The underlying scientific results should be taken with a grain of salt because of the reproducibility crisis in psychology, Gary Marcus’s and Annie Duke’s Wall Street Journal article “The Problem with Believing What We’re Told” are otherwise a fascinating rundown of indications that people are often quite sloppy about deciding whether something is true or false.

On that grain of salt, as I discuss in “Let's Set Half a Percent as the Standard for Statistical Significance," the article "Redefine Statistical Significance" on Psyarchive notes that—of results in psychology that according to the author’s statements, supposedly had only between 1/2 % and 5% of being due to chance, only 24% could be replicated in a follow-up study designed to verify those results.

But with that grain of salt, following are the suggestive results of psychological research on discernment of true/false as summarized by Gary Marcus and Annie Duke. I have added bullets to their words to separate different passages:

• The simple act of repeating a lie can make it seem like truth … Test subjects became more likely to believe things as they were repeated, regardless of whether they were true or false. The third time they heard a false statement, they were just as likely to believe it as a true statement that they heard once.

• When pictures were attached, people were more likely to believe the statements, including the fake ones.

• … the presence of moral and emotional words like “hate,” “destroy” or “blame” acted like an accelerant, increasing the chance that a message would spread by about 20% for each additional emotional word. The study also found that most of the sharing was done within political parties rather than across political divides, creating an echo-chamber effect.

• Savvy propagandists have long exploited the tendency of the human brain to take shortcuts. But social networks make it far easier, because they feed on a further human vulnerability: our need for approval, affection and positive feedback.

• When first asked to assess the believability of true and false headlines posted on social media, the 68 participants—a mix of Democrats, Republicans and independents—were more likely to believe stories that confirmed their own prior views. But a simple intervention had an effect: asking participants to rate the truthfulness of the headlines. That tiny bit of critical reflection mattered, and it even extended to other articles that the participants hadn’t been asked to rate. The results suggest that just asking yourself, “Is what I just learned true?” could be a valuable habit.

• … prompting people to consider why their beliefs might not be true leads them to think more accurately.

If borne out, this set of results provides a road map for creating successful fake news—and a road map for not being one of the people who is snookered.