Assuming the Best: Individual Differences in Compensatory “Green” Beliefs Predict Susceptibility to the Negative Footprint Illusion

Abstract

Recent years have seen a marked increase in carbon emissions despite pledges made by the international community at the Paris Accord in 2015 to reduce fossil fuel production and consumption. Rebound effects could contribute to this phenomenon as, in which attempts to curb carbon emissions might have inadvertently led to an upswing in fossil fuel usage. The present study hypothesizes that rebound effects are driven by a misapplication of compensatory balancing heuristics, with the unintended outcome of producing inaccurate estimates of the environmental impact of “green” or environmentally friendly labelled products or behaviors. The present study therefore aims to investigate the relationship between participants’ degree of compensatory thinking (e.g., “Recycling compensates for driving a car”) and their susceptibility to the Negative Footprint Illusion, a widely replicated phenomenon demonstrating that the presence of “green” products biases carbon footprint estimations. One hundred and twelve participants were asked to complete a 15-item Compensatory Green Beliefs scale and to estimate the total carbon footprint of a set of 15 conventional houses, followed by a set that included 15 “green” houses in addition to 15 conventional houses. Results indicated that participants, on average, believed that the “green” houses were carbon neutral, and that susceptibility to the Negative Footprint Illusion was predicted by performance on the Compensatory Green Beliefs scale. This is the first study confirming that individual differences in cognitive processes (i.e., Compensatory Green Beliefs) are indeed related to inaccurate estimates of “green” products, providing a foundation for further investigation of the influence of “green” and compensatory beliefs on carbon footprint estimates.

1. Introduction

Recent years have seen urgent calls from the international scientific community for anthropogenic carbon emissions to be drastically reduced so as to avert or mitigate a number of environmental threats (e.g., global temperature increases, extinctions, warming and acidifying oceans, sea level rise, extreme weather events) [1]. The 2015 Paris Accord seemed to hail a significant international political response accompanied by strong social will to enact changes, represented by the rise of climate movements (e.g., 350.org) [2] and strikes (e.g., Fridays for Future) [3]. However, empirical measurements taken in subsequent years indicate that changes have been ineffectual, leading to some of the highest climate-change indicator averages in recorded history [4,5,6]. A 2019 UN report (“The Production Gap”) [7] evaluating countries’ planned production of fossil fuels in light of the Paris Agreement Goals concluded an extreme mismatch between intention and action at a global scale. The report indicated that by 2030, the production of fossil fuels will be 50% higher than what would be consistent with the 2 °C pathway and 120% higher than the 1.5 °C pathway. Why the necessary alterations in human behavior have not yet been made to avert this growing existential risk, despite an abundance of information, is a question on which the survival of many species may rest. Although many structural barriers to sustainable behavior exist such as economic reliance on comparatively cheap fossil fuels, psychological and cognitive barriers have also been identified as partly responsible for inaction on anthropogenic climate change, and are increasingly being recognized as an important focus of investigation [8,9,10,11].

So-called “rebound effects” occur when attempts to get consumers to reduce energy consumption, such as encouraging the use of LED lightbulbs, inadvertently leads to an increase in the behavior they were attempting to reduce, such as leaving the lights on longer and thereby consuming even more energy than they originally would [12]. A cognitive explanation of why these attempts at sustainability often have the opposite effect on people’s judgements, beliefs and behaviors has yet to be clearly articulated. Kaklamanou, Jones, Webb and Walker [10] have suggested that Compensatory Green Beliefs (CGBs) could be cognitive precursors to rebound effects; namely, beliefs that sustainable behaviors can be used to compensate for unsustainable ones. Sörqvist and Langeborg [11] have posited that compensatory beliefs probably evolved in the social domain as a form of moral cleansing, in which good deeds can be used to cancel out the effect of prior bad deeds. These beliefs are misapplied when used in the context of calculating sustainability, with potentially far-reaching implications. For example, people seeking alleviation for their harmful environmental behavior might seek to compensate for it and restore balance with comparatively ineffectual pro-environmental behavior (e.g., signing a climate-change awareness petition). Researchers found that CGBs were related to other measures of environmental beliefs; they also assessed green identity, ecological behavior and worldview, and found that those with a higher endorsement of CGBs had weaker pro-ecological viewpoints, behaviors and identities [10]. CGBs were also particularly over-represented in those most skeptical of anthropogenic climate change.

However, CGBs are not only evident in climate-change deniers and people with an anti-environmental agenda. A more general consideration indicates that CGBs are actually widespread in those actively claiming to mitigate the negative effects of climate change. Most notably, “carbon offsetting” has been adopted by companies wishing to make a difference by indirectly reducing emissions, a process whereby greenhouse gas emissions are “paid for” by implementing carbon-reducing projects elsewhere [13]. “Carbon trading” of emissions in the form of “carbon credits”, tradeable certificates that can be purchased to offset emissions [14], are seen by companies as an alternative to reducing emissions and are even looked to as a much hoped-for solution by environmentalists for repairing environmental damage [15]. For example, 2016 saw the launch of a carbon offsetting scheme for the air transport sector (CORSIA) that reduces emissions indirectly by investing in emission-reducing projects or purchasing carbon credits. This has been accompanied by consumer research targeting flyers, which attempts to optimize messaging that encourages “voluntary carbon offsetting”, namely, the purchasing of a more expensive ticket in which the extra money pays for the environmental damage caused by air travel [16,17]. Countless other examples exist, which, taken together, indicate large-scale, systematic encouragement of CGBs.

According to Kaklamanou, Jones, Webb and Walker [10], these compensatory incentives could be responsible for rebound effects, such as feeling licensed to fly more frequently because one has paid for carbon-offset tickets rather than the possibly more appropriate response of flying less. This response could be a result of difficulties encountered when making the complex calculations required to determine relative environmental impacts. Further, the Negative Footprint Illusion (NFI) is an increasingly corroborated phenomenon that has indeed provided some empirical evidence that people are unable to accurately evaluate environmental impacts [9,18,19,20,21]. Generally, the NFI emerges when an environmentally friendly addition to a set of conventional items leads to the perception that the carbon footprint of the combined set decreases rather than increases. For example, in Holmgren, Andersson and Sörqvist [19], participants were asked to evaluate the environmental impact of a set of 15 conventional houses as well as a set of 20 houses of which one-quarter were “green” houses (i.e., 5 houses with a smaller carbon footprint). Resulting estimates indicated that participants believed the carbon footprint of the set which included “green” houses had less of a carbon footprint than the smaller set consisting of conventional houses.

A number of suggestions have been made to account for the NFI. On the most basic level, it could be influenced by poor mathematical skills in participants. However, as the NFI was also demonstrated in a study which had only masters-level engineering students for participants [20], this provides some evidence that it might not be related to mathematical ability. It has been posited that an averaging bias could underlie the NFI [10,19], namely that the environmental impact of a set of items which include a green product is calculated as the average rather than the sum of those products (when the calculation would ordinarily have been additive). Beliefs about environmentally friendly labels, such as a preference bias for eco-labelled products like hybrid cars or organic food (see the “eco-label effect” [21,22]), could underlie this averaging bias by inducing a type of “quantity/scope insensitivity” when making calculations of “green” products [23]. In practice, the perceived benefits of the pro-environmental product are believed to cancel far more of the underlying negative effects than they reasonably should. In the case of the NFI phenomenon, this could mean that the participants’ beliefs cause them to underestimate the true environmental impact of the environmentally friendly houses.

Although previous studies have indicated an influence of pro-environmental values on cognitive judgements and performance [24], few studies have measured the relationship between environmental beliefs and the NFI. Therefore, the present study provides the first evaluation of the role that individual differences in environmental beliefs play in cognitive judgements of environmental impacts operationalized in the NFI experiments. The NFI was assessed within subjects using a paradigm in which participants were asked to compare the environmental impact of 15 conventional houses with a set of 30 houses consisting of 15 “green” plus 15 conventional houses. CGBs were assessed using a 15-item scale adapted from Kaklamanou, Jones, Webb and Walker [10] for determining the degree of endorsement of compensatory beliefs and behaviors. The scale includes questions such as “Limiting your household water consumption can compensate for not better insulating your home”, and “Composting food waste can make up for buying imported food”. It was hypothesized that CGBs would explain some of the variance in judgements of environmental impact, and that those rating highly in the CGB group might be more susceptible to the NFI than those with lower compensatory beliefs due to the p.

2. Method

2.1. Participants

A total of 112 participants (46 women) with a mean age of 39.4 years (SD = 15.37) were invited to participate in the study. Participants were convenience sampled from multiple sites in the city of Gävle, Sweden. All subjects provided informed consent and did not receive any reimbursement for participating in the study. As the study did not contain the risk of injury or deal with any physical intervention that could subsequently affect the subject, sensitive personal data were not collected (data were additionally anonymized) and external ethical review was not required in accordance with Swedish law. However, the study adhered to the American Psychological Association’s and the declaration of Helsinki’s ethical guidelines.

2.2. Materials, Design and Procedure

Data were collected in questionnaire format and the study used a within-participant design. The questionnaire took approximately 15 minutes to complete and participants were required to sit in a quiet place without interruption while completing the tasks. The first page of the questionnaire informed participants of the purpose of the survey, namely to investigate estimations of houses’ carbon footprints. The concept of “carbon footprint” was explained and they were additionally informed that a lower carbon footprint resulted in less impact on the environment. On the second page, participants were presented with Figure 1a, depicting 15 yellow houses that represented the so-called “conventional” houses, and asked to “estimate the combined carbon footprint (i.e., environmental impact) of all the houses”. Thereafter, participants were asked to judge the green addition condition, presented as a figure (see Figure 1b) showing the 15 conventional houses with 15 additional “green” houses, accompanied by the following statement and task: “These yellow houses represent 15 houses in a community. The community has decided to build 15 additional houses, all of which will be environmentally certified. Please estimate the combined carbon footprint (environmental impact) the houses have together when the 15 additional environmentally certified houses have been built”.

Participants were also given an example with which to anchor their calculation: “As a point of reference you can consider while making your estimate is that an apartment building with 18 apartments would have scored 5 on the [footprint] scale”. The purpose of the reference point was to give, as far as possible, the participants a common understanding of the scale to decrease the individual interpretation of the scale. The reference point (18 apartments) was deliberately ambiguous so as to recreate the sort of scenario participants would be likely to encounter when trying to evaluate their carbon footprint in the real world, where the information required to make this complex calculation would often be similarly unclear. In the third part of the questionnaire, participants filled in a slightly modified form of Kaklamanou, Jones, Webb and Walker’s [10] Compensatory Green Beliefs questionnaire (see

Table 1. The 15-item Compensatory Green Beliefs questionnaire. Each statement had response options from 1 (strongly disagree) to 9 (strongly agree).

If you have a low flush toilet, then it is okay to use more water in other ways Recycling compensates for driving a car You do not need to worry about which country your food comes from if you use energy-efficient appliances in the home It does not matter how much energy you use if you are on a green energy tariff Limiting your household water consumption can compensate for not better insulating your home It is okay to drink bottled water if you limit the number of car journeys that you make Composting food waste can make up for buying imported food Walking to the supermarket can compensate for buying highly packaged food Having a water butt can compensate for using the oven It is okay to leave electrical goods turned on if they are modern and efficient Not driving a car compensates for flying on holiday Flying abroad can be made up for by being a vegetarian (i.e., not eating meat) If you have energy-efficient electrical equipment, then it is okay to leave it on standby Not using a dishwasher can compensate for taking longer showers It is okay to leave the lights on if you use low-energy light bulbs

Cronbach’s alpha = 0.92

) in which one item was removed from the questionnaire to increase Chronbach’s alpha. The intention of the CGB questionnaire is to determine how prone people are to compensatory thinking.

3. Results

The carbon footprint estimates for the ‘green addition condition’ (M = 5.88, SD = 1.87, 95% confidence interval [5.5:6.2]) did not differ from the estimates for the ‘control condition’ (M = 6.11, SD = 1.38, 95% confidence interval [5.9:6.4]). This was confirmed by a paired sample t-test, t(111) = 1.17, p = 0.245, d = 0.14. These results suggest that the participants did not judge the combined set of 30 conventional and green houses as having a significantly greater carbon footprint than the set of 15 conventional houses, as was predicted. A Bayesian analysis was then conducted in order to determine whether there was any evidence for the null hypothesis, namely that there would be no difference between the two conditions. The estimated Bayes factor (BF₀₁) indicated there was substantially more evidence for the null hypothesis [25], which was 4.918 times more likely than 0.203:1 in favor of the alternative hypothesis.

Difference scores for the two carbon footprint estimate conditions were calculated to provide a measure of participants’ susceptibility to the NFI. Furthermore, the average of the responses to the 15 CGB questions was used to create an index of compensatory thinking. The linear relationship between these NFI-susceptibility difference scores and their CGB results was analyzed using a linear regression model. As illustrated in Figure 2, participants’ CGB scores significantly predicted their susceptibility to the NFI, R² = 0.07, F(1, 110) = 8.72, p = 0.004. Participants’ susceptibility to the NFI increased significantly by 0.39 (β) for every 1 unit increase in their CGB score.

As the regression indicated that a large number of participants had indeed been susceptible to a NFI, the median split depicted in Figure 3. was used in a final exploratory step to the analysis to divide participants into high CGB (n = 57) and low CGB (n = 55) groups so as to test the hypothesis that those with higher CGBs were more susceptible to the NFI. A 2 (display of houses: green addition condition vs. conventional condition) × 2 (Compensatory Green Beliefs group: high vs. low) repeated-measures analysis of variance showed a significant interaction between display of houses and CGB groups, F(1, 110) = 6.85, p = 0.010, $\mathsf{ƞ}$_p² = 0.06. A post-hoc t-test indicated that those falling into the high CGB group [conventional condition mean (SD): 6.11 (0.19); green addition condition mean (SD): 5.4 (0.26)] were susceptible to the NFI; t(57) = 2.64, p = 0.01. However, this difference was non-significant, t(55) = −1.05, p = 0.299, in the low CGB group [conventional condition mean (SD): 6.11 (0.17); green addition condition mean (SD): 6.38 (0.23)].

4. Discussion

The Bayesian analysis confirmed sufficient likelihood that the null hypothesis was true, namely that it was highly likely that the participants did not think that the carbon footprint of the set increased whatsoever when “green” houses were added to the original set of conventional houses. This suggests that participants might assume that a “green” or “environmentally certified” house has net-zero carbon emissions (i.e., are carbon neutral) rather than just being less environmentally harmful than conventional houses. Although not impossible, carbon neutrality of houses is extremely difficult to achieve; most houses, however energy efficient, will still have a carbon footprint incurred through the production of their materials. It could be debated under present global environmental conditions that a truly environmentally certified label should refer to carbon neutrality, but this is unfortunately not yet the standard for certifiability. However, as people seem to assume the best of such a product, this suggests the presence of a “green halo” [26]. This phenomenon, which was initially shown as overly positive attitudes towards organic products (e.g., eco-labelled coffee), is extended here to exaggeratedly positive expectations of the environmentally friendliness of “green” houses, which influence subsequent carbon footprint calculations.

Previous research on the NFI has proposed an averaging bias as the mechanism underpinning the illusion (e.g., [9]), possibly as a result of people misapplying the previously mentioned social balancing heuristic proposed by Sörqvist and Langeborg [11]. The “averaging account” is also supported by research indicating that the NFI is not dependent on the number of “green” items (e.g., [21]), which it would be if any type of additive calculation was being made. It is further reasoned that the presence of environmentally friendly versus non-environmentally friendly buildings might nudge participants into a vice–virtue categorization (e.g., good vs. bad, healthy vs. unhealthy, and eco-friendly vs. conventional; see [27]), influencing people to estimate the average carbon footprint of the set rather than their aggregated sum. The present paper argues that compensatory thinking, as measured by the CGB scale, contributes to this faulty reasoning process because the process of calculation is sensitive to people’s beliefs. CGBs are based on an erroneous assumption that all people are allotted a certain amount of “acceptable” environmental damage, and that one’s allotment of “acceptable” damage can be increased in proportion to how much one compensates. This type of judgement could be susceptible to an averaging bias; when participants are asked to evaluate the acceptability of having a car if one also recycles, they might also evaluate the average of a single set consisting of an environmentally harmful product/behavior plus an environmentally beneficial product/behavior. An alternative explanation is that those with higher CGBs believe that environmentally friendly products are carbon neutral or even carbon negative, and therefore cleanse the effects of non-environmentally friendly behaviors. In the latter case, even if the resulting calculation was additive rather than the result of an averaging process, the neutral or subtractive effect of the “green” buildings would lead to equal estimations or a NFI between conditions.

It was therefore hypothesized that those with higher potential for estimating “green” houses as carbon negative or carbon neutral might (i.e., susceptible to the NFI) would also be more likely to believe that environmentally harmful behavior could be compensated for. In such a case, susceptibility to CGBs would be more likely to be correlated with the NFI. Indeed, the findings of the regression analysis comparing CGBs to the carbon footprint estimate difference scores indicated that higher CGBs were significantly related to increasing susceptibility to the NFI. Furthermore, the results of the subsequent explorative ANOVA analysis show that those who were especially susceptible to the NFI clustered in the high CGB group. In the low CGB group, almost identical means for the two conditions showed that this group’s participants still believed that the “green” houses were carbon neutral. Although a causal connection cannot be determined through such a study, the implication could be that cognitive processes underlying the NFI and CGBs share similarities, and that compensatory thinking could therefore contribute to the resulting footprint estimate.

Different from most NFI investigations, the present study used a within-participant design and purposefully chose not to counterbalance presentation conditions, both methodological factors which could have had some bearing on the results. In previous studies, the NFI is diminished or absent when using a within-participant design, but it was not possible in the present study due to the correlational design in which each participant’s NFI-susceptibility difference score needed to be matched with their performance on the CGB scale. However, the NFI effect would be more robust if it were to be present in a within-subjects study design for the following reason. If participants were presented sequentially with both conditions (in either order), this would allow for comparison between the two conditions, potentially drawing attention to the differences in the quantity of houses and thereby lessening the chances of quantity insensitivity and an accompanying averaging bias. In sum, it could encourage the conclusion that an additive calculation was required to accurately estimate the footprint. The strength of a within-participant design was that those who still demonstrated a NFI effect could more certainly be said to have done so in error.

Furthermore, although it is customary in studies to counterbalance conditions so as to avoid unwanted order effects, counterbalancing is counterproductive when the order of presentation of conditions is important, or when counterbalancing could lead to very different outcomes and thereby obscures the results. For example, in counterbalanced NFI studies, half the participants see the conventional condition first and the green addition condition second, while the other half see the green addition condition first and the conventional condition second. In the first group, participants judge the footprint of conventional houses and then calculate the additional footprint of the “green” houses which are presented next. In the second group, participants judge the whole group of “green” and conventional houses together first and then subtract the footprint of the “green” houses from the conventional houses second. These two quite different calculations would have different predicted outcomes. In order to control for this difference, we chose to specify the order of presentation in the present experiment, as we primarily wanted to know what happens when people add environmentally friendly items to a set consisting only of conventional items.

5. Conclusions and Future Directions

The current investigation of the relationship between CGBs and the NFI represents the first demonstration of the role of individual differences in cognitive processes and the NFI, suggesting that this phenomenon is partly related to beliefs about “green” products and behaviors. Having accounted methodologically for potential points of confusion, the fact that the NFI could be predicted by hypothesized cognitive processes (i.e., CGBs) provides further evidence for the validity, reliability and robustness of the NFI phenomenon, which would be unlikely to co-vary with CGBs had participants, for example, misunderstood the task. Therefore, we suggest that participants’ estimates were either the result of an averaging bias or believing that a “green” label on a house equates to it being carbon neutral or carbon negative. That people might presume something with a “green” label to be carbon neutral or even carbon negative can have potentially damaging implications, the most obvious being the rebound effect in which people assume that green products and services can be used with impunity. For example, the frequent flyer might believe that they can fly twice as much because their plane’s carbon emissions are offset by the purchasing of carbon credits. In extreme cases, they might even think they are making a positive contribution to the environment by their behaviour, considering that flying directly leads to the financing of projects involving the extensive planting of forests elsewhere. The potential for such a view to be exploited by industry is large, highlighting the fact that the use of such a label therefore comes with an added responsibility to ensure that the actual environmental costs of manufacture and usage are highly transparent. In conclusion, modelling the respective or combined roles of compensatory thinking and other cognitive processes in inaccurate estimates of carbon footprints represents a promising future direction for NFI research.