January 20, a moving average provides information on what the voters had thought about a party or a leader, on average, for an arbitrary period prior to January 20. In presenting his “stand alone” graph, Senator Murray appears to have decided that it was not the fortnightly polls that mattered, but the average poll result for the six month period prior to any poll date. (We can’t say for certain how long Senator Murray chose to smooth the results over as he didn’t think it was worth informing his readers -  a bit of experimenting suggests that 6 months looks about right though.)

It is often said that a week is a long time in politics. Six months would therefore seem to many to be an eternity. But Senator Murray’s manipulation of the original data does much more than smooth out an otherwise bumpy graph, it gives Meg Lees some credit for Cheryl Kenot’s superior performance in the polls and it drags down the jump in support that followed the election of Senator Stott Despoja. A cynic might argue that Senator Murray appeared to be looking around for a set of numbers that boosted the performance of his friend Meg and dragged down the performance of the leader he pledged to bring down.

An example helps explain the lengths to which Senator Murray has gone in his attempt to convince himself, if not others, that Meg Lees was a political asset and Senator Stott Despoja was as unattractive to voters as he himself was to Democrat members.

Senator Murray states, “After a sustained period of attack, the Party votes out Lees in favour of Stott Despoja in March 2001. Stott Despoja briefly lifts the Democrats an encouraging 2 percentage points until September 2001 when a continuous downward slide begins”. Leaving aside the fact that Senator Murray is mistaken about the date that the Members rejected Senator Lees in favour of Senator Stott Despoja (it was April not March), his determination to rely on a moving average rather than the actual data not surprisingly understates the extent of the lift in the polls associated with the change in leadership.

Between February 9 and February 25 Newspoll had the Democrats at what was then the near record low level of 3 per cent. In the first poll taken after the change of leadership Newspoll’s measure of support for the Democrats rose to 9 per cent.

Of course you could argue that all new leaders would expect a bounce in the polls but unfortunately that wasn’t true for Senator Lees who saw her stocks fall from 5 per cent in December 1997, when she became leader, to 2 per cent in March the following year.

Senator Murray is no doubt proud of his decision to support the GST, and he no doubt feels that his decision to do so was in the national interest. But his determination to convince himself that the GST was a vote winner for the Democrats, and his continued belief that the decision to replace Meg Lees in the lead up to the 2001 election was a vote loser, is no longer just misguided, it is a dangerous folly that his party can ill afford.

At the time of the GST deal Senator Lees commissioned polling to test voter sentiment and Democrat voter sentiment in particular. It found that 71 per cent of Democrat voters were opposed to the idea of a deal on the GST. Either Senator Murray was not privy to that polling at the time, which would be a pity, or he was shown it but continues to keep his head in the sand. Either way, the sad reality for the Democrats is that Andrew Murray is one of the few Australians who can say that the GST was a vote winner for the Democrats and keep a straight face. Even sadder is that it seems his inability to move on from his misjudgement is still holding back the whole party.