Inconsistency between attitude and behavior is a major obstacle to research on the predictive power of attitudes on behavior. To clarify the mechanism underlying such inconsistency, we combined event-related potential [ERP] and questionnaires to explore the relationship between contextualized attitudes and retrospective attitudes in the context of illusion of privacy empowerment [IPE]. When the participants read the IPE events [including platform empowerment intention, technique, result, etc.] on slides, we measured retrospective attitudes with questionnaires and recorded contextualized attitudes with ERPs. We found that individuals’ retrospective attitudes were different from contextualized attitudes: retrospective attitudes were mainly affected by the individual’s analytic system, while contextualized attitudes were mainly affected by the direct stimulus-response [i.e., heuristic system]. Therefore, retrospective attitudes may not accurately reflect individual cognition in the immediate context, and inconsistency between attitudes and behavior may be caused by the mismatch between retrospective attitudes and immediate behavior. Our findings provide a more reasonable account of the relationship between attitudes and behavior.
Keywords: attitude-behavior inconsistency, contextualized attitude, retrospective attitude, illusion of privacy empowerment, event-related potentials
1. Introduction
Attitudes and behaviors have long been a focus of psychological research [,,], with researchers primarily investigating the relationship between attitudes and behaviors to reveal the predictive power of attitudes on behavior [,]. However, numerous studies have demonstrated that individuals often express their attitudes but do not act in accordance with those attitudes []. For instance, some individuals may have negative views of smoking but still engage in smoking behaviors, some may state their intention to lose weight but eat high-calorie food, and some users may be concerned about privacy but are willing to sacrifice it for convenience.
The existing explanations for attitude–behavior inconsistency mainly focus on exploring the characteristics of attitude. First, based on the theory of attitude strength, it is argued that when attitude strength is strong, the attitude is clear and can help individuals resist persuasive messages and maintain a high degree of behavioral consistency. When attitude strength is weak, the attitude is vague, and therefore no corresponding behavior will be produced [,,]. Second, based on the theory of planned behavior, it is argued that subjective norm and perceived behavioral control will both affect the predictive role of attitude on behavior [,,]. Third, based on the system view of attitude, it is argued that attitude is overall. When one component of an individual’s attitude [i.e., affective, cognitive, or behavioral] is inconsistent with the overall evaluation or other components, attitude ambivalence exists, and attitude is difficult to guide behavior [,,,].
Although these studies have explained the inconsistency between attitude and behavior from multiple perspectives, there are still some problems due to the limitations of research perspectives and methods. First, the importance of specific contexts has been ignored, and the relationship between attitude and behavior is only discussed in the background of psychological systems. Second, the existing studies mostly adopt questionnaire measurement methods, but the questionnaire results are only hypothetical responses to hypothetical situations, which are overall evaluations after individuals’ retrospective, processing and rational analysis, subject to recall and the influence of individual subjective factors []. It is difficult to scientifically and accurately reflect the cognitive process in a real situation.
The illusion of privacy empowerment [IPE] means that platforms grant users invalid privacy control rights and data sovereignty, giving them the illusion of autonomous choice. For instance, even if they refuse to authorize their contacts, they may still be recommended related friends, or nearby information even if they refuse to authorize their location []. As IPE emerges, users claim to be concerned about it, yet they still accept the recommended information with “manipulative” attributes, which is a new attitude–behavior inconsistency phenomenon in the digital era.
Event-related potential [ERP] technology, which has a millisecond-level high temporal resolution, can better simulate immediate privacy decision-making scenarios. In addition, this technology is viewed as a “magnifying glass” for observing psychological processes without directly asking users about their thoughts, memories, evaluations, or decision strategies [], which can precisely locate individual cognitive processes without being easily influenced by individual subjectivity. Different from questionnaire measurement, neuroscience experiments provide different research questions that can be solved. The results of questionnaire measurement are based on the individual’s comprehensive analysis and are mostly general and holistic perceptions and evaluations []. ERP technology is used to measure behaviors and cognitive processes directly generated by stimulus-response, which is based on heuristic processing and is the embodiment of instantaneous and emotional attitudes [].
Based on this, this study takes IPE as its research context and employs a combination of questionnaires and ERP to measure the willingness to accept recommendation advertisements as a reflection of retrospective attitudes and cognitive resources such as attention to reflect contextualized attitudes. The study aims to explore the relationship between contextualized and retrospective attitudes to explain the inconsistency between individual attitudes and behaviors.
2. Literature Review and Research Hypothesis
2.1. Attitude–Behavior Inconsistency and the IPE
The attitude–behavior inconsistency has received widespread attention from scholars once it was proposed. Scholars have attempted to identify the causes of inconsistency, explore the factors restricting the relationship between attitude and behavior, and clarify the role of attitude in predicting behavior [,,].
The research on attitude–behavior inconsistency mainly focuses on exploring the characteristics of attitude itself. First, the Dual Attitudes Model states that people can simultaneously hold two different evaluations for the same attitude object, one being an automatic, implicit attitude, and the other being an aware, explicit attitude []. Studies have demonstrated that behavior is the result of the combined effect of implicit and explicit attitudes. When the motivation level of the explicit attitude is low, individuals will take actions that are inconsistent with it [,,]. Secondly, the Tripartite Model of Attitude Structure viewpoint suggests that attitude consists of cognition, emotion, and intention, and the consistency of evaluation among these components should be taken into consideration when examining the relationship between attitude and behavior [,,]. Thirdly, the concept of attitude strength proposes that strong attitudes will affect the selection and processing of information, impede attitude changes, and influence the duration of attitude, thus exhibiting better behavior prediction ability [,,].
However, some studies have also shown that individual behavior is often a product of the interaction between attitudes and their perceived context, arguing that even a specific attitude is not guaranteed to determine a person’s behavior in a given situation. It has led to the concept of contextualized attitudes, which refers to forming immediate attitudes based on changes in context [,,]. Although the critical role of context in attitude–behavior inconsistency has been recognized, the limitations of research methods have made it difficult to further explore this topic.
The essence of the IPE is to manipulate users covertly by giving them the illusion of privacy control and autonomy of choice, then depriving them of their rights and interests. Individuals often show inconsistent attitudes and behaviors in the face of IPE. In IPE situations, users’ actual decision-making behavior is specific and immediate [], and is only influenced by contextualized attitudes induced by contextual cues. Therefore, this study takes IPE as the research context to investigate the relationship between the contextualized attitude and retrospective attitude induced by different regulatory focus and social distance. By measuring the changes of contextualized attitude and retrospective attitude, this study aims to provide a new explanation for the individual’s attitude–behavior inconsistency phenomenon.
2.2. Regulatory Focus and Social Distance
Regulatory focus and social distance play a key role in affecting individuals’ privacy decisions. Therefore, scholars often use them in studies on privacy disclosure, information risk, and privacy-protective behavior [,,].
Regulatory focus is an individual’s tendency to change or control their thoughts and reactions in a specific way and tendency to achieve a goal. It is divided into promotion focus and prevention focus [,,]. Individuals with promotion focus are more likely to rely on emotional factors, focus on the potential benefits, and often ignore potential risks [,,]. In contrast, those with prevention focus are more cognitively oriented, tend to avoid negative outcomes, and are more likely to search for relevant material and analyze potential risks rationally, even if they can benefit [,,]. Therefore, it is hypothesized that promotion-focused individuals are more likely to change their attitudes toward privacy events when learning content related to privacy events because of the apparent hazards, while prevention-focused individuals will sort and judge the learned content, and their attitudes are less likely to change.
Based on this, this paper proposes the following hypothesis:
H1.
The change in willingness to accept recommended information before and after learning of IPE events is more significant in individuals with promotion focus than in those with prevention focus.
Social distance is a concept that refers to the perceived distance between an individual and other object in their environment. When making behavioral decisions, individuals frequently consider the proximity of interaction objects to themselves [,]. When the social distance is distal, individuals have fewer fair considerations and fewer emotional responses, which allows them to make rational decisions. When the social distance is proximal, individuals have higher expectations of social norms. When they are treated unfairly, triggering emotions such as surprise and frustration that usually lead individuals to make irrational decisions []. Compared to promotion-focused individuals, prevention-focused individuals who want to change their IPE attitudes require more redundant information to refine them. However, even in the case of proximal social distance, it is not easy for individuals to obtain enough information, making it difficult to have a large attitude change.
Based on this, this paper proposes the following hypotheses:
H2.
In the promotion focus group, individuals’ willingness to accept recommendation information exhibits a greater change in the proximal social distance group than in the distal social distance group.
H3.
In the prevention focus group, social distance has no significant impact on the change of willingness to accept recommendation information.
2.3. EEG Component Hypothesis
According to Cognitive Resource Theory, human cognitive resources are limited and individuals allocate certain cognitive resources to learning and evaluating each task [,]. When the individual is highly invested in the learning task, all available cognitive resources are occupied, making them less able to effectively inhibit task-irrelevant stimuli and more susceptible to interference from irrelevant stimuli. Conversely, individuals with a low level of engagement in the learning task have sufficient cognitive resources to suppress the interference of irrelevant stimuli []. Therefore, the present study employed the ERPs technique to scan and record subjects’ EEG signals using the single-stimulus experimental paradigm, replacing the non-target word sounds in the oddball paradigm with silent voice. The N1 and P2 component wave amplitudes reflect the degree of attentional distraction from irrelevant factors during the learning process, thereby laterally reflecting the input of cognitive resources in the individual learning process from another angle.
The N1 and P2 components have been reported to be related to attentional input and risk perception. As their attentional alertness increases, decision-makers are proposed to dedicate more attentional resources to stimuli. In addition, a higher level of attentional alertness is suggested to be associated with a greater N1 amplitude []. Meanwhile, the more attentional resources an individual puts into information processing, the higher the P2 amplitude is expected to be. It has also been claimed that negative stimuli may necessitate more attentional resources, leading to a higher P2 amplitude []. Moreover, when it comes to assessing the riskiness of safety sign warnings, research has shown that high-risk warnings tend to generate more P2 amplitude than low-risk warnings [].
Both promotion and prevention focus individuals increased their attentional alertness in the face of the negative event of IPE, so there were no significant differences in N1 and P2 wave amplitudes between the promotion and prevention focus groups. Promotion-focused individuals are susceptible to simple information cues that do not occupy or occupy less cognitive resources []. Thus, the hazards embodied by social distance have no significant effect on them. On the other hand, prevention-focused individuals devote a large number of cognitive resources to identifying and evaluating the presented information [], and proximal social distance events are more likely to cause them to perceive the threat and thus devote more attentional resources to them.
Based on this, the study proposes the following hypothesis:
H4.
There is no significant difference in the N1 amplitude between the promotion focus group and the prevention focus group.
H5.
For promotion-focused individuals, there is no significant difference in the N1 amplitude between the distal distant group and proximal social distance group.
H6.
For prevention-focused individuals, the N1 amplitude in the distal social distance group is higher than that in the proximal social distance group.
H7.
There was no significant difference in the P2 amplitude of the promotion focus group and the P2 amplitude of the prevention focus group.
H8.
For promotion-focused individuals, there is no significant difference in the P2 amplitude between the distal distant group and proximal social distance group.
H9.
For prevention-focused individuals, the P2 amplitude of the distal social distance group was higher than the P2 amplitude of the proximal social distance group.
The research model in this study is shown in Figure 1.
![An external file that holds a picture, illustration, etc. Object name is behavsci-13-00223-g001.jpg][////i0.wp.com/www.ncbi.nlm.nih.gov/pmc/articles/PMC10045271/bin/behavsci-13-00223-g001.jpg]
3. Materials and Methods
3.1. Participants
In this study, subjects willing to participate in the EEG experiment were randomly recruited through various social platforms, posters, and other publicity methods. Applicants were required to fill out a questionnaire covering personal information, contact information, regulatory focus items, and willingness to accept the recommendation information items. One hundred and sixty questionnaires were collected during the preliminary recruitment process. Among respondents, 88 were males and 72 were females, with an average age of 21 years old; 61.25% were undergraduate students, and 38.75% were graduate students. The regulatory focus questionnaire was adapted from the Regulatory Focus Questionnaire [Chinese Version] revised by Yao et al. [2010] [], which is suitable for the Chinese context. It contained 10 items, including 6 promotion focus items and 4 prevention focus items. The score of regulatory focus is the score of the prevention focus questionnaire minus the score of the promotion focus questionnaire. The higher the score is, the more prevention-focused the individual tends to be, and the lower the score is, the more promotion-focused the individual tends to be. The scale of willingness to accept recommendation information is based on the willingness to accept recommendation advertisements. It includes three items: “I would be willing to accept information sent to me by this platform”, “I would be willing to click and browse information sent to me by this platform”, and “I would consider sharing information sent to me by this platform with others” []. Seven-point Likert scales were employed in both questionnaires, ranging from 1 [strongly disagree] to 7 [strongly agree].
G* Power3.1 was used to calculate the sample size required for research []. According to Cohen’s standard [], F tests were used as the test family, and the parameters were set as follows: Repeated measures, within-between interaction, Effect size f=0.25, α err prob =0.05 Power 1−β err prob=0.8, Number of groups =4, Number of measurements=3, Corr among rep measures =0.5, Nonsphericity correction ε =1. The total sample size was 40. So, grouped by the level of regulatory focus scores, 20 participants with promotion focus and 20 participants with prevention focus were selected.
Then the participants of the EEG experiment were selected according to the questionnaire results. The specific selection method was as follows: We first ranked the regulatory focus scores of 160 respondents from highest to lowest, selected the 20 respondents with the highest scores as the participants in prevention focus group, and selected the 20 respondents with the lowest scores as the participants in promotion focus group. Among them, the score for promotion focus group was between −1.08~0.83 points, M=−0.17, SD=0.58, and the score for prevention focus group was between 1.25~2.41 points, M=1.59, SD=0.36. Among the participants, there were 17 males and 23 females, with an average age of 23. Undergraduate students accounted for 57.5%, and graduate students accounted for 42.5%.
The promotion focus group and the prevention focus group were randomly divided into two groups: the distal social distance group and the proximal social distance group. There was no significant difference in regulatory focus scores of the distal/proximal social distance participants in promotion focus group, t18=−0.146, p>0.05. There was no significant difference in the regulatory focus scores of the distal/proximal social distance participants in prevention focus group, t18=−0.312, p>0.05. There was no significant difference in age among the four groups, F3,27=0.727, p>0.05. All subjects were right-handed, had no history of mental illness, and had normal or corrected-to-normal vision.
3.2. Procedure
We randomly selected 40 events [intentions, methods, and consequences] with distal social distance and 40 with proximal social distance. Subsequently, 50 participants were randomly chosen to read the materials and respond to the question of “To what extent do you think this event is an IPE event?” on a seven-point Likert scale [ranging from 1 [strongly disagree] to 7 [strongly agree]]. The 20 materials with the highest scores were selected for the formal experiment.
Formal Experiment
In the single-stimulus experimental paradigm, this experiment used a silent speech to replace the non-target word sounds in the oddball paradigm, recorded the participants’ attention and cognitive resource input in the reading process, and explored the cognitive neural mechanisms in the real-time situations from the cognitive and attentional perspectives. The experiment adopted an inter-group design based on regulatory focus [promotion/prevention] * social distance [distal/proximal].
Participants were asked to watch a 10-min slide with an electroencephalogram [EEG] cap on. The content of the slide is typical IPE materials selected by the subject personnel through a strict manipulation test. Participants in the proximal social distance group read the domestic IPE events, while participants in the distal social distance group read foreign IPE events. The learning time, environment, content, and cognitive ability of the participants were all controlled at the same level.
Participants were asked to watch the content in the slide carefully. They were asked to fill out a questionnaire about their willingness to accept recommended information at the end of the experiment. Moreover, they were asked to recall the content of the IPE event in the slide and record it on paper in the form of keywords to measure their perception of the exogenous learning content.
3.3. Data Collection and Analysis
An American Neuroscan EEG recording and analysis system with a 64-channel Ag/AgCl electroencephalogram cap [according to the extended international 10–20 system] was employed in this study. The filter bandpass was set to 0.01–100 Hz, the sampling frequency to 100 Hz, and the resistance was maintained below 5 kΩ throughout the experiment. Both VEO and HOE were recorded.
Prior to the experiment, the participants were instructed to sit comfortably in a quiet, softly lit room with sound insulation. The distance between the participant’s eyes and the computer screen was approximately 1 m, and the horizontal and vertical viewing angles were set to less than 5°. During participants’ reading materials, a 1000 Hz [100 ms duration, 10 ms rise/fall time, and 60 dB SPL] audio stimulus was delivered binaurally at a distance of 60 cm from each ear, according to the single-stimulus experimental paradigm.
4. Results
4.1. Behavioral Data
t-test was carried out to compare the 40 participants’ willingness to accept the recommendation information before and after the experiment. There was no significant difference for the pre-experimental willingness to accept information between the promotion focus group M=4.06, SD=0.87 and the prevention focus group M=3.86, SD=0.73, t38=0.82, p=0.420>0.05, see Table 1. The willingness to accept recommendation information in the promotion focus group [t19=6.15, p