Pity the political analysts who cover special elections. A sample size of one House of Representatives race makes for a challenging exercise in interpretation. Yet, depending on the result, the race can be spun as a negative referendum on the president or a confirmation of his policy decisions.
Despite their relatively low predictive power, special elections are worth a second look. In particular, I will show that resignations, which have caused 101 (55.2 percent) of special elections since 1970, have crept up markedly in the past five years. For this article, I use a dataset of 183 special elections for the House of Representatives (all elections held from 1970 onward). [cont.]
Kristin Blagg, Georgetown Public Policy Review