Wasting Despite Motivation: Exploring the Interplay of Perceived Ability and Perceived Difficulty on Food Waste Behavior Through Brehm’s Motivational Intensity Theory

Abstract

Household food waste remains a persistent challenge despite widespread pro-environmental intentions. Drawing on Brehm’s Motivational Intensity Theory, this study examined how perceived difficulty and perceived ability interact with motivation to predict self-reported food waste. We surveyed 939 participants in Flanders and Spain, measuring motivation to avoid waste, self-rated perceived ability to manage food, meal planning perceived difficulty, and food waste. Moderated moderation analyses revealed that motivation and perceived ability each independently predicted lower waste. Crucially, a significant three-way interaction showed that motivation most effectively reduced waste when perceived difficulty was low and perceived ability was high; when perceived difficulty exceeded perceived ability, motivation had no mitigating effect. These findings underscore that effort mobilization influenced by both individual capacity and situational demands is key to closing the intention–behavior gap in food waste. Practically, interventions should go beyond raising awareness to simplify tasks and bolster consumers’ skills, aligning action demands with realistic effort levels.

1. Introduction

Food waste represents a significant global challenge with far-reaching environmental, economic, and social implications. In 2021, the European Union generated approximately 131 kg of food waste per person, with households responsible for 54% of this total (~70 kg per person); the remainder arose across other supply chain stages [1]. Despite numerous campaigns and policies, reductions in household food waste remain modest and inconsistent [2,3]. Psychological models have emphasized attitudes, norms, and awareness, yet typically explain only a limited share of actual behavior [4,5], while logistical/economic interventions (e.g., portion sizes, pricing) deliver mixed effects [2,6]. A common omission is the difficulty of food management tasks and the conditions under which people can translate pro-environmental intentions into sustained, effortful action [7,8,9]. Even consumers with strong sustainability values may view systematic meal planning or leftover reuse as complex or time-consuming, which can prompt disengagement [10].

This intention–behavior gap is well documented: Intentions explain only a fraction of variance in pro-environmental actions [11,12,13,14,15]. We propose that a mismatch between perceived difficulty and perceived ability undermines effort mobilization in household food waste reduction. Accordingly, we apply Motivational Intensity Theory (MIT) [16], which predicts that effort rises with perceived task difficulty up to the point at which demands exceed perceived ability, after which effort collapses. Using large-scale survey data, we test whether perceived ability buffers the demotivating effect of high perceived difficulty and whether, under manageable difficulty, motivation translates more reliably into lower food waste. Beyond awareness-raising, the findings inform interventions that lower perceived difficulty (e.g., simplifying planning) and raise perceived ability (e.g., tools, reminders, skill-building).

1.1. Literature Review

Household food waste is a multifaceted, goal-directed behavior shaped by cognitive, affective, and contextual factors. Intention-based models such as the Theory of Planned Behavior (TPB) link attitudes, subjective norms, and perceived behavioral control to waste-related intentions, but they typically explain only a modest share of actual disposal practices [4,5]. Habits and routines can further erode intention–action consistency as shopping and discard behaviors become automatized over repetition [4,17,18]. Emotions—guilt, disgust, anticipated regret—also sway discard versus reuse decisions, while self-regulatory supports (e.g., implementation intentions, storage and portioning skills) can help yet often show context-dependent effects [5,10,11,19]. Taken together, these strands suggest that reducing food waste requires sustained effort under shifting constraints, not merely concern.

The Motivation–Opportunity–Ability (MOA) framework argues that behavior occurs when people both want to act and can act within situational opportunities (time, logistics) and with sufficient abilities (skills, resources) [20]. Recent household studies combine MOA with TPB (and related models) and show that opportunity and ability condition whether intentions are enacted—i.e., intentions translate into behavior when people have time/logistical opportunity and adequate capability [21,22]. At the same time, MOA is largely static: It enumerates prerequisites for action without formalizing the interplay between a person’s perceived capacity and the perceived demands of a specific task at a given moment.

This gap can be addressed by bringing Motivational Intensity Theory (MIT) into the food studies context. MIT specifies how much effort people will mobilize for a goal, given (a) the importance of the goal (potential motivation), (b) the perceived task difficulty (required effort), and (c) the expectancy of success (whether success seems attainable). The effort rule is simple: Individuals invest effort up to the level required by perceived difficulty, as long as success is feasible; if required effort exceeds what seems feasible (relative to one’s capacity), effort collapses [16,23,24,25]. This yields a characteristic pattern: Effort increases with difficulty while the task is seen as doable but drops sharply once difficulty is appraised as greater than one’s perceived ability.

Whereas MOA lists what is needed (motivation, opportunity, ability), MIT formalizes the moment-to-moment trade-off between subjective demands (perceived difficulty) and subjective capacity (perceived ability). In other words, MOA is descriptive and largely static, while MIT is mechanistic and dynamic: It predicts when people scale effort up and when they disengage, given their appraisal of difficulty relative to ability.

Laboratory research across cognitive and self-regulatory tasks shows that cardiovascular and behavioral indicators of effort scale with task difficulty when success is plausible but decline when difficulty is (or is believed to be) too high [16,23,25]. These effects arise with both objective difficulty (task parameters) and perceived difficulty (instructions, uncertainty) and are moderated by perceived ability/expectancy, such that higher perceived ability sustains effort at higher levels of difficulty.

In everyday food management, planning uncertainty (e.g., fluctuating headcounts), time pressure, storage/portioning complexity, and safety judgments can make waste reduction tasks feel difficult; if people also doubt their capacity—skills, routines, or tool use—they may disengage despite strong values [7,8,9,10,11]. Accordingly, we treat perceived difficulty (subjective task demand) and perceived ability (subjective capacity) as joint regulators of effort in household food waste.

Prior evidence suggests that stronger motivation and higher perceived ability relate to lower waste, whereas greater perceived difficulty relates to higher waste [4,5,10,11,19,24]. MIT yields a specific interactive prediction: Motivation most strongly reduces waste when perceived ability is high and perceived difficulty is manageable, but its effect weakens or collapses when perceived difficulty exceeds perceived ability [16,23,24,25].

1.2. Overview of the Study

This study advances the psychological account of household food waste by applying Motivational Intensity Theory (MIT). We treat food waste reduction as goal-directed action regulated by effort mobilization, which depends not only on motivation but also on whether people perceive they can meet the task demands—that is, on perceived ability relative to perceived difficulty. Using a large survey (CHORIZO), we analyze motivation to avoid waste, perceived ability to manage food, perceived difficulty in meal planning, and covariates. Objectives: (i) test an MIT-based model in which motivation predicts lower waste conditional on perceived ability and perceived difficulty; (ii) quantify the main associations of motivation, perceived ability, and perceived difficulty with two outcomes (Household Food Waste Frequency; Perceived Relative Food Waste); and (iii) assess robustness with standard covariates and derive actionable levers for interventions (reduce perceived difficulty; build perceived ability). Key hypotheses: (1) Higher motivation relates to lower food waste, (2) higher perceived ability relates to lower food waste, (3) higher perceived difficulty relates to higher food waste, and (4) the core MIT test, a three-way interaction (motivation × perceived ability × perceived difficulty): Motivation most strongly reduces waste when perceived ability is high and perceived difficulty is manageable, but its effect weakens when perceived difficulty exceeds perceived ability. The tested model is presented on Figure 1.

2. Materials and Methods

In total, 1005 participants (491 females, $M_{age}$ = 48.43, $SD$ = 16.56) completed the online survey. The Flemish sample included 800 participants (351 females, $M_{age}$ = 49.69, $SD$ = 17.43) and the Spanish one included 205 participants (140 females, $M_{age}$ = 43.55, $SD$ = 11.47). The survey was conducted in Dutch for the Belgian participants and Spanish for the Spanish participants. No exclusion criteria were applied to the samples beyond the recruitment quotas. The same data was used in other publications with different variables, theoretical backgrounds, and research questions. In Flanders, Belgium, iVox (market research company) was contracted to recruit respondents from their database. In Spain, the survey was deployed via the CNTA Mundor Sabor consumers database.

We used only a subset of the variables collected in the full study for the analyses presented here. The final sample comprised 939 participants, after excluding those who did not complete all the questions of interest.

The Committee for Ethics in Social Science and Humanities of the Flanders Research Institute of Agricultural, Fisheries and Food (ILVO) has confirmed that the research was conducted according to the ethical rules presented in its Ethical Principles and Procedures document. Ethical clearance number: ComESSH_2026_06(1).

2.1. Measures

2.1.1. Motivation

To measure the motivation to not waste food, we used a 3-item scale ($M$ = 5.74; $SD$ = 1.12). Participants rated their agreement with three statements, each reflecting a decreasing level of motivation. The items were “In my daily life, I try very actively to avoid food waste,” “I think throwing away food is very irresponsible,” and “I feel bad (e.g., guilty) when I throw away food.” Participants responded to the questions on a 7-point Likert scale from 1 (totally disagree) to 7 (totally agree). The answers were averaged, and the reliability of the final measure was high ($\alpha$ = 0.76).

2.1.2. Perceived Ability

Perceived ability is defined here as not only internal skills and knowledge (e.g., label literacy, portioning routines), but also access to—and confidence in using—external resources and tools (e.g., measuring cups, storage solutions, planning aids), because all of these shape one’s sense of “can do” in everyday food management [10,11,19,24]. To assess participants’ perceived ability, we asked them to answer seven questions, like, “We often use tools (e.g., scale, measuring cup) to prepare just the right amount/portion size per person,” “Before going to the store, we always check the food stock at home (e.g., in the refrigerator, pantry),” and “We always think carefully about how much exactly we need to prepare so that everything gets eaten.” Participants responded the questions on a 7-point Likert scale from 1 (totally disagree) to 7 (totally agree), and mean of their answers was saved as the perceived ability index ($M$ = 5.31; $SD$ = 1.01). The reliability of the measure was high ($\alpha$ = 0.77). The full list of questions can be found in the Supplementary Materials.

2.1.3. Food Waste

Household Food Waste Frequency Index. To assess the frequency of household food waste behaviors across a range of categories, we used a composite index based on seven specific food types. Participants were asked, “How often does your household throw away food? Please rate by category.” For each of the seven categories, respondents indicated the frequency of disposal using an 8-point scale from 1 (never) to 8 (every day). Some of the categories included in the index were “Completely unused foods with long shelf life (e.g., sealed bag of crisps, unopened cookie package),” “Meal leftovers left on plates, pots, or pans immediately after meals,” and “Partly used foods with long shelf life (e.g., half-opened cookie package).” Responses across all seven categories were averaged to form a Household Food Waste Frequency Index ($M$ = 2.61; $SD$ = 1.18), with higher scores indicating more frequent food disposal. This index provides a general measure of self-reported household food waste behavior by incorporating both perishable and non-perishable items and accounting for both completely unused and partially consumed food. The use of a mean score assumes equal weighting of each item and enables straightforward comparison across individuals and subgroups. The reliability of the measure was high ($\alpha$ = 0.88). The full list of questions can be found in the Supplementary Materials.

Perceived Relative Food Waste. To assess participants’ self-perception of their household food waste, we used a single-item measure with an informational anchor. We established this informational anchor at 1.7 kg based on a report by the Departement Omgeving van de Vlaamse overheid and GfK Belgium Social and Strategic Research [26]. Respondents were asked the following question: “Compared to an average household in [Flanders/Spain], how much solid food and beverage waste do you think your household generates in a typical week? On average, households in [Flanders/Spain] waste approximately 1.7 kg of solid food and beverages per week.” Participants responded using a 7-point Likert scale ranging from 1 (much less than other households) to 7 (much more than other households). This measure ($M$ = 2.62; $SD$ = 1.37) captures participants’ subjective evaluation of their food waste behaviors relative to the perceived norm. The 1.7 kg weekly reference served as a common informational anchor to reduce between-respondent variability in scale interpretation. We note that anchors can shape judgments [27,28,29]; thus, this value should be considered a heuristic benchmark, not a factual country-specific average for every respondent.

2.1.4. Meal Planning Perceived Difficulty

Guided by our motivational framework—where difficulty denotes a subjective task demand (Motivational Intensity Theory) and, in the TPB tradition, perceived control indexes the ease or difficulty of performing a behavior—we treat contextual unpredictability as a source of perceived difficulty in meal planning [23,24,25,30,31,32]. Accordingly, we used a single-item indicator targeting the participation variability facet of perceived difficulty: “How many people will join for the meal is always subject to last-minute changes.” Respondents answered on a 1–7 Likert scale (1 = totally disagree; 7 = totally agree). Higher scores indicate greater perceived unpredictability and, by extension, higher perceived planning difficulty. Descriptives were M = 2.79, SD = 1.74. This item captures logistical uncertainty common in shared meals—especially in dynamic or non-routine households—where unanticipated changes in diners can prompt over-preparation and subsequent waste. We acknowledge that this is a single-facet, single-item operationalization; its psychometric limitations are noted, and we outline a roadmap for a multi-item perceived-difficulty scale in the Section 4 [2,33,34,35,36].

2.2. Control Variables

In all analyses, we included a set of covariates that prior research has linked to household food waste behaviors. First, we controlled for location (Flanders vs. Spain). Because the sample sizes for Flanders (n = 800) and Spain (n = 205) are imbalanced, country is used solely as a covariate to adjust mean differences. We do not interpret country-specific effects, and our analyses are not designed to support cross-national comparisons. Any country-level patterns should therefore be viewed as exploratory and potentially unstable given unequal cell sizes. We also adjusted for demographics: respondent gender, age (in years), and educational level (higher education vs. up to secondary). To account for economic circumstances, we included household income satisfaction, based on the six-point item “Which of the descriptions comes closest to how you feel now about your household income?” (from 1 = living comfortably to 4 = having a very difficult time). We further controlled for respondents’ typical food provisioning patterns, measured by five items asking how often they (1) cook for themselves, (2) order takeout or delivery, (3) use food rescue apps/websites (e.g., Too Good To Go), (4) go out to eat in restaurants, and (5) subscribe to a meal kit or food box service (e.g., HelloFresh) (all on 1 = never to 5 = daily scales). Finally, we accounted for hometown urbanicity, with a single item asking, “How would you describe your hometown?” rated from 1 = very rural to 7 = very urban. Including these controls helps isolate the unique effects of perceived difficulty and perceived ability on food waste outcomes.

3. Results

3.1. Household Food Waste Frequency Index

We tested a pre-specified three-way moderated moderation using Hayes’s PROCESS Model 3 [37] with 20,000 bootstrap samples, treating motivation as the focal predictor (X), perceived ability as moderator W, and perceived difficulty as moderator Z. All continuous variables were mean-centered (low/mean/high = −1 SD/0/+1 SD). The model included the covariates listed in Section 2.2. The overall model accounted for 29.03% of the variance, R² = 0.2903, F(18, 920) = 20.9045, p < 0.001.

Main effects. Motivation negatively predicted food waste, b = −0.2644, SE = 0.0341, t(920) = −7.7483, p < 0.001, 95% CI [−0.3313, −0.1974]. Perceived ability also negatively predicted food waste, b = −0.2347, SE = 0.0359, t(920) = −6.5382, p < 0.001, 95% CI [−0.3051, −0.1642]. Perceived difficulty was not a significant main effect, b = 0.0346, SE = 0.0201, t(920) = 1.7198, p = 0.0858, 95% CI [−0.0049, 0.0740].

Two-way interactions. The motivation × perceived difficulty interaction was significant, b = 0.0428, SE = 0.0189, t(920) = 2.2665, p = 0.0237, 95% CI [0.0057, 0.0799]. The motivation × perceived ability interaction was not significant, b = −0.0248, SE = 0.0269, t(920) = −0.9210, p = 0.3573, 95% CI [−0.0777, 0.0281]. The perceived ability × perceived difficulty interaction showed a marginal trend, b = −0.0396, SE = 0.0206, t(920) = −1.9244, p = 0.0546, 95% CI [−0.0799, 0.0008].

Three-way interaction. Critically, the motivation × perceived ability × perceived difficulty term was significant, b = −0.0489, SE = 0.0158, t(920) = −3.0893, p = 0.0021, 95% CI [−0.0800, −0.0178], adding variance beyond lower-order terms, ΔR² = 0.0074, F(1, 920) = 9.5436, p = 0.0021.

Probing the interaction. Consistent with a priori expectations, the conditional motivation × perceived ability interaction depended on perceived difficulty: at low perceived difficulty (−1.74 SD), effect = 0.0604, F(1, 920) = 2.5559, p = 0.1102; at the mean, effect = −0.0248, F(1, 920) = 0.8483, p = 0.3573; at high perceived difficulty (+1.74 SD), effect = −0.1100, F(1, 920) = 7.8262, p = 0.0053. Simple-slope estimates for the focal predictor (motivation) showed that when perceived ability was low and perceived difficulty was high, the slope was small and non-significant, b = −0.0796, SE = 0.0543, t(920) = −1.4650, p = 0.1433, 95% CI [−0.1861, 0.0270]. In contrast, when perceived ability was high and perceived difficulty was high, motivation significantly predicted lower food waste, b = −0.3001, SE = 0.0683, t(920) = −4.3971, p < 0.001, 95% CI [−0.4340, −0.1662]. For completeness, at mean perceived difficulty the slope of motivation was negative at both low perceived ability, b = −0.2395, SE = 0.0393, t(920) = −6.1009, p < 0.001, 95% CI [−0.3165, −0.1624], and high perceived ability, b = −0.2892, SE = 0.0474, t(920) = −6.1031, p < 0.001, 95% CI [−0.3822, −0.1962]; at low perceived difficulty, the slope remained negative at both low perceived ability, b = −0.3994, SE = 0.0577, t(920) = −6.9237, p < 0.001, 95% CI [−0.5126, −0.2862], and high perceived ability, b = −0.2784, SE = 0.0634, t(920) = −4.3901, p < 0.001, 95% CI [−0.4028, −0.1539].

Interpretation. These findings align with the opening claim that concern and intentions alone explain only a limited share of actual behavior and that interventions focused purely on information or logistics often yield modest and inconsistent effects. Framed through Motivational Intensity Theory (MIT), the results indicate that household food waste reduction is not simply a function of “wanting” to act; rather, it depends on whether people believe they can meet the perceived demands of the task. When food management activities are perceived as difficult, motivation translates into action only if perceived ability is sufficiently high. In contrast, when the task feels easier, motivation tends to predict lower waste regardless of perceived ability. This asymmetry is precisely what MIT anticipates: People scale their effort up with difficulty only up to the point where demands remain within their perceived capacity; once difficulty is judged to exceed perceived ability, effort collapses even in the presence of strong motivation [16,20,22,28]. The results are presented in Figure 2. The regression analysis outcomes for covariates predicting the Household Food Waste Frequency Index are depicted in

Table 1. Regression coefficients, standard errors, t-values, p-values, and 95% confidence intervals for covariates predicting the Household Food Waste Frequency Index.

Covariate	b	SE	t	p	LLCI	ULCI
Location (Spain = 1; Flanders = 0)	0.0990	0.0942	1.0510	0.2935	−0.0858	0.2838
Gender (female = 1; male = 2)	0.1588	0.0671	2.3667	0.0182	0.0271	0.2906
Age (years)	−0.0088	0.0023	−3.8500	0.0001	−0.0133	−0.0043
Education (1 = up to secondary; 2 = higher)	0.0854	0.0685	1.2465	0.2129	−0.0490	0.2198
Income satisfaction (1 = living comfortably to 4 = very difficult)	0.0110	0.0316	0.3472	0.7285	−0.0510	0.0729
Cook at home (1 = never to 5 = daily)	0.0420	0.0512	0.8198	0.4125	−0.0585	0.1425
Takeout/delivery (1 = never to 5 = daily)	0.2156	0.0467	4.6174	<0.0001	0.1240	0.3073
Food rescue apps (1 = never to 5 = daily)	−0.0350	0.0432	−0.8092	0.4186	−0.1197	0.0498
Eat out (1 = never to 5 = daily)	0.2025	0.0543	3.7304	0.0002	0.0960	0.3090
Meal kit service (1 = never to 5 = daily)	0.0321	0.0560	0.5738	0.5662	−0.0778	0.1421
Urbanicity (1 = very rural to 7 = very urban)	0.0080	0.0183	0.4365	0.6626	−0.0279	0.0439

.

3.2. Perceived Relative Food Waste

We tested a pre-specified three-way moderated moderation using Hayes’s PROCESS Model 3 [37] with 20,000 bootstrap samples, treating motivation as the focal predictor (X), perceived ability as moderator W, and perceived difficulty as moderator Z. All continuous variables were mean-centered (low/mean/high = −1 SD/0/+1 SD). The model included the covariates listed in Section 2.2. The overall model accounted for 27.49% of the variance, R² = 0.2749, F(18, 920) = 19.3802, p < 0.001.

Main effects. Motivation negatively predicted FW_17, b = −0.3535, SE = 0.0402, t(920) = −8.7987, p < 0.001, 95% CI [−0.4324, −0.2747]. Perceived ability also negatively predicted FW_17, b = −0.2476, SE = 0.0423, t(920) = −5.8565, p < 0.001, 95% CI [−0.3305, −0.1646]. Perceived difficulty was a positive predictor, b = 0.0909, SE = 0.0237, t(920) = 3.8386, p = 0.0001, 95% CI [0.0444, 0.1373].

Two-way interactions. The motivation × perceived ability interaction was significant, b = −0.0740, SE = 0.0317, t(920) = −2.3309, p = 0.0200, 95% CI [−0.1362, −0.0117]. The motivation × perceived difficulty interaction was not significant, b = −0.0026, SE = 0.0222, t(920) = −0.1191, p = 0.9052, 95% CI [−0.0463, 0.0410]. The perceived ability × perceived difficulty interaction was not significant, b = −0.0387, SE = 0.0242, t(920) = −1.5987, p = 0.1102, 95% CI [−0.0863, 0.0088].

Three-way interaction. Critically, the motivation × perceived ability × perceived difficulty term was significant, b = −0.0687, SE = 0.0187, t(920) = −3.6817, p = 0.0002, 95% CI [−0.1053, −0.0321], adding variance beyond lower-order terms, ΔR² = 0.0107, F(1, 920) = 13.5552, p = 0.0002.

Probing the interaction. Consistent with a priori expectations, the conditional motivation × perceived ability interaction depended on perceived difficulty: at low perceived difficulty (−1.74 SD), effect = 0.0456, F(1, 920) = 1.0519, p = 0.3053; at the mean, effect = −0.0740, F(1, 920) = 5.4329, p = 0.0200; at high perceived difficulty (+1.74 SD), effect = −0.1935, F(1, 920) = 17.4638, p < 0.001. Simple-slope estimates for the focal predictor (motivation) showed that when perceived ability was low and perceived difficulty was high, the slope was negative and significant, b = −0.1641, SE = 0.0640, t(920) = −2.5668, p = 0.0104, 95% CI [−0.2897, −0.0386]. When perceived ability was high and perceived difficulty was high, motivation strongly predicted lower relative food waste, b = −0.5521, SE = 0.0804, t(920) = −6.8694, p < 0.001, 95% CI [−0.7098, −0.3944]. For completeness, at mean difficulty the slope of motivation was negative both at low perceived ability, b = −0.2794, SE = 0.0462, t(920) = −6.0439, p < 0.001, 95% CI [−0.3701, −0.1887], and high perceived ability, b = −0.4276, SE = 0.0558, t(920) = −7.6628, p < 0.001, 95% CI [−0.5372, −0.3181]; at low perceived difficulty, the slope remained negative at both low perceived ability, b = −0.3946, SE = 0.0679, t(920) = −5.8089, p < 0.001, 95% CI [−0.5280, −0.2613], and high perceived ability, b = −0.3032, SE = 0.0747, t(920) = −4.0603, p = 0.0001, 95% CI [−0.4497, −0.1566].

Interpretation. In line with Motivational Intensity Theory and the intention–behavior gap outlined in the introduction, motivation consistently predicts lower perceived relative food waste, but its protective effect is larger when perceived ability is higher, and this perceived ability-based strengthening intensifies as perceived difficulty rises (i.e., the motivation × perceived ability interaction becomes increasingly negative with increasing perceived difficulty). This pattern underscores the need for interventions that both raise and lower perceived difficulty (to convert pro-environmental motivation into sustained household practices). The results are illustrated in Figure 3. The regression analysis outcomes for covariates predicting perceived relative food waste are depicted in

Table 2. Regression coefficients, standard errors, t-values, p-values, and 95% confidence intervals for covariates predicting perceived relative food waste.

Covariate	b	SE	t	p	LLCI	ULCI
Location (Spain = 1; Flanders = 0)	0.0122	0.1109	0.1100	0.9124	−0.2055	0.2299
Gender (female = 1; male = 2)	0.0173	0.0790	0.2186	0.8270	−0.1378	0.1724
Age (years)	−0.0148	0.0027	−5.4791	<0.0001	−0.0201	−0.0095
Education (1 = up to secondary; 2 = higher)	−0.1148	0.0806	−1.4237	0.1549	−0.2731	0.0435
Income satisfaction (1 = living comfortably to 4 = very difficult)	0.0491	0.0372	1.3201	0.1871	−0.0239	0.1220
Cook at home (1 = never to 5 = daily)	0.1036	0.0603	1.7181	0.0861	−0.0147	0.2220
Takeout/delivery (1 = never to 5 = daily)	0.1276	0.0550	2.3201	0.0206	0.0197	0.2355
Food rescue apps (1 = never to 5 = daily)	−0.0133	0.0509	−0.2612	0.7940	−0.1131	0.0865
Eat out (1 = never to 5 = daily)	0.1270	0.0639	1.9865	0.0473	0.0015	0.2524
Meal kit service (1 = never to 5 = daily)	0.0497	0.0660	0.7539	0.4511	−0.0797	0.1792
Urbanicity (1 = very rural to 7 = very urban)	−0.0398	0.0215	−1.8470	0.0651	−0.0820	0.0025

.

These findings suggest that both motivation and perceived ability are key predictors of food waste. However, their impact is significantly influenced by the level of the moderator variable, i.e., perceived difficulty. Specifically, motivation is most effective at reducing food waste when individuals have a higher perceived ability, and this effect is stronger when values of perceived difficulty are higher. It means that when perceived ability is low and the action to perform is perceived as difficult, motivation is not enough to mitigate food waste.

4. Discussion

The present study explored whether the relationship between motivation to reduce food waste and actual self-reported food waste is influenced by both individuals’ perceived ability to act on that motivation and the difficulty of food-saving behaviors. The results indicated a moderated moderation effect, suggesting that the influence of motivation on food waste was contingent not only on individuals’ perceived ability but also on the perceived difficulty of the situation. As expected, higher motivation and higher perceived ability were independently associated with lower food waste. However, the three-way interaction revealed that the protective effect of motivation was significantly stronger when perceived ability was high and perceived difficulty was low. In contrast, when difficulty was high and perceived ability was not, the motivation was insufficient to mitigate food waste.

These findings emphasize the importance of considering contextual and individual resource-related factors when designing interventions aimed at reducing food waste. The results suggest that motivational interventions alone may not be sufficient, especially for individuals who face higher perceived barriers or possess fewer relevant skills or resources. Furthermore, these findings may explain intra-individual variance, explaining why the same person may sometimes act on environmental intentions and sometimes not (depending on the difficulty of the task, while motivation and perceived ability remain constant). In line with theoretical models of effort mobilization (e.g., [16,23,24,25]), motivation may translate into behavior only when individuals perceive that the effort required is manageable and that success is within reach. Interestingly, the interaction suggested that people may disengage from acting on their goals when the perceived effort outweighs their confidence in success. In such cases, even strong motivation may not be sufficient to produce behavioral change, especially if the task is appraised as too demanding.

Practical and policy implications. From a practical standpoint, interventions aimed at reducing food waste should go beyond motivational appeals and consider the “match” between motivation, perceived ability, and contextual difficulty. This approach is more tailored, focusing not only on building stronger intentions but also on making reducing waste accessible within the bounds of realistic human effort investment. These findings provide a new perspective on why this gap occurs and suggest a “more fundamental” motivation mechanism at work: assessing and appraising the required effort in relation to perceived ability and difficulty, particularly in the context of meal planning. For instance, simplifying tasks, providing scaffolding or skill-building, and reducing perceived complexity can strengthen the link between intention and behavior.

Our MIT-based pattern suggests two complementary levers: lower perceived difficulty and raise perceived ability. Policy and program designers can (i) support headcount-aware planning tools (e.g., household-size toggles in public apps, default flexible recipes) to reduce planning uncertainty, (ii) standardize portioning aids (free or low-cost measuring cups/scoops distributed through municipalities, food banks, schools), (iii) simplify date-label communication (clear “use by” vs. “best before” campaigns with iconography and fridge/freezer guidance), (iv) promote FIFO-friendly storage (simple bin/label kits; shelf talkers at retailers to cue planned quantities), (v) target out-of-home eating routines (nudges in delivery apps for smaller portions, leftover plans, or “pause order” prompts tied to home stock checks), and (vi) fund skill-building micro-modules (90 s videos/checklists embedded in retailer and municipal platforms) to increase confidence in meal planning, portioning, and safe reuse. These actions are low-cost and scalable, and map directly onto the effort mechanism identified here.

Additional insights. When controlling for the important variables, we obtained additional insights; for instance, age emerged as a significant predictor. Beyond the focal MIT pattern, ancillary covariate patterns (age ↓, takeout/dining out ↑, a small gender difference) offer targeting cues but remain secondary to the core effort mechanism. In both models, older respondents reported lower perceived relative food waste and self-reported lower actual waste. In other words, the propensity to waste food declines with age. Food-related behaviors outside the home showed the strongest effects. More frequent use of takeout/delivery services was associated with increases in both perceived waste and self-reported actual waste. Likewise, more frequent dining out elevated perceived waste as well as self-reported actual waste production. Gender had a differential impact depending on the model: In the analysis of perceived relative waste, there was no significant difference between women and men, whereas in the model of self-reported actual waste, women reported higher levels of waste than men. These results should be interpreted with caution, as neither the Spanish nor the Flemish sample was representative of the general population, and thus the tests and inferences may have limited generalizability.

Study limitations. One limitation of our study is that we relied on self-reported food waste (FW) frequency. However, since our goal was to explore how intention, motivation, perceived difficulty, and perceived ability influence FW behaviors, not to compare absolute waste levels across individuals, this poses little concern. Any social desirability, recall, or intentionality bias in these self-reports is likely to affect all respondents similarly, resulting in a common measurement error that does not compromise our analysis of predictor–outcome relationships. Another limitation is that all measures (motivation, perceived ability, perceived difficulty, and self-reported waste) were collected in a single, cross-sectional survey, so we cannot definitively establish causal direction. While MIT suggests that perceived difficulty and perceived ability drive effort (and thus waste), it is equally plausible that prior waste experiences shape how participants perceive task demands or their own capability. Additionally, using a single instrument at one time increases the risk of common method bias inflating observed effects.

Our findings should be interpreted in light of a measurement limitation. We intentionally focused on a single facet of perceived difficulty—logistical planning difficulty arising from last-minute variability in household participation—because, in our framework, contextual unpredictability acts as a source of subjective task demand (Motivational Intensity Theory) and aligns with the TPB convention that perceived control indexes the ease or difficulty of performing a behavior [30,31,32]. This single-item operationalization is parsimonious but psychometrically constrained and does not capture other difficulty sources highlighted in the food waste literature—e.g., time scarcity, storage/portioning complexity, food-safety judgments, and preference coordination [2,7,34,36]. First, future work should conduct domain mapping along the consumer “food waste journey” (planning → shopping → storage → preparation → consumption → leftovers) to identify core facets of perceived difficulty (e.g., planning uncertainty, time scarcity/schedule volatility, storage and portioning complexity, food safety judgments/date labels, household preference coordination, and leftover management) [6,38,39]. Subsequent studies should develop a concise, multi-item scale spanning these facets and related “types” of difficulty (e.g., spoilage uncertainty, applying FIFO in complex home inventories, adapting meals to what is on hand, and time constraints around cooking, storage, or reheating). Such a scale would better reflect the subjective nature of difficulty and enable formal tests of Motivational Intensity Theory’s predictions (including potential curvilinear effects) across a broader range of situational constraints [39]. By doing so, researchers can gain a more nuanced understanding of how diverse task demands interact with motivation and perceived ability in shaping food waste behaviors.

The unequal cell sizes (Flanders vs. Spain) limit the stability of country-specific estimates; we therefore treat country as a covariate and avoid cross-national inference. The 1.7 kg anchor standardizes respondents’ frame of reference but may also induce anchoring/assimilation; importantly, our focal interactive pattern also emerges on the non-anchored waste frequency index, reducing the risk that conclusions depend on the anchor [27,28,29].

Further directions. Future research should use longitudinal or experimental designs to test directionality and mechanisms, and—where feasible—pair self-reports with objective indicators to curb shared-method variance. Longitudinal work could combine experience sampling of perceived difficulty with weekly kitchen diaries and smart-bin weights to model within-person change (e.g., cross-lagged or RI-CLPM tests of whether spikes in difficulty precede waste). Experiments can manipulate specific facets of perceived difficulty (e.g., headcount unpredictability, time pressure, storage complexity) via planning prompts, schedule volatility scenarios, or portioning tools, then measure downstream waste using barcode/app logs, bin weights, or photo-verified leftovers. Micro-randomized trials (just-in-time nudges before shopping/cooking) and A/B tests in meal-planning apps can probe short-horizon causal effects. Where randomization is impractical, natural experiments (e.g., sudden schedule shocks), instrumental variable strategies (e.g., exogenous school/work timetable changes), or difference-in-differences around policy or retailer interventions can support causal inference. Across designs, preregistered analyses, multi-source measurement (self-report + behavioral + sensor data), and tests of mediation/moderation (e.g., motivation/perceived ability as mediators; household volatility as a moderator) will sharpen causal claims and directly evaluate MIT’s predictions.

Our findings suggest that people differ not only in how they perceive difficulty and ability, but also in the maximum effort they are willing to invest before disengaging. Future work could measure each person’s “effort threshold,” the point at which perceived demands exceed their willingness to try. Additionally, researchers could examine how this threshold varies with personality traits (e.g., conscientiousness, need for cognition) or past experience with food waste reduction. Understanding individual differences in effort mobilization would allow for more personalized strategies. For instance, digital tools could dynamically adjust the complexity of suggested actions (e.g., from simple portion-control tips to more involved meal-planning routines) based on a user’s real-time engagement. This approach would keep tasks just below each person’s disengagement point, maximizing sustained behavior change.

5. Conclusions

Reducing household food waste is not just a matter of “caring,” but of mobilizing effort when tasks feel doable. Grounded in Motivational Intensity Theory, our results show that motivation lowers waste most reliably when perceived ability is high and perceived difficulty is manageable; when perceived difficulty exceeds ability, motivation alone does not translate into action. This interactive pattern was replicated across two outcomes and held after adjusting for standard covariates, indicating a robust, effort-based mechanism behind the intention–behavior gap.

Theoretically, we foreground a dynamic appraisal account in which perceived difficulty (subjective task demand) and perceived ability (subjective capacity encompassing skills and confidence with tools) jointly regulate effort. Practically, interventions should pair task simplification (e.g., headcount-aware planning, clearer date-label cues, easier storage/portioning) with ability-building (skills, checklists, prompts, user-friendly tools)—an approach likely to outperform awareness campaigns alone. Looking ahead, developing a multi-item perceived difficulty scale and testing causal designs with behavioral/sensor data will help tailor supports to keep actions just below each person’s disengagement threshold.