A quantitative analysis of the influence of unemployment on the results of parliamentary and presidential election in Poland (1993–2001) is presented in the paper. The statistical data on province level is used. Election results were published by the State Election Committee, and the unemployment rate was taken from the Central Statistical Office publications. For the parliamentary election only the lower chamber (Sejm) results were considered. Presidential election results are analyzed on the basis of the first round. Correlation and regression analysis were used in the statistical analysis. The coefficient of the relative importance has been proposed. The significance of relations has been tested with proper statistical tests. It has been found that the unemployment was an important factor which influenced the election results within analyzed period. Generally speaking, left-wing parties and candidates, are those who “gained” from the unemployment while the right-wingers appeared to be the losers. A special „historical” mentality can be one of the reasons. People have the tendency to remember rather good things from the socialist past, like full employment and social security (not counting their actual level). These thoughts are transfer to the favor of political groups connected with the past. There was no evidence that being a ruling party is better (or worse) that being an opposition, as far as the influence of unemployment is concerned. Maybe there is another (latent) factor which influences both unemployment and election results giving artificial correlations. Observed symptoms of non-linear relation between unemployment and SLD popularity, in case of growing unemployment can result in shifting more votes toward “new” populist left-wingers.