The paper, published in the peer-reviewed journal World Development, applied a multivariate regression analysis and logit model to areas of the country to identify why people voted the way they did.
The level of higher education in an area was far more important than age, gender, the number of immigrants, or income in predicting the way an area voted, the researchers found.
Age and gender were both significant but not as important as education level, the researchers found. Income and number of immigrants in an area were not found to be a significant factor in how people voted.