When it comes to moral decisions, testosterone doesn’t seem to do much

Image showing actors at the controls of a trolley.
Enlarge / The trolley problem brought to (after) life.
NBC

The Trolley Problem is a staple of ethics courses and has even made its way into prime-time television. It’s a problem because it forces people to decide between two options that are both considered highly ethical: not choosing to kill someone, and minimizing the total number of deaths. It even has real world correlates, like whether it’s better to shoot down a hijacked airliner filled with innocents or allow it to be used as a weapon that kills even more people.

Some of us would like to think that we’d be able to step back and evaluate the situation dispassionately, but the reality is that our emotions often drive important decisions (and besides, as the clip from The Good Place linked above shows, there’s not always time for careful evaluation). Since testosterone influences both emotions and decision-making, many people had ideas about how it might alter the decisions made by people weighing these moral issues. But when a team of researchers from the University of Texas, Austin, decided to test those ideas, it turned out none of them was right.

That doesn’t mean testosterone does nothing, but it certainly indicates we don’t understand what it might do.

Planned in advance

There has been a lot of discussion about the issues of reproducible research, p-hacking for any publishable results in data, and an inability to publish negative results. So it’s a pleasure to report on a paper that actually does something about it. The authors—Skylar Brannon, Sarah Carr, Ellie Shuo Jin, Robert Josephs, and Bertram Gawronski—actually pre-registered their research plan, including the hypotheses they were going to be testing and the experiment that would test them.

The hypotheses were based on previous suggestions that testosterone would make people less empathetic and therefore more likely to choose a purely utilitarian solution, minimizing the total number of deaths. They also hypothesized that people given testosterone would prefer taking action, which would bias them toward changing the status quo. To test these hypotheses, they got 200 volunteers and randomly assigned some to receive a boost of testosterone, while the rest got a placebo. The participants were then asked to consider a moral dilemma and come to some decision on what they felt the right course of action was.

Rather than simply looking at what the subjects chose, the researchers recognized that there were several potential tiers of decision-making at play. These included a general preference for action vs. inaction and possible preferences for moral prohibitions or utilitarian views. The participants’ responses were plugged in to a model that determined where they fell on these various values. The outcome of that analysis was used to test the various hypotheses.

“The results of the current study yielded no evidence in support of any of the four preregistered hypotheses,” write the authors. Those given testosterone were no more likely to prefer taking action than those given the placebo. It was predicted that they’d also be more prone to make utilitarian judgements that minimize total casualties; this wasn’t true, either. Another hypothesis suggested that those given testosterone would be less sensitive to moral prohibitions, such as not choosing to kill someone. In fact, the results suggest the exact opposite is true.

Before and after

While the experiments nicely wrecked all the researchers’ expectations, the paper didn’t end there. The researchers had also gathered some additional information on most of their participants and decided to do what they termed an “exploratory” experiment. Most of the results of this experiment aren’t especially interesting. For example, the subjects weren’t successful in determining whether they got a placebo or the real thing, and women and men were mostly indistinguishable in these experiments (women had a slight preference for respecting moral prohibitions). The one exploratory result that did stand out was when the researchers looked at pre-experiment testosterone levels.

To do this analysis, they handled males and females separately, since they have different levels of the hormone, and divided each into high and low testosterone groups. The high testosterone group in this case showed a weaker affinity for moral prohibitions—the exact opposite of what was seen when they were given additional testosterone.

Given this apparently contradictory result and the complete absence of other significant changes, it’s fair to say that we don’t know if testosterone is involved in this sort of decision-making, much less what its impact is if it is. It’s not even entirely clear that more work is warranted, given that there are more compelling indications of effects that might be better to follow up on.

On the plus side, however, the paper does show the advantage of pre-registering experimental designs. The fact that the hypotheses were out there compelled the authors to be up front about them not being supported and undoubtedly contributed to their ability to get these negative results published in a high-profile journal. The only downside is that rather than focusing on the negative results, the authors’ title happily describes the one significant result they got from the main experiments, even though it falsified their ideas.

Nature Human Behavior, 2019. DOI: 10.1038/s41562-019-0641-3  (About DOIs).

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *