Machine learning, optimism, and ethics
Artificial intelligence, machine learning, algorithms—data-powered technology is embedded all around us, and its influence and use will likely only grow from here. We’ve featured plenty of pieces on these types of technology (check out this recent piece and this one from our archive), but this was a new application even for us. In this study, researchers trained a deep-learning model to comb through decades of World Values Survey data and then generate novel hypotheses about antecedents of unethical behavior. Based on that data, the model came up with the hypothesis that optimism about the future would predict less unethical behavior. The researchers tested this hypothesis in small correlational and experimental studies, both of which provided support. That means that this paper has two novel, though unrelated, preliminary contributions: the finding that optimism might help reduce unethical behavior and the utility of machine-learning methods to generate novel hypotheses. [Psychological Science]
Analyzing twitter politics en masse
In other data- and tech-related research news, scientists have created the Twitter Parliamentarian Database (TPD), an attempt at a rigorously vetted repository that contains the tweets of all tweeting members from 27 different legislative bodies. Social media is a powerful new tool for researchers, but these authors argue that social science research lacks rigorous standards for using these platforms. By compiling and validating Twitter-centric data from multiple sources, the TPD could make it “possible to compare different countries, political parties, political party families, and different kinds of democracies.” This article’s authors took their new database out for a research spin and examined differences between countries in how lawmakers interact on Twitter, differences between political parties in their use of hashtags, and how elected officials interact during transnational debates. [PLoS One]
Psychological needs for me but not for thee
It’s a fundamental idea from the early days of psychology: physical needs are more important than psychological ones. While that could be the case (empirical evidence for Maslow’s hierarchical organization of our needs is mixed), researchers argue that diminishing the importance of others’ psychological needs in comparison to physical ones is demeaning—and potentially even dehumanizing. That, explains the authors of a new study, is “because a sophisticated humanlike mind is required for an agent to be motivated by psychological needs, whereas any agent with a body—including nonhuman animals—can be motivated by physical needs.” The authors used the example of someone taking umbrage when people without housing seek dignity or respect in addition to a place to sleep. Across six studies, this research found initial evidence for this hypothesis, and found that people tend to demean peers’ needs more than their own by “inferring that one’s own behavior is guided more strongly by psychological needs than identical behavior in others.” [Journal of Personality and Social Psychology]
Do no harm . . . unless there are big benefits
Research has shown that the overwhelming majority of people are harm averse—that is, they make choices with the intention of doing the least amount of harm. But what happens when harm can’t be avoided? New research suggests that when decision-makers must choose between committing less harm for less benefit (think destroying a small part of the rainforest to feed a small number of hungry people) and committing more harm for more benefit (bigger chunk of the rainforest gone, but feeding more hungry people), aversion to harm diminishes significantly and they choose the latter. The authors interpret their findings to mean that when individuals can no longer completely avoid violating the “do no harm” principle, they may prefer more harm so that the violation of a moral principle is “worth it.” [Psychological Science]
Does the door-in-the-face technique replicate?
A team of researchers report in a working paper that they were able to successfully replicate Cialdini’s famous 1975 study documenting the door-in-the-face technique—whereby someone becomes more likely to agree to a request if they’re first presented with an extreme request that they decline.
In the case of Cialdini’s study, participants were first asked if they’d volunteer for a delinquent youth counseling program two hours per week for a minimum of two years (which participants would presumably see as an extreme request and reject). Then they were asked if they’d be willing to chaperone a group of delinquent youth to the zoo for two hours. In Cialdini’s study those that had been first presented with the extreme request were more likely to agree to the smaller request than those who simply received the smaller request.
The current study replicated the effect in a German population 55 years after the original study, which provides evidence that the effect may hold up across time and in different cultures. That said, their sample was still very limited, so further replication attempts in more diverse populations could shed light on the strengths or limitations of the effect. [Working paper]
High rates of replication are possible in the social and behavioral sciences
More rigorous methods do lead to higher rates of replication, argue the authors of a new working paper. Researchers from four labs teamed up to replicate a sample of one another’s work over the course of five years. They employed the most rigorous methods at their disposal—including preregistration, methodological transparency, and high statistical power. They achieved a replication rate of 86 percent and found that effect sizes were largely consistent from the original study to the replications. This indicates that low replication rates are not a feature of social and behavioral science research, but the result of a combination of factors that diminish the quality of research—like publication bias, p-hacking, and low statistical power. This is encouraging evidence that the call for more rigorous and open science can achieve its aim of improving the quality and reliability of science. [Working paper]
Doing science during COVID: hardest for female, lab-dependent, and parent scientists
“COVID-19 has not affected all scientists equally. A survey of principal investigators indicates that female scientists, those in the ‘bench sciences’ and, especially, scientists with young children experienced a substantial decline in time devoted to research. This could have important short- and longer-term effects on their careers, which institution leaders and funders need to address carefully.” [Nature Human Behavior]
Gender gaps in psychological science
While the pandemic might have made things harder for female scientists, disparities certainly existed beforehand. Although “women make up a large and growing proportion of today’s psychological scientists,” these gaps—publishing less, being cited less, and earning less, among others—still exist in psychology. Researchers systematically reviewed ten issues that are relevant for career advancement and argue that the field of psychological science is uniquely positioned to address them, in large part because it is a field that focuses on studying and intervening in human behavior. [Perspectives on Psychological Science]
That intervention talks the talk, but could it walk the walk?
Psychologists Bharathy Premachandra and Niel Lewis Jr. reviewed two decades of academic articles published across five broad categories of psychological interventions—self-affirmation, identity-based motivation, growth mindset, utility value, and belonging—and found that the majority of these studies provided only about two-thirds of the information you’d need to implement the intervention.
“Many interventions are developed and reported in our journals but do not make their way into the broader world they were designed to change,” they write.
“Though it is wonderful for psychologists to develop and test interventions and publish them in our peer-reviewed journals, if we do not share the information that is required by implementors, the immense potential our work has for fostering change will not be realized.” [Perspectives on Psychological Science, forthcoming]
“Reflections on the past two decades of neuroscience”
To mark its twentieth year, Nature Reviews Neuroscience reached out to neuroscientists for their reflections and on how the field has developed over the past two decades. Over a dozen neuroscientists weigh in on what’s changed, the big questions on their mind, and what they’re most excited about for the future. [Nature Reviews Neuroscience]