2. Background and hypotheses

The seminal works of Daniel Kahneman and Amos Tversky have demonstrated that perceptions and decision-making are frequently influenced by various heuristics and biases, leading to systematic and predictable errors (e.g., Tversky & Kahneman, Reference Tversky and Kahneman1974, Reference Tversky and Kahneman1981). Actually, because individuals often have a difficulty in processing numeric data, they rely on the numerosity heuristic (Pelham et al., Reference Pelham, Sumarta and Myaskovsky1994; Shrivastava et al., Reference Shrivastava, Jain, Nayakankuppam, Gaeth and Levin2017; West et al., Reference West, Azab, Ma and Bitter2020). Accordingly, individuals are influenced by (i) larger numbers and distracted by unit manipulations (120 minutes are perceived greater than 2 hours) and (ii) phonetically longer numbers that are perceived greater than phonetically smaller ones (two thousand eight hundred seventy-five greater than three thousand). Thus, we formulate the following hypotheses:

•

• H1. Food waste expressed in a higher numeric value is perceived greater than an equivalent quantity in a lower numeric value.

• H2. Individuals are more likely to reduce food waste when it is framed in a higher numeric value, compared to its lower equivalent.

• H3: Lower food waste expressed with a phonetically longer denomination is perceived greater than a higher food waste with a phonetically smaller denomination.

• H4. Individuals are more likely to reduce food waste when it is framed in a phonetically-longer denomination, compared to a phonetically lower one.

In order to test our hypotheses, we conducted two experimental surveys. In Study 1 (S1), there are two treatments (T1 and T2) with an identical and realistic food waste quantity expressed in different units (2.5 kg in S1T1 versus 2500 g in S1T2). In Study 2 (S2), we compare a higher quantity expressed with few syllables (three kilograms in T1) and a smaller quantity expressed with more syllables (two kilos eight hundred seventy-five grams in T2). Table 1 below depicts the flow across studies.

Table 1. Experimental design

3. Study 1

3.2. Results

Mean responses (Table 2) suggest that perception of food waste volume and likelihood to reduce it were not affected by our numerosity manipulation. Analyzing the effect of the treatment on the likelihood to reduce food waste in a regression controlling for the perceived level of food waste, participants’ own practice, age, and gender (Table 3), confirms that H1 and H2 are not supported. Interestingly, individuals who perceive the food waste volume as high are more likely to reduce it, regardless of the treatment (the interaction between the two variables is not significant).

Table 2. Mean responses for food waste quantity and likelihood to reduce it

Table 3. Effect of larger numbers on food waste reduction

*** and ** stand for parameter significance at the 1% and 5% levels, respectively.

4. Study 2

4.2. Results

Unlike our prediction (H3), two kilos eight hundred seventy-five grams of food waste are perceived lower than three kilograms and intentions to reduce food waste do not vary significantly across treatments (Table 4). Examining the effect of the treatment on the likelihood to reduce food waste reveals that H4 is not supported (Table 5).

Table 4. Mean responses for FW quantity and likelihood to reduce it

Table 5. Estimation of the effect of phonetical denomination on food waste reduction

*** stands for parameter significance at the 1% level.

5. Conclusion

Despite its promising features to address food waste issues, our manipulations based on the numerosity bias failed to change food waste perceptions and raise awareness. Unlike previous studies, we considered the numerosity bias in a domain unrelated to money. In addition, our sample consisted of participants with backgrounds in agriculture and environmental matters, likely to hold strong normative beliefs, which may have reduced their susceptibility to numerosity bias related to food waste.

The ineffectiveness of our manipulation may also be attributed to two other key factors.Footnote ¹ Firstly, the numerosity bias, typically observed in evaluative situations, may not apply to food waste scenarios that focus on divesting rather than assessing value. Interestingly, the domain analyzed in our study pertains to a very concrete ‘bad’ (Grolleau et al., Reference Grolleau, Mzoughi and Weber2025), whereas previous studies often examined more abstract ‘goods’ or less negatively connoted items, such as price discounts or sharing. Secondly, food waste is likely associated with disgust – an instinctual reaction that serves as a protective mechanism to avoid potential sickness. Discarded food may evoke feelings of contamination or decay, reinforcing negative perceptions (see, e.g., Davey, Reference Davey2011; Horberg et al., Reference Horberg, Oveis, Keltner and Cohen2009; Rozin & Fallon, Reference Rozin and Fallon1987; Rozin et al., Reference Rozin, Haidt, McCauley, Lewis and Haviland-Jones2000). Disgust is often linked to the experience of nausea and prompts a need to expel possible harmful objects, potentially overriding numerosity appraisals. Moreover, when considered through the lens of affects-as-information or as-cognitive feedback models, negative emotions serve as a ‘Stop!’ signal and significantly influence decision-making (Huntsinger et al., Reference Huntsinger, Isbell and Clore2014; Schwarz & Clore, Reference Schwarz and Clore1983, Reference Schwarz, Clore, Fiedler and Forgas1988). This mechanism may explain why disgust exerts a stronger influence on food waste judgments than numerosity assessments. In sum, our findings suggest that the numerosity bias may have well-delineated scopes and that strong emotional responses can significantly shape decision-making processes, often superseding more abstract cognitive mechanisms.