You would have to redo the study and make sure to compare liberal and conservative voters happiness levels. Not everyone in Alaska, Utah, and Wyoming are conservative voters. If they polled people in big cities for example, it would be more likely that most of the respondents are liberal