1 The degree of change in voting behaviour between elections. There is a distinction between net volatility and total volatility. Net volatility relates to the change in the shares of the vote for each party. It is usually measured by the index of dissimilarity, or Pedersen index, which is the sum of the absolute changes in vote shares divided by 2. This gives a lower bound on the proportion of voters who changed their votes, assuming a stable voting population. Total or overall volatility is the total proportion who changed between the two elections. This can be measured with respect to those who voted at both elections, those who were eligible at both elections, or those who were eligible at either election. Since the 1960s in Britain overall volatility has been relatively stable despite large changes in net volatility.
2 The idea that voters have become more willing to switch between parties. This is related to the dealignment thesis that voters are no longer consistent strong party identifiers with correspondingly stable voting behaviour, but free to choose who to vote for depending on the issues of the day.