Heading into these decisive weeks, we will hear much about the uninviting weirdness of Ed Miliband and the uncaring poshness of David Cameron. Party strategists are ramping up the politics of personalities in an attempt to break the current stalemate in the polls.

But do we, as an electorate who are bombarded by these incessant character assassinations, really conceive of general elections as mere popularity contests?

In these most uncertain of election periods, it’s wise to remember that the best lessons of where we are going in the future often come from studying patterns from the past.

Below is a graph compiled by Cicero Elections from the monthly political barometer surveys of Ipsos MORI. Here we aim to measure the effects, if any, of the popularity of party leaders against the general elections of the past two decades, and beyond.

Leader Approval Ratings GE 1997 – Present:

Leader popularity graph

There are three key takeaway points.

First, whichever party leader, from either the Conservatives or Labour, has been more popular at the time of a general election, that party has won. This holds true going back even further than the data shown here, to when the monthly political barometer started in 1977. So, assuming a concurrent pattern based on the data of the last 8 elections since 1977 is maintained, David Cameron’s greater popularity over Ed Miliband forecasts a Tory win.

Second, the margin of popularity between the main two party leaders seems to bare very little indication as to the scale of the victorious party in terms of vote share on polling day. Little can be causally measured between Tony Blair’s approval rating of more than 50% over John Major in 1997 and the Labour victory by a margin of 13% in the popular vote, particularly since Blair’s net approval against William Hague dropped to around 20% in 2001, with Labour’s margin of victory that year still being a huge 9%. Even more stark still are the Liberal Democrat figures, which between the five years of 2005 and 2010 jump between a leadership approval low of below 20% and a high of close to 70%, only for the party to secure roughly the same number of seats at the two elections.

The scale of a leader’s popularity seems to fluctuate too widely to attempt to translate these numbers into a solid forecasting model for vote share ultimately won. So, where each leadership approval rating lands at the time of an election is the only consistently meaningful facet of this data, and only insofar as it suggesting who will come out on top between Labour and the Conservatives.

Third, the public has never had such low regard for the main three party leaders when the group is taken as a whole. Even going back to when the data begins in 1977, at no point has the cumulative approval of the leaders tracked for the parties shown been lower. This point is likely married to the fact that we are now seeing historically high levels of public disinterest in the main three parties in voting intention polling.

Perhaps the reality here is that party leader popularity is more of a shadow on the wall rather than a real indicator of how new Parliaments will be shaped. Only on May 8th will we know whether the last eight consecutive elections showing the correlations highlighted here were just fluke, or whether it proves to be one of the few constant patterns to emerge in even the most uncertain of election cycles.

Whatever will come to pass at GE 2015, let’s at least enjoy the fact that we have this rich history of data to plough into.