1.5 The Excess Choice Effect
1.5.1 Evidence Against the ECE
However, the universality of the ECE has been challenged. Some
researchers have found no evidence of the ECE. For example, in Arunachalam et al.,’s (2009) experiment, participants were found to be significantly more likely to opt for a soda over cash as payment for participation in a study if they were shown an extensive selection of 24 different flavours, rather than a limited choice of just 6 flavours. The authors state this demonstrates an ‘anti-ECE’, by
supporting the standard economic model that more choice is beneficial to chooser motivation.
Berger, Draganska and Simonson (2007) also found no evidence for the ECE. In their experiment, participants were asked to select the chocolate they would most like to sample from one of two different brands. One brand contained an intermediate (10) selection of chocolates, whilst the second brand contained an extensive (30) number of options. After choosing and sampling a chocolate, participants rated perceived quality of the product and brand, as well as their likelihood of purchase. The authors found that despite the fact that participants rated choosing from the extensive choice as more difficult and frustrating; the majority of participants were attracted to the brand offering the extensive, rather than the limited selection of chocolates. In addition, quality ratings, and likelihood of purchase were both found to be higher for the brand offering the extensive selection of chocolates, in comparison to the intermediate choice brand. The authors’ interpret this as demonstrating that the size of a brand’s assortment may be seen as a quality cue.
Kahn and Wansink (2004) found that assortment size, as well as the perceived variety and structure of that assortment could influence consumption levels. In their experiment, participants were offered jelly beans to eat, from either a limited (6 colours) or an extensive (24 colours) selection. These assortments were either organised by colour, or scrambled altogether (disorganised
selection). Kahn and Wansink (2004) observed that consumption levels increased as choice level increased, for organised assortments (no difference was found for disorganised assortments). However it may be questionable to assume that increased consumption necessarily equates to increased
satisfaction. This point was demonstrated by Reibstein, Youngblood, and Fromkin (1975) who also found that increased choice lead to increased
consumption. In their experiment, participants were asked to select which soft drink they would like to try from either a limited (2) or an increased (4) selection of flavours. Participants’ consumption and satisfaction levels were then examined. The authors found that participants who chose from the increased selection drank more of their chosen drink, than participants who chose from the more limited selection. However, even though there were increases in consumption levels, no differences were found in terms of satisfaction levels. As such this raises doubt over the use of either of these two studies showing increased consumption with increased choice as providing evidence against the ECE, even though the former was considered relevant in Scheibehenne, Greifeneder and Todd’s (2010) meta- analyses (discussed momentarily below).
Scheibehenne et al., (2009) found no evidence for the ECE except in a study which called for choice justifications. The authors conducted a series of studies to examine potential moderators of the ECE. In the first study option
variety was manipulated. Participants read descriptions of either a large (30) or small (5) variety of restaurants, and were then asked whether they would prefer cash or a voucher to dine at one of the restaurants. The authors found no evidence of the ECE, with participants in both large and small variety conditions being equally likely to opt for a restaurant voucher over cash. The authors then examined the moderating impact of prior preferences, cultural differences, further increases in choice set size and choice justification on the ECE. Of these four manipulations only choice justification was found to elicit an ECE. The authors claim this fits with the assumption that justification becomes more difficult when options become more difficult to distinguish from one another – which naturally increases as assortments grow larger. This assumption is furthered supported by Sela, Berger and Liu (2009), who found that because choosing from larger
assortment sizes is more difficult, people will select options which are easiest to justify. In their experiment, participants presented with a large selection of 10 ice cream flavours were much more likely to select a reduced fat option than
participants who were presented with a limited selection of only 2 options. The authors suggest that this is because when presented with increased choice, people will choose options that are easier to justify.
Finally, Scheibehenne et al., (2010) conducted a meta-analysis of 50 studies which had examined the ECE, and found an overall effect size of virtually zero, meaning that no evidence was found for a reliable occurrence of the ECE across the choice literature. Nevertheless the large degree of variance in study outcomes led the authors to conclude that a theoretical account which could help to cover the divergent findings was needed. It is suggested that this may depend on the prevalence of particular moderator variables, such as a call for post-
decisional justifications (see Scheibehenne et al. 2009), and the interaction between the structure of the choice assortment and decision processes. For example, this could involve the interaction between choice complexity, in terms of the number of attributes that alternatives are differentiated on (see for example, Greifeneder et al., 2010), with individual differences in the degree to which decision makers make use of complex versus heuristic decision processes (see for example, Scheibehenne & Todd, 2009).
The main aim of the current research was to investigate the potentially causal role of counterfactual thought in driving the ECE. The rationale for this is provided in Section 1.6 below. If supportive evidence is found for this overarching hypothesis, it is possible that this in turn may go some way to explaining some of the divergent findings noted in Scheibehenne et al.,’s (2010) meta-analyses. Specifically, it may be the case that variability in the experimental circumstances of studies considered in the meta-analyses may have made counterfactual alternatives more or less readily available. As such, via predicted influence on counterfactual generation, it is possible that there are certain situational factors which may be influential in determining the prevalence of the ECE. Evidence for factors known to influence the availability of counterfactual thoughts, alongside the predicted impact of each upon the ECE is considered in Sections 1.8 and 1.9 below, whilst the potential for the current research to explain the divergent
findings noted in Scheibehenne et al.,’s (2010) meta-analyses is a point to which I will return in greater detail in Chapter 6.