Partitioning menus to nudge single-item choice

Decision makers are often called on to choose a single item from a menu of options: employees pick health plans from provider-sponsored lists, citizens vote for a representative among candidates, physicians select a treatment from an order set. Because options are often organized or grouped into subsets, the question arises whether the grouping of menu items affects the options ultimately selected by decision makers. Across ten experiments, we provide robust evidence of single-item partition dependence — decision makers are more likely to choose from response categories that are more finely partitioned, holding the underlying set of options constant. For example, an employee choosing a health plan from a menu that lists HMOs individually but groups PPOs into a single category with sub-options may be biased toward choosing an HMO. Unlike prior work on multi-item allocation decisions, the traditional explanation of partition dependence — a bias toward even allocation — cannot apply to single-item choice, because singular choices are not divisible. Instead, we find that finer partitions receive greater choice share because decision makers infer that separately listed options are more popular on average than grouped ones. More broadly, strategic partitioning of the menu space may serve as a simple and effective tool for managers, policymakers, and choice architects: listing options separately rather than grouping them together tends to increase their choice share, without fundamentally altering the underlying set of alternatives or the effort required to choose among them.