A brief overview of coin flips

Coin flips or other decision-making aids with random outcomes have been studied before, but in a more classical situation such as the old standing practice in soccer games to determine which team will play the first half of the game on which side and who will kick the ball first [1]. In this classic situation, the outcome of the coin flip determines the decision, and individuals cannot decide on their own.

In this context, Keren and Teigen [1] investigated individuals’ willingness to toss a coin in situations of indeterminacy, where there are no decisive arguments that favor one alternative over another. The authors focus on the finding that for trivial decisions a coin flip is considered an efficient and fair means of making decisions, but when the outcome is more important, a decisive coin toss becomes less acceptable, as this approach seemed to conflict with traditional ideas about argument-based rationality and personal responsibility of the decision maker [1].

In situations that are less dramatic and more irrelevant, individuals may instead choose to use a coin flip to delegate the choice to randomness [2] or to avoid being responsible for implementing inequity [3]. In a more applied setting, Levitt [4] conducted a large field experiment and asked participants about an important decision they needed to make. He then let them flip a coin. After two and six months participants were contacted again and asked about their decision and associated satisfaction. His study revealed that individuals who were told by the coin toss to make a change (compared to maintain the status quo) were much more likely to make the change and were happier six months later.

Flipping a coin to help make a decision

When individuals flip a coin, they do not necessarily need to follow the outcome. Sometimes they report liking or disliking the outcome and this feeling then holds information in itself, as individuals follow the coin’s suggestion, when they like it, but go for another option, when they don’t.

To illustrate this theoretical process, we briefly summarize the story of a friend of one of the authors. This friend once had to make a difficult career decision between two jobs after graduating from his law studies. On the one hand, there was a very well-paid position in a prestigious law firm, which would probably require high commitment and a lot of overtime. On the other hand, there was a position in a smaller company, which was not as well-known and also paid less, but allowed for more flexibility in working hours. Both jobs had distinct advantages and disadvantages, making it difficult to come to a conclusion. The friend drafted lists of pros and cons, thoroughly considered the options, and talked to other people. As these strategies did not allow for a clear decision, he simply tossed a coin in the end. The coin suggested going for the well-paid job with long working hours. He disliked the outcome of the coin flip and felt that he actually did not want to take that job. Instead, he decided to take the other job.

But why would a random device, such as a coin flip, lead to feelings of liking or disliking? In brief, we hypothesize that as the coin flip suggests making a certain choice by pointing to one option over the other, the individual focuses on the option the coin is pointing to. More specifically, the individual concretely imagines obtaining this option (taking the prestigious job). This process of mental simulation results in a more vivid representation of the choice. The vivid representation makes previously existing feelings stronger. Like a magnifying glass, the coin flip amplifies positive and negative affect so that previously undetectable differences between positive and negative reactions may become apparent. In our example, these feelings are of negative valence (negative feelings about overtime overpowered the positive ones elicited by the high pay), individuals might dislike the outcome of the coin flip. If, instead, the stronger feelings are of positive valence, individuals may like the outcome of the coin flip. These feelings may then have an impact on subsequent processes, arguably resulting in a clearer preference and decision.

Perhaps the role the random outcome decision-making aid plays in such a situation is best described as a catalyst, since it allows for a decision where individuals were undecided before. We hypothesize that random devices may catalyze decisions by rendering feelings stronger, eventually resulting in more affect-driven judgment and decision processes.

Why a catalyst may elicit feelings that matter in decision-making

Our novel perspective of decision aids as catalysts is different from the classic perspective of using them as deciders, as it highlights that decision makers decide on their own, but are informed by feelings strengthened during the coin flip.

We assume when using a decision-making aid such as a coin flip, the suggestion of one option over the other is not binding, but it may feel very real. Therefore, using a decision-making aid may result in a more vivid representation of the suggested option, which may be linked to stronger feelings. This is consistent with previous research, which has shown that emotion is positively linked to mental imagery [5], and that concrete mental simulations evoke stronger emotional reactions than more abstract ones [6]. Moreover, Hsee and Rottenstreich [7] showed that providing more vivid information, such as pictures compared to text, leads to increased salience of feelings.

Strengthened feelings can then serve as pieces of information in their own right. The feeling-as-information account suggests that such feelings, just like pieces of content information, can be used as information to make decisions [8,9]. In accordance with the “How do I feel about it” heuristic [10,11], individuals can inspect the feelings that the representation of a certain target may elicit when being pursued [12]. If the feelings that were strengthened by vividly imagining obtaining the suggested option are positive, individuals can use this information as an indicator for their actual preference. If, however, the feelings that were strengthened by vividly imagining obtaining the suggested option are negative, individuals can use this information to conclude that they might prefer the alternative option. Therefore, using a catalyst may result in an immediate feeling of which option individuals prefer.

In consequence, depending on whether the strengthened feelings are positive or negative towards the suggested option, feelings such as liking or disliking the decision-making aid’s outcome may result. Positive feelings should result in stronger feelings of liking or approving of the coin flip’s suggestion, while negative feelings should result in stronger feelings of disliking or disapproving of the coin flip’s suggestion. Next, these affective reactions towards the decision-making aid can be used to form a clear preference for one option over the other and therefore allow individuals to come to a decision. Suddenly, they may report having a feeling of knowing what they want.

Summary of studies

To summarize, decision-making aids that produce random outcomes may function not only as deciders (as investigated in other research), but also as catalysts. As decision-making aids point to one option over the other, individuals may vividly imagine and mentally experience obtaining the outcome. By this visual and sensory representation, feelings related to the decision option become stronger and, depending on their valence and the suggestion of the coin flip, result in satisfaction or dissatisfaction with the outcome. Against this background, we suggest:

• Hypothesis 1: Participants using a catalyst (catalyst-participants) compared to participants without a catalyst (control-participants) report stronger affective reactions and feelings of satisfaction (liking) and/or dissatisfaction (disliking)
• Hypothesis 2: These affective reactions are helpful in clarifying one’s preference, and therefore result in a feeling of knowing what one wants.

Two studies that use different catalysts are reported in support of these hypotheses. We received an IRB approval from the IRB of the University of Basel, Faculty of Psychology, with the approval number 001–14 and obtained written consent in both online studies. All data and materials are available under: https://drive.switch.ch/index.php/s/1HzUVPqmdhT1HSq.

Method

Results

To compare answers of participants in the catalyst-condition and control-condition, we calculated t-tests for independent samples. When a significant Levene test indicated that variances were not equal, we report the corrected results.

Discussion

Study 1 provides first experimental evidence that flipping a coin may strengthen feelings. Specifically, catalyst- compared to control-participants reported stronger immediate feelings of which option they would (not) prefer. Furthermore, catalyst-participants reported stronger feelings of disapproval versus approval and dissatisfaction versus satisfaction concerning the outcome of the coin. Complementarily, control-participants were more indifferent to the outcomes of the coin tosses.

Method

Results

To check whether participants understood the set-up, we analyzed whether participants indicated that the die made a suggestion when it actually pointed to one option over the other (manipulation check). Out of 87 participants, 77 answered these questions correctly for all five trials, eight individuals made one mistake, one participant made two mistakes, and one made four mistakes. We checked whether excluding the participants changed the pattern of the results, which was not the case. We thus report the results for the entire sample.

Discussion

Study 2 was conducted to replicate the findings from Study 1 with a different catalyst, a die, to gauge the results’ generalizability. Furthermore, Study 2 was designed to rule out alternative explanations. This was achieved by providing all participants with the decision aid in all trials, yet varying whether the aid provided a suggestion and thus served as a catalyst (catalyst-trials) or not (control-trials).

Results of Study 2 corroborate the hypotheses as participants reported stronger feelings in catalyst- compared to control-trials. Moreover, feelings of satisfaction were reported significantly more often in catalyst-trials, and significantly less often in control-trials, than would be expected by chance. Furthermore, the more suggestions participants received, the more feelings of satisfaction were reported, although our exploratory analyses show that participants did not simply followed the die’s suggestion. This indicates that receiving a recommendation could strengthen feelings but would not necessarily determine the decision. The more suggestions participants received, the fewer feelings of annoyance were reported. This could indicate that using a die that did not recommend an option and therefore was useless was annoying for participants.

Interestingly, feelings of dissatisfaction did not significantly vary between catalyst- and control-trials. At first glance, this appears to be in conflict with our hypothesis that when the die points to the option that feels wrong, dissatisfaction arises. However, dissatisfaction may also arise when the aid does not make a suggestion at all. As a result, dissatisfaction may be present in both catalyst- and control-trials, but for different reasons.

Limitations and options for future research

Studies 1 and 2 provide support for our hypothesis, but leave room for improvement for a set of future research studies. Throughout our work, we aimed to find a particularly strong manipulation of a catalyst by either showing a spinning coin or a rolling die. To further illustrate the suggestion that the device made, the device was displayed above the option it was suggesting. This operationalization however leads to a confound between catalyst trials and alignment of the device above one of the decision options (while in control trials the device is not positioned above one of the decision options). Although we doubt that the simple alignment elicits a stronger immediate feeling of knowing what one wants, it could direct participants’ attention to the decision options, which could alternatively explain why they might have had stronger emotional reactions. Future research might use different visualizations of catalysts that resolve this confound. Having said this, we speculate that the menu options’ exclusiveness captured participants’ attention regardless of whether the catalyst pointed to the options or not.

In our research we investigated the impact of using decision making aids on affective reactions and argue that the changes in affective reactions impact decision making processes. However, one could also wonder whether flipping a coin and looking at its suggestion may trigger more cognitive reactions, too. As we hypothesize that a coin flip leads individuals to imagine the options more vividly, one could assume that they might also think about them in more concrete (instead of abstract) terms. Construal Level Theory [16,17] provides a framework that highlights differences in weighing information when construing options more concretely or abstractly. As an example, individuals have been shown to weigh feasibility considerations more when construing concretely, while desirability considerations are weighed more strongly when construing abstractly [18]. This reasoning could have implications for the usage of decision making aids as well, as a catalyst may trigger differential weighing of attributes (e.g., prestige versus convenience of a job). The present studies do not allow for testing differences in weighing. Future research may therefore fruitfully explore this more cognitive pathway.

Next to investigating the more cognitive side of catalyzing processes, one could also speculate to what extent decision difficulty moderates whether flipping a coin is perceived as helpful or not. Moreover, it may prove interesting to systematically gauge the impact of prior preferences.

Future research could also investigate how personality traits or states might affect the consequences of using a catalyst for decision making. As an example, studies could include a trait measure such as tendency to feel regret [19], analyzing whether individuals who might find it more difficult to make decisions due to anticipated regret might benefit from using a catalyst. Moreover, studies could also focus on the potential regret participants could experience in a certain decision scenario, investigating whether potential regret (just as with decision difficulty) might moderate the catalyzing effects of the random decision making aids.

Why does affect matter in judgment and decision-making processes?

The aim of the present research was to investigate whether using a catalyst can result in feelings of satisfaction and dissatisfaction. Going beyond the scope of the studies presented here, it is interesting to speculate about downstream consequences of these elicited feelings. Study 2 hints at this question as the results show that catalyst trials were not only associated with stronger affective reactions, but also with an immediate feeling of preferring one option over the other. Apparently, the elicited feelings may allow individuals to form a clear preference and therefore also impact subsequent decision-making processes.

The role of affect in judgment and decision-making has long been investigated and is the topic of thriving research. A plethora of findings suggests that feelings are used in decision-making, and may even constitute a critical ingredient [20–25]. Extant research suggests that reliance on feelings (next to thoughts) in judgments and decisions is not a phenomenon confined to very limited circumstances, but rather the default operation of everyday life [26,27]. Furthermore, reliance on feelings can actually increase under uncertainty [28], which is often the reason for why individuals rely on decision-making aids in the first place.

As described before, the feeling-as-information account suggests that feelings, just like pieces of content information, can be used as information to make decisions [8,9]. This account might be best illustrated by the “How do I feel about it” heuristic [10,11]. This heuristic involves holding a representation of a target in mind and inspecting the feelings that this representation may elicit [12], when being pursued. These feelings, in turn, are used to decide.

Should individuals use a catalyst to aid decision making?

We have shown that in the setting of asking individuals to make inconsequential menu choices, using a catalyst strengthens feelings and results in an immediate feeling of knowing what to choose. We carefully speculate that the same strategy may successfully be applied to more consequential decision context, too, where helping individuals to make difficult decisions may prove beneficial both on the individual and on the societal level. For instance, choosing a financial savings plan for retirement is an important decision. Individuals, however, often postpone such a decision. However, not making the decision results in a reduced amount of savings for the individuals, as interest and compound interest are forfeited, and there are higher risks for society, as the public may need to cover gaps in individual saving plans. Moreover, not making the decision may produce a very aversive and unpleasant state. If using a catalyst enables individuals to assess their feelings, get information about their preference, and eventually get started on difficult decisions, they and society may benefit on the levels of finance and well-being.

Although we think that there is good reason for these speculations to prove true, it is important to note that the specific circumstances under which the use of a catalyst can be beneficial await further empirical investigation. We assume that using a catalyst strengthens feelings, and others have suggested that relying on feelings can influence and even improve decision-making [23]. However, it is important to note that feelings can serve as a basis of accurate as well as mistaken inferences, depending on the relationship between feelings and the target [26]. Feelings may prove beneficial, for instance, when sufficient background knowledge is possessed, in that individuals who trust their feelings can predict outcomes of future events better than individuals with lower trust in feelings [29]. In sum, if decision quality benefits from including feelings, using a catalyst could be advantageous.

Interestingly, having a strong feeling about one’s own preferences could have benefits that go beyond decision accuracy. An immediate feeling could also increase certainty and perceived ease of deciding. Reversely, post-decisional regret and dissonance may be decreased. Speculating further, decision certainty or ease could even constitute one reason as to why individuals use a decision-making aid with a random outcome in the first place: although they do not know if their decision benefits from using a catalyst, considering their feelings may ease the decision process and increase certainty in regard to the outcome. Further research should investigate the nuances in affective and cognitive reactions towards the usage of catalysts in different decision-making scenarios.