Sentiment Analysis as a Means to Estimate Individual Ministerial Popularity

Students of ministerial elites have dedicated considerable energy examining the relationship between career paths and public opinion. Taking a broad swash, one can simply ask whether ministerial resignations affect government survival. A more focused lens may pinpoint the precise effects of turnover on government popularity. Indirectly, government and/or leader satisfaction is a reasonable control to include in models of ministerial turnover. However, existing research is not able to pinpoint the effects of ministerial satisfaction on ministerial turnover because measures of individual level ministerial satisfaction seldom exist. In this paper we try to overcome this hurdle by developing a measure of ministerial satisfaction using sentiment analysis, a natural language processing technique typically used in market research analysis. We apply sentiment analysis to the most recent Canadian ministry and discuss results and implications for further application.