Research

Below are links to some recent papers that include research on attitudes and beliefs about free will, determinism, and moral responsibility (my thanks to the Garden of Forking Paths for listing these). Ideally, a large cross-sectional population survey on free will should be conducted to see how beliefs about it vary by demographics, economic strata, education, and religious affiliation.

John Monterosso, Edward B. Royzman, Barry Schwartz, "Explaining away responsibility: Effects of scientific explanation on perceived culpability", forthcoming in Ethics and Behavior.
Shaun Nichols and Joshua Knobe, "Moral Responsibility and Determinism: The Cognitive Science of Folk Intuitions" (revised version as of April 2005)
E. A. Nahmias, S. G. Morris, T. Nadelhoffer, and J. Turner, "Is Incompatibilism Intuitive?", forthcoming in Philosophy and Phenomenological Research.

For a preliminary version of a brief survey on free will, click here.

And here's a methodological note on free will research posted to the Garden of Forking Paths:

What’s typically done in research on beliefs and attitudes is a two stage process. First, conduct extensive focus groups and interviews with open-ended questions (e.g., about free will and moral responsibility) in which people get to ramble on at length about the issues in question, without much if any guidance from the moderator or interviewer. The resulting material is then subjected to a rigorous content analysis which extracts the salient features of the belief and attitude concepts, their dimensions, and their variability. Of course it’s important to involve people from different walks of life in order not to get a biased slice of the conceptual variation. And it’s equally important to get more than one person’s take on the content analysis. What this stage does, ideally, is get a relatively unbiased empirical assessment of the landscape of belief and attitudes.

Then, on the basis of the content analysis, questionnaires are constructed using yes/no, multiple choice, and Likert scale type questions (strongly disagree, disagree, agree, strongly disagree) which can efficiently test for variation on the concepts that have been elicited in the focus groups and interviews. [for an example, click here] The data are then analyzed to see if the questions capture consistent patterns of variation that indicate real dimensions of belief, and if so, what these are best interpreted as being (e.g. belief in contra-causal freedom vs. a desire that people be held accountable). When someone asks you to justify your close-ended (as opposed to open-ended) questionnaire items, you can point to the earlier stage of research and give an empirical rationale for phrasing the questions the way you did. But of course there’s always going to be room for interpretive disagreement at all stages of this sort of research.

It’s great that Eddy Nahmias and others are delving into the empirical question of what people really think about free will. But I wouldn’t be surprised if people’s beliefs are a mess of contradictory intuitions that will take some pretty cagey research to untangle. [And indeed, that's what the papers listed above seem to show.]