1. Local by-elections take place for authorities with differing responsibilities. To state the obvious, people are more likely to change their votes e.g. because of the qualities of the candidates standing, in partish or town council by-elections than those for county councils or large unitary authorities. The more local the election, the less good it is likely to be as a guide to national party standings.
2. The salience of national issues as opposed to ideosyncratic local ones varies greatly from place to place. For instance, in general my experience is that the swing is more likely to depend on the individuals standing in more rural ares with a lower turnover of electors than in more amorphous cities. And to some extent things just depend upon local political culture. For instance the South Wales valleys, whilst until recently overwhelmingly Labour at General Elections, were always quite prepared to vote for independents. Plaid or even the Communists in local by-elections (admittedly, recently they seem to have adapted this philosophy to some Regional and Westminster elections too).

One of the most valuable aspects of this site is that contributors often report not only the results of by-elections within individual consitutencies, but also their causes and issues arising. These observations can be used to adjust, WITHIN INDIVIDUAL SEATS, predictions made from national polling as to what the swing will be, and hence who will win the seat.

Interpretation of local by-election results by psephologists would be greatly facilitiated if the pollsters conducted, on the lines of the poll last Autumn, a large scale survey of Westminster voting intentions at the time of the ordinary council elections, using samples large enough to predict for individual seats (at any rate marginals). More generally, given the number of polls recently that have produced similar results, this would fit a suspicion I have that we would get better value if the pollsters conducted fewer surveys but with sample sizes enabling more detailed analysis.

P.S. Whilst I no longer belong to a party, I have sympathy from past experience with Anon Activist’s observations about some councillors who do not slog around the streets, although this only applies in a minority of cases.

Thanks for the research Anthony.

so just as well local elections don’t reflect general election results.