Partition dependence in consumer choice: Perceptual groupings do not reliably shape decisions

Abstract

The partitioning of options into arbitrary categories has been shown to influence decisions about allocating choices or resources among those options; this phenomenon is called “partition dependence.” While we do not call into question the validity of the partition dependence phenomenon in the present work, we do examine the robustness of one of the experimental paradigms reported by Fox, Ratner, and Lieb (2005, Study 4). In three experiments (N = 300) conducted here, participants chose from a menu of perceptually partitioned options (varieties of candy distributed across bowls). We found no clear evidence of partition dependent choice in children (Experiment 1), and no evidence at all of partition dependence in adults’ choices (Experiments 1-3). This was true even when methods were closely matched to those of Fox et al.’s Study 4 (Experiment 3). We conclude that the “candy-bowl” choice task does not reliably elicit partition dependence and propose possible explanations for the discrepancy between these findings and prior reports. Future work will explore the conditions under which partition dependence in consumer choice does reliably arise.

Partition dependence in development: Are children’s decisions shaped by the arbitrary grouping of options?

“…Previous work in our lab used the candy choice paradigm to test for partition dependence in children's choices. However, not only did we fail to find evidence of partition dependent choice in children, but in three experiments we did not replicate the previously reported findings of adult partition dependence in this task (Reichelson, Zax, Bass, Patalano, & Barth, 2017). This work shows that the candy choice task itself does not reliably elicit partition dependence, even in adults-leaving open the question of whether partition dependence in decision making occurs in children.…”

“…Most findings of partition dependence concern complex, formal decision situations. Perhaps partition dependence is less likely in the decision situations available to children (an idea that finds some support in the repeated lack of partition dependence in both children's and adults' choices in candy choice tasks; Reichelson et al, 2017). It is also possible that attention to higher order partitions (vs. individual choice options) does not occur early in life.…”

“…Participants were asked to help a puppet distribute 12 tokens representing treats to four different types of zoo animals arrayed across three boxes. Two boxes were each arbitrarily associated with one type of animal, while the third box was associated with two animals (Figure 1), to create an arbitrary threeway partition of containers and a four-way partition of specific elements associated with the containers (as in previous consumer choice studies , Fox, Ratner, et al, 2005;Reichelson et al, 2017). There were therefore two "singleanimal boxes" and one "double-animal box," with counterbalancing of the specific animal types appearing across these box types.…”

Partition dependence in financial aid distribution to income categories

“…Simple choices of candies from an arrangement of bowls were reported to exhibit partition dependence in one study [1], but no such effect was found in three recent replication attempts, including one with very closely matched methods [11]. None of the established factors known to reduce partition dependence appeared to play a role in this study (e.g., expertise or greater intrinsic preferences; [1,12]), showing that physical partitions like those utilized in the candy bowl study did not reliably influence consumer choice [11]. Simple selections for consumption from menus of physically partitioned options, therefore, do not necessarily produce partition dependent behavior as had previously been reported.…”

“…One goal of the present work, therefore, is to replicate the findings of a previous study that reported partition dependent behavior. The consumer choice studies in which partition dependence findings were not replicated [11] made use of a relatively simple, straightforward selection task. For the present work, we aimed to move from consumer choice to an allocation task, while continuing to use a relatively simple, straightforward paradigm to attempt to replicate findings of partition dependence in adults' decisions.…”