Influence of Emotions on the Acceptance of an Oil Drilling Project

Method

Results

Descriptive statistics among the variables included are displayed in Table 1. The results indicated that most participants held interdependence beliefs and clearly rejected oil exploration. Also, participants disagreed with the legal procedure carried out, considering that they had received limited information about the project. Moreover, they perceived more risks than benefits and had a tendency to experience negative emotions to oil drilling (Table 1).

OPEN IN VIEWER

Table 1.

Descriptive Statistics for Included Variables.

	M	SD	Range
Environmental beliefs	8.74	1.58	0-10
Risk perception	5.62	3.61	0-10
Benefits perception	3.53	2.39	0-10
Procedural justice	2.15	2.26	0-10
Information	3.77	2.69	0-9
Negative emotions	5.94	2.95	0-10
Acceptance	2.62	2.53	0-10

The proposed model in Figure 1 was checked by SEM. In the first step, results indicate a satisfactory fit of the model for the entire sample (Baseline Model). Although a significant χ²(334, N = 559) = 833.94, p < .001, was obtained, the other fit indexes led to acceptance of the model: root mean square error of approximation (RMSEA) = .052, 90% confidence interval (CI) = [.047, .056]; normed fit index (NFI) = .93; non-normed fit index (NNFI) = .95; comparative fit index (CFI) = .95. The covariance of errors of benefits perception and procedural justice had to be released (permit values not equal to 0) to achieve the model fit, indicating a joint error variance not explained by our model. The covariance of the measurement errors of Items 1 and 2 of procedural justice (Items 1 and 2 in the procedural justice scale) were also released. In structural equations models, the exogenous latent variables can covariate freely (phi parameters) but the endogenous latent variables cannot covariate at all (by model). Nevertheless, these endogenous latent variables have not explained variance (error variance or diagonal psi parameters) than can covariate (psi parameters out of the diagonal), probably because they have a common source of not explained variance. The effect of incomes over acceptance, age over negative emotions, information over negative emotions and over risk perception, risk perception over acceptance, and environmental beliefs over acceptance, predicted in Figure 1, were nonsignificant and were eliminated in the final model.

This model was used in the second step with multigroup estimation to check whether the proposed model was the same for males and females. First, a Configural Model (two groups without any constraints) was estimated. Although the fit indexes were now smaller, they were still adequate: χ²(668, N = 554) = 1,257.6, p < .001, RMSEA = .056, 90% CI = [.052, .061]; NFI = .89; NNFI = .94; CFI = .95. Subsequently, a model with constrained factor loadings (Measured Model) for both groups was estimated, showing no decrease in fit. The difference in the chi-square in the two multiple-groups models was not significant (χ² = 25.308, df = 19, p > .05). After measurement model invariance was established, structural differences were examined (Structural Model). The previous model was compared with a new one with loading and regression parameters constraints. The differences between this Structural Model and the Configural Model was nonsignificant after five structural parameters were released (χ² = 8.4781, df = 9, p > .05). The five structural parameters with significant difference between males and females were as follows: environmental beliefs over risk perception and age over acceptance were nonsignificant for males; environmental beliefs over benefits perception, incomes over negative emotions, and information over acceptance were nonsignificant for females. The final Structural Model and these differences are shown in Figure 2, a slash (/) between values indicating when there are differences between males and females.

OPEN IN VIEWER

Figure 2.

Structural Model obtained.

The model accounted for adequate proportions of variance in both endogenous final factors, for both males/females: negative emotions (53/50%) and acceptance (65/71%). Environmental beliefs were a significant predictor of benefits perception for males but not for females, while environmental beliefs were a significant predictor of risk perception for females but not for males. However, environmental beliefs were a significant predictor of procedural justice, for both males and females. For both groups, the β value for environmental beliefs over negative emotions, though significant, was very low (.07), and there was no direct effect of environmental beliefs over acceptance. Benefits perception, risk perception, and procedural justice were significant predictors of negative emotions, although Benefits perception explained the highest percentage of variance of negative emotions with a β value of −.57. The contribution of risk perception to the variance of negative emotions was significant, but so low (.10) that it might be attributed to a power excess due to large sample size rather than to a real influence (spurious effect). Incomes were a significant predictor of negative emotions only for males. Information showed a significant relation with benefits perception, procedural justice, and acceptance, but in the latter, only for males. Acceptance was largely explained by negative emotions with a β value of −.29, and benefits perception with a β value of .48. The contribution of procedural justice to the variance of acceptance was no longer significant when the two groups (males and females) were included. Finally, the contributions of Information to the variance of acceptance was significant only for males, and age only for females, although in both cases, they were so low that, once again, we considered the need for future studies to ascertain whether they might be attributed to an excess of power because of large sample size and hence an unavoidable spurious effect.

Finally, a multigroup SEM moment analysis (means comparison between groups) allowed us to compare group means differences in every latent variable. None of six latent variable means differences were significant between males and females (p > .05). Figure 2 shows all the β values and the direction of relations between the factors (see Online Appendix B for correlations, means, and standard deviations of observable variables for males and females).

Discussion

In the current study, we tested a theoretical model that explained the acceptance of oil drilling based on psychosocial factors. In general, the results obtained indicate that the local community showed a low level of acceptance of this kind of energy project. This is consistent with the results obtained in the previous survey undertaken by the Canary Islands government also involving a population from different islands (EFE, 2014), thereby demonstrating the validity of the results.

The proposed model was adapted for the entire sample and separately for male and females, although there were small differences between each group. For females, beliefs about interdependence were associated to risk perception but not benefits perception; the opposite occurred for males. Similarly, income was associated to negative emotions for males but not for females. Age does not affect the final acceptance of the project for males, but carries a certain weight for females. Finally, the received information explained part of the variance of acceptance for men, but had no effect in the case of females.

Therefore, the final model is similar for males and females, in terms of the acceptance, or not, of oil drilling: benefits perception and negative emotions were what explain acceptance in both groups. Likewise, no differences were found in the means of any of the variables in the study. Therefore, it cannot be said that these results show females to have more negative emotions associated with the oil drilling project, as found in studies on everyday emotions (Mackinnon et al., 1999), nor that there are differences in the level of acceptance of the project for men, as may have been suggested in previous studies (Devine-Wright, 2007). Age and income were not shown to be important in the final variance of the acceptance of oil drilling. The weight of demographic variables in the acceptance of environmental projects continues to be contradictory, as already indicated by other studies (Carlisle, Feezell, Michaud, Smith, & Smith, 2010; Devine-Wright, 2007).

With this model, we want to highlight the importance of emotions for the acceptance of an energy project, as many studies on decision taking and pro-environmental behaviors have only included cognitive variables. Based on Ellis’ (1973) model and appraisal theories of emotions (Moors et al., 2013), we have considered emotions to be the bridge between the beliefs and perceptions of individuals, and the decisions they make. Although cognitive variables can also exert a direct influence on the acceptance of a project, the emotions aroused by the interpretation of that proposal have been shown to be an important factor behind the acceptance/rejection of the oil drilling project. Indeed, in the final model, the load of risk perception, environmental beliefs, and procedural justice was only significant on negative emotions and not directly on acceptance.

These results concur with those of other research on the importance of emotions in environmentally relevant behavior (Carrus et al., 2008; Chaudhuri, 2002; Durán et al., 2007; Durán, Ferraces, Rodríguez, & Sabucedo, 2016). Emotional bonds, such as place attachment, have also been related to acceptance of environmental intervention (Devine-Wright, 2009) or the human response to natural disasters (Ruiz & Hernández, 2014). For example, it has been shown that risk perception is not only based on physical elements that might be dangerous but also on other emotional, social, and cultural factors that may affect end perception and alter how people adapt to the demands of the situation (Gaillard, 2008; Tobin et al., 2011). The intensity of the emotional response has also been linked to the proximity of residence to the environment where the changes take place (Ruiz & Hernández, 2014) or to the magnitude of the changes that have or are expected to occur (Lai & Kreuter, 2012).

In our study, the other factor associated with acceptance and, at the same time with this negative emotional state, was benefits perception: The fewer the benefits perceived, the greater the negative emotional response and lower acceptance. Some authors have found that support for energy projects depends largely on the economic benefits that they bring to the community (Bidwell, 2013). In our study, the local population perceived few benefits, whether economic or otherwise, from oil drilling, hence the general rejection and nonacceptance of the project. Perceived benefits had greater impact on acceptance than perceived risks, which was consistent with previous studies (Visschers & Siegrist, 2014).

However, negative emotion was not explained only by the perception of benefits. Perceived procedural justice, the extent to which participants considered that legal and information procedures and collective participation had been adequately managed, contributed to a higher negative emotional response. Our results showed that the local population was not satisfied with the management of energy issues by the different levels of government (procedural justice). As we have said previously, much of the rejection was due to the modus operandi of the project developers rather than the project itself. The proportion of individuals who would support an energy project increases when the proposal comes from local government rather than private developers (Firestone & Kempton, 2007). The oil drilling project in the Canary Islands was proposed by private business and backed by the national government. However, local government had voiced its objection to the project and its support for renewable energies in the islands. The information transmitted by the media may have encouraged a negative opinion about benefits of the project and the procedure followed to introduce it. The need to release the covariance of errors of benefits perception and procedural justice to fit the model shows that these two variables may both be related to others not included in the model. Future studies could explore this further.

Moreover, in our study, the population did not perceive a high environmental risk associated with the oil drilling project. In fact, this factor explained a small part of the negative emotions, and the relationship with acceptance is not significant. Previous research has found that people evaluate risk through affective reactions or rational analysis (Chaudhuri, 2002; Tobin et al., 2011). When people have positive or favorable feelings toward a situation, they judge risks as low and benefits as high; an opposite pattern occurs when negative or unfavorable feelings appear (Finucane, Alhakami, Slovic, & Johnson, 2000). In our study, emotion and perceived benefits seems to have been the salient factors in the assessment of whether or not to accept the oil drilling project over and above the real or perceived risk.

The other variable that showed a direct effect on the acceptance of the oil drilling project was the assessment of the amount of information available to participants, but only for males. Our results confirm a positive—albeit minimal—relation between these variables. Our results therefore coincide with prior studies that have shown that the information provided has a direct relationship with support for technology (Cacciatore et al., 2012; Frank et al., 2011). What remains clear in our study is that information levels were positively linked to perceived benefits and procedural justice. Appropriate and reliable information may enhance acceptance of this type of intervention. It is important to bear in mind that participants mainly relied on information from the local media, which highlighted the fact that most island residents rejected this kind of activity. This information may have acted as a perceptual filter for the evaluation of the perception of benefits and attitudes toward technologies, as shown in other studies (Binder et al., 2011).

Contrary to what was expected, interdependence beliefs had little bearing on the model. No direct relation was found between beliefs and acceptance of the oil project, although a moderate connection between beliefs, risk and benefits perception, and procedural justice perception was observed, but with differences by sex. Beliefs of interdependence led to increased risk for women and a decrease in benefits for men. Previous studies pinpointed the influence of beliefs on behavioral intention and decision making (Arroyo-López, 2012; Durán et al., 2016; Steg et al., 2005; Stern et al., 1999). In our study, however, beliefs only exerted an indirect effect. People who mostly believed that there should be an interdependent relationship between humans and the environment were more sensitive to the possible risks or absence of benefits to the environment of an oil drilling project. Therefore, to a certain extent, environmental beliefs influence people’s interpretation of the situation, their perception of the performance of the agents involved and the possible consequences, and how information about a new environmental project is organized (Sjöberg, 2000).

Furthermore, a minor direct relation between beliefs and emotions was found, but beliefs essentially were related to emotions through perceived benefits and risks, and procedural justice. The interpretation of the reality in which we are immersed affects the emotions we feel (Briñol et al., 2010). In this case, as we have already seen, beliefs in interdependence affect the perception of reality regarding risks, benefits, and procedural justice. The greater the risks, the fewer the benefits, and the less procedural justice perceived, the more intense will be the emotions aroused by oil drilling.

A possible limitation to be taken into account in the interpretation of the results is the low level of acceptance of oil drilling by most participants. These results are therefore related to a context of low acceptance of an energy project, and the weight of emotions that we found may not be replicated in a situation in which the population has no prior negative attitude. It is therefore necessary to verify the weight of emotions in contexts of positive assessment of the project. In our study, the emplacement of oil drilling in islands whose economy is based on tourism and where these installations would have a visual impact on the landscape could be a compelling reason for rejection of the project. This may limit extending the results to other contexts. Furthermore, despite our attempt to produce a sample of all the islands, with sufficient variability depending on sex, age, and socioeconomic level, the sample cannot be considered representative, thus restricting the generalizability of the results.

Another limitation to consider is the measurement used to evaluate information about oil drilling. We measured the extent to which participants perceived that they had been given sufficient information but not the objective amount of information that they might have had. In future research, this variable should be measured with more indicators. It would also be interesting to analyze whether participants are given information about community benefits in terms of creating jobs, income, and about specific risks, as well as the sources of such information, thereby improving measures of risk and benefits perception. However, we believe that this model has enabled us to establish that benefits exert greater effect on acceptance than risk perception and that emotions seem to play a central role in decision making in the face of environmental change.

Useful future research would involve further qualitative exploration of individuals’ reasons for opposing such projects (whether they trust the veracity of the benefits and costs of the project, whether there is a clearly negative attitude to traditional energies in favor of renewable energies). This would improve the measurement of acceptance, which is rather limited in this study, which included only two questions. The inclusion of more specific measurements of behavioral support would also be advisable.

In our study, the perception of the unfair nature of the process had a moderate effect on rejection of the project, and when males and females are compared, this effect was not significant. Some studies have found that the outcomes fairness, perceived benefits, and general attitudes are more related to acceptance of an energy project when compared with procedural fairness (Visschers & Siegrist, 2012). Therefore, it is important that the project lead to a favorable outcome for the population, whose participation should be counted upon before it is launched. Future analysis could examine whether there was more acceptance from those who actually participated in decision making compared with those who had not. In this study, it was not possible to check whether participants had been called upon by the agents involved in developing the project to ensure they would have meaningful participation in the decision making.

In sum, this study has shown that several factors can influence public opinion about an oil drilling project. Negative emotions and low perceived benefits generated a rejection of this kind of project. In addition, perceived benefits and procedural justice and, to a lesser extent, perceived risk contributed to the negative emotions. Appropriate information would create a more favorable attitude to an oil drilling project. Environmental changes relating to energy management require a precise knowledge of the specific action, as well as a better understanding of the effects on the population. The engagement of the local community in environmental conservation and sustainable tourism may have been determining factors for the negative emotions aroused, and that led to the rejection of this type of drilling.