School choice can only thrive when families are well-informed about their options. Unfortunately, good information about different schools in a district can be hard to come by, and even when it is accessible, how it is presented and the level of detail provided can heavily influence the decisions parents make. To understand how to display data to nudge parents, particularly low-income parents, in the direction of high-performing schools, researchers at Mathematica Policy Research designed a series of school-data presentations and tested their effects on parents’ selections.
For the experiment, the authors designed a hypothetical district with sixteen different schools, each of which had different strengths and weaknesses, and asked 35,000 low-income parents to select a school for their child to attend based on online displays of information about each. The schools and their characteristics remained constant, while the displays varied across five factors: numbers-only format versus a visual one; whether a district average was included for comparison; whether parent opinions were included; low- or high-information summaries; and whether the default sort order was by distance from home or by academic performance. After evaluating the schools, parents picked their top three and answered a few questions about their final choices and the usability of the platform. The authors then estimated the effects of each factor on a parent’s decision.
Of the five factors, the simple default order of the list most strongly affected parents’ choice of school. When the options were listed in order of academic achievement, parents chose academically stronger schools (about 5 percentile points higher) that were farther away from home (about 0.6 miles farther). Even when parents could manually re-sort the list by a number of different criteria, the default order was still the most powerful influence on the outcome.
Other factors had significant effects, too. For example, using icons to represent academic performance and safety skewed choices towards schools with higher ratings in those areas. High-information displays, which cited twenty-two data points instead of four, encouraged parents to choose better-resourced schools. And when ratings from other parents appeared in the display, safety ratings had a stronger effect on the ultimate choice.
Overall, the best display model for encouraging parents to pick high-performing schools sorted the list of schools by performance, displayed a limited amount of information on each school, used icons to represent academics and safety, and included ratings from other parents and a district average for comparison. Using this model, parents picked schools with academics nineteen points higher than those selected by parents using the worst, or opposite, model.
The experiment has its limitations, though. Asking parents about a hypothetical school district cannot fully replicate a real school choice environment, where parents may already have developed good or bad impressions about particular schools from their neighbors, the news, or advertising. The stakes of picking a fake school are also lower, and parents may be less swayed by external factors when their child’s actual education is on the line.
Even if each factor’s influence is relatively small, the effects are cumulative and can add up to make a big difference in a family’s final decision. That, the authors explain, is why this study is so important to improving the school choice environment, especially for low-income families. There’s no such thing as a no-nudge display, so reform advocates should seek to understand which nudges lead to the best outcomes for students.
SOURCE: Steven Glazerman et. al, “Nudging Parents to Choose Better Schools: The Importance of School Choice Architecture.” Mathematica Policy Research (November 2018).
