The study says that libertarians are "libertarians were moderately more utilitarian than conservatives, and slightly more utilitarian than liberals." But they also value liberty as a terminal value more than liberals or conservatives. I don't see how this can be reconciled.

Assume conservatives and liberals have other terminal values besides liberty. Even if they value liberty less, other terminal goals can still pull them further away from utilitarian positions.

If liberty (however defined) is a highly weighted term in a utility function, or rather in whatever approximation of a utility function people actually use to do consequential reasoning, then we can reconcile a consequentialist worldview with prioritizing it above other values. Generally when people talk about utilitarianism they mean happiness/suffering utilitarianism, but I can forgive a little imprecision in a news article.

Most people don't have a systematic moral theory, where everything follows consistently from a small number of fundamental principles. Instead, they have a bundle of intuitions, emotions, principles, etc., which they use to varying extents in various situations. "More utilitarian" just means that they make moral judgments consistent with utilitarianism a larger percent of the time (which is consistent with having other terminal values besides well-being).

This study used several trolley-type problems, and found that libertarians were somewhat more likely to approve of flipping the switch or pushing the fat man (or throwing a person overboard from a lifeboat that is sinking because it has too many people on board). These dilemmas come from the utilitarianism vs. deontology debate, which suggests that libertarians might rely on cost-benefit reasoning more often and rights- or rules-based reasoning less often, but that is not necessarily true. Joshua Greene's take on these dilemmas is that the utilitarian response is based more on calculating reasoning (System 2) while the non-utilitarian response is based more on emotional aversions (System 1). So this result may just fit with the storyline that libertarians rely more on reasoning and less on emotions, with cost-benefit reasoning among the types of reasoning that they use more of.

The confusion might come from thinking that libertarians are a homogenous group. To make a generalization: some groups of libertarians do value liberty as an end in itself, or the ultimate moral value; others see it as the logical extension of acting on a utilitarian calculus. Given my previous involved in the libertarian community and what I know about multi-system moral psychology, this seems plausible, if not obvious. [PDF warning]

But to an outsider researcher not already familiar with the libertarian movement/mindset, I can see how they wouldn't know that and how that lack of knowledge could skew the data into the contradiction you identified.

Within libertarian circles, I've seen folks differentiate between "capital L" libertarians who think liberty is an end in itself and "lower case l" libertarians who are libertarian on utilitarian grounds.

The study says that libertarians are "libertarians were moderately more utilitarian than conservatives, and slightly more utilitarian than liberals." But they also value liberty as a terminal value more than liberals or conservatives. I don't see how this can be reconciled.

It can be reconciled if there are several other values aside from liberty that liberals and/or conservatives value and libertarians don't. (As the article suggests.)

Hey, if they could spot those kinds of contradictions, they'd be libertarians.