People often simplify complex political choices by giving decisive weight to a single characteristic, according to a new study by social scientists at UC Berkeley. In experimental choices involving immigration, climate policy and candidates for the US Senate, the researchers identified the use of mental shortcuts in 68% of cases. The research, published in Proceedings of the National Academy of Sciences, primarily examines how preferences take shape, rather than which political position prevails.
Nearly 3,800 people took part in the study, choosing on a computer between pairs of hypothetical immigrants, candidates or climate policy packages. Each option was described by six characteristics, combined at random. For immigrants, these were region of origin, education, gender, age, English proficiency and whether they were seeking economic opportunities or asylum. This section presented 30 different pairs, allowing repeated choices to reveal which characteristics most influenced each participant.
Political scientist Kirk Bansak and cognitive science researcher Nidhi Banavar developed a model to distinguish between different decision-making strategies. The tool does not directly observe participants' thoughts: it infers their likely approach to evaluation from patterns in their choices. If, for example, someone consistently chose the younger immigrant or the one with better English, the researchers could identify the particular weight given to that factor. A similar approach was applied to the other two topics.
Nearly 54% of cases showed a choice pattern dominated by one of the six factors. In another 14%, participants appeared to count how many characteristics they preferred in each option and choose the one with the most positive attributes. Just under 30% involved an overall assessment of all six characteristics, without the shortcuts examined by the model. These differences show that the same final preference can emerge from different ways of processing information.
Banavar points out that using shortcuts does not automatically mean people are irrational. The concept of rationality used in the study requires computational abilities different from those we usually use in everyday life. In addition, most participants did not consistently follow the same strategy across all topics: someone might use a single criterion to decide on climate issues but weigh more factors when considering immigration. The reasons for this shift require further research, she says.
The research also recorded certain preferences within the experiment: participants preferred younger political candidates and climate policies that limited the financial burden on their households. In the immigration choices, they preferred people seeking asylum over those seeking economic opportunities, as well as people below retirement age. These preferences, however, were not the study's main focus. The central question was how much information participants used and how they combined it.
The design has limitations: each characteristic had only two possible values, such as male or female, white or Black, or an age of around 50 or 70. The researchers explain that more categories would have made the cognitive model harder to use and the patterns in the data less distinct. They also left out the hypothetical candidates' party affiliation because it could trigger strong, immediate reactions that would overshadow the other characteristics. When both options had the preferred characteristic, participants had to consider other factors, while rating each profile beforehand made them more willing to weigh all the factors.
For the authors, the findings matter for opinion polls and political communication because choosing between complex proposals does not prove that a respondent assessed every aspect of them. If, for example, concern about household finances dominates decisions on climate policy, this may influence how proposals are presented and how much attention is paid to their scientific dimension. The study thus questions the assumption that access to information is enough for a full assessment of political issues. The researchers argue that understanding public opinion requires knowing both what citizens choose and how they arrive at their choices.
Comments