We have a theory for how moral opinions change that was recently published1. According to this theory, liberalization of moral opinions should be almost inevitable as long as people are free to argue their views.
Our theory is based on two things: that different opinions rely on different kinds of arguments such as fairness considerations, tradition or religious beliefs; and that people sometimes exchange arguments about moral issues and are more likely to be swayed by an argument that they find intuitively acceptable.
So what arguments do people find acceptable? To answer this, we turn to research in moral psychology on the moral arguments for which we find a stark difference between conservatives and liberals. Conservatives tend to be equally accepting of many kinds of moral arguments (including appeals to authority, loyalty, and purity but also harm and fairness), whereas liberals mainly tend to accept moral arguments that appeal to harm and fairness (universal arguments)2.
Taken together, this means that for any issue liberals will tend to move towards the position with more harm-fairness support. Conservatives on the other hand are equally likely to be swayed by all arguments and will therefore move towards the average opinion in the population as this reflects how often they hear arguments for any given position. For each position (eg. being in favor of gay marriage) we calculated its harm-fairness connection advantage. By this we mean its harm-fairness support compared to the harm-fairness support of the opposite position (eg. being against gay marriage). This logic is spelled out in a formal model in our paper and results in three specific predictions.
Prediction 1. On any issue, the liberal position should tend to be the position that has a harm-fairness connection advantage.
Prediction 2. Public opinion should trend toward the position with harm-fairness connection advantage, that is, the liberal position.
Prediction 3. The stronger the harm-fairness connection advantage a position has, the faster public opinion should move towards that position.
All of these predictions found strong support in the data but the format did not allow us to go through the results issue by issue so we have now published a tool that allows you to interactively investigate how well the predictions fit each issue.
1Strimling, P., Vartanova, I., Jansson, F., & Eriksson, K. (2019). The connection between moral positions and moral arguments drives opinion change. Nature human behaviour.
2Graham, Haidt, & Nosek 2009; Haidt & Graham 2007.