It Is Probably Fake but Let Us Share It! Role of Analytical Thinking,Overclaiming and Social Approval in Sharing Fake News

Abstract

This study investigates whether and how analytical thinking, overclaiming, and social approval are associated with the intention of sharing fake news on social media. To randomize each respondent to a group and treatment and to test of several hypotheses simultaneously, two by two factorial design was used. An online survey (N = 1160) on Iranian social media revealed that overclaiming and social approval are positively related to sharing fake news on social media. Surprisingly, analytical thinking yielded no significance. We believe that in order to show more knowledge users tend to share information with high social approval irrespective of their credibility. Although CRT proved no relation with sharing, significant differences among male and female users were found. The proven relation between sharing more and overclaiming more reveals a marketing opportunity. Gamification of communication which provides a vehicle for users to overclaim their knowledge to their peers on social media might be a suitable strategy on social media to spread the message.

Keywords

Fake news analytical thinking overclaiming social approval social networks

Introduction

Fake news and post-truth are now familiar terms in different strings of research. The 2016 American presidential election (Allcott & Gentzkow, 2017; Fallis & Mathiesen, 2019; Molina et al., 2019) and COVID-19 (Goswami et al., 2022) brought fake news to the mainstream. However, fake news is not a new subject and has been around for millennia. For example, in the Roman Empire, the Procopius of Caesarea used fake news to defame his political rival Emperor Justinian (Browning, 1962).

The advances in communication tools such as the internet, social media and online platforms have changed our public domain dramatically. Unlike in the past, today fake news is spread with a few taps and clicks faster and wider than credible information. A growing body of empirical and theoretical research on fake news has addressed different aspects, such as the theoretical consequence of fake news (Gupta et al., 2013), financial motivations (Allcott & Gentzkow, 2017), and political and ideological motivations (Marwick & Lewis, 2017). The threats of fake news are not limited to these domains; stock markets (Ferrara et al., 2016) and health (Fernández-Luque & Bau, 2015) are also among the social circumstances affected by fake news. Besides these negative outcomes of fake news, perhaps the most influential and dangerous effect of fake news is on people’s perception and interpretation of reality (Lewandowsky et al., 2012; Silverman, 2016, p. 16).

The majority of the research has focused on the topics such as fake news fabrication and the dissemination process (Allcott & Gentzkow, 2017), the difficulty of distinguishing between credible news and fake news (Tandoc et al., 2018), and the vulnerability of individuals to fake news (Shao et al., 2018). Therefore, the important question about the psychological aspects of fake news is partly unanswered.

To address this gap, this article aims to investigate the different psychological factors affecting fake news. Factors such as analytical thinking in the dual-thinking process, overclaiming and social factors, including social ties and social approval, are tested together for the first time in this study in for groups.

Social Ties and Social Approval

Before the Internet, people were merely seen as an audience whose contribution to the communication process was limited to passively absorbing whatever the media and corporation had to offer (Mudgal, 2019). But the social network has democratised the communication paradigm, and users now can be participants instead of mere observers (Constantinides, 2009). The most fundamental effect of the social network is transforming the static nature of the internet into an interactive dynamic sphere (Brennan & Schafer, 2010). This change enabled users to generate their content (user-generated content [UGC]) and post mentions on the posts, and, based on new features of the social network, they could share these contents (Lai & Turban, 2008), resulting in ever-changing content controlled by users (Kaplan & Haenlein, 2010) instead of credible sources. These possibilities have changed online behaviour from passive to active (Law et al., 2003; Stewart & Pavlou, 2002). Participation, openness, conversation, community and connectedness are all new traits of modern online people (Mayfield, 2008).

Leonardi et al. (2013) use the ‘social lubricant’ metaphor for this feature. Social networks lubricate how we communicate by easing connection and communication. Now users are able to maintain social ties that were impossible before. Also, the depth, density, frequency and quality of communication are easily manageable via social networks. The direct effect of social ties on information dissemination was seldom studied in the past (Lai & Wong, 2002). The majority of the previous research on social ties was situational in which purposive action needed to be taken.

The importance of interpersonal and social ties in the dissemination of information is not a new domain and is evident in the communication research literature. The intention to share information on social networks is one of the main motivations for joining and forming online communities (Ridings & Gefen, 2004), but what are the factors affecting this innate urge? Trust, (Andrews et al., 2002; Ridings et al., 2002), group ties (Shen & Gong, 2019) and personal ties (Wang et al., 2015) are among those factors predicting information-sharing on social media.

One of the new features of modern social media is liking (Hong et al., 2017), which is the process of expressing feelings towards a post by pressing the like button (heart, thumb-up, star, etc.). This liking behaviour is an exclusive feature of social media that influences every aspect of social media communication. Studies show that liking is an interaction between users, exhibiting some extent of approval and impression (Skågeby, 2009, 2010). Liking shows some types and extent of awareness about the topic as it is viewed by others (Segev et al., 2013). This awareness might be a behaviour of controlling information dissemination to avoid social disapproval to protect one’s self-presentation (Hong et al., 2017). Self-presentation is closely tied with sharing, posting and disclosing information (Lee-Won et al., 2014). The findings of previous studies highlight the role of social approval by asserting that one might get social approval and acquire self-presentation by sharing socially approved positive information (Lee-Won et al., 2014). We can propose the following hypotheses:

H₁: On social media, high social approval is positively connected to the intention of sharing content with one’s peers and acquaintances.

H₂: On social media, low social approval is negatively connected to the intention of sharing content with one’s peers and acquaintances.

Fake News and Dual-Thinking Process

Contrary to the common belief, ‘fake news’, is not a new phenomenon, and it has been studied from different perspectives such as motivations (Allcott & Gentzkow, 2017; Subramanian, 2017) and outcomes of consumptions (Ferrara et al., 2016; Gu et al., 2017; Silverman, 2016). Many have defined fake news with different approaches. Some have focused on the untruthfulness of the information. Leeder (2019) argues that fake news is a story that is fabricated with the intention to deceive viewers into thinking it is real news.

Finneman and Thomas (2017) define fake news as the intentional deception of a mass audience by non-media actors via a sensational communication that appears credible but is designed to manipulate and is not revealed to be false. This definition omits certain types of fake news such as click-bait, which is used by media actors to gain audience and monetise it (Fallis, 2015; Wardle, 2017; Wardle & Derakhshan, 2017). Allcot and Gentzkow define fake news in a differentiated way from credible news: ‘news articles that are intentionally and verifiably false and could mislead readers’ (Allcot & Gentzkow, 2017, p. 213). This simple definition points out to substantial elements of fake news: first, fake news is a form of news fabrication and second, fake news has deception with the intention to mislead. Since the definition suggested by Allcot and Gentzkow (2017) is succinct and encompassing, we follow this definition and, in our experiment, we formulate fake news that are intentionally and verifiably false and could mislead the readers.

Advances in communication technologies and the availability of the internet proliferate fake news to an unprecedented measure. Also, change in information-consumption patterns is another problematic matter (Goyanes, 2014; Shu et al., 2017). The increasing popularity of digital information platforms as the information source is an outcome of the drastic increase of popularity of the internet in the 1990s, which eventually led to the decline of hard news (Ahlers, 2006; Meyer, 2004). Recently, other types of media lost ground to social media as an information source. For example, a recent study reveals that in 2017, more than half (about 54%) of people online across 36 countries use social media as a source of news and 14% use it as their main source of information (Newman et al., 2017).

Nowadays, people glance at the news and rarely read anything beyond the headlines (Rochlin, 2017). Previous studies show that almost 50% of the news shared on the Twitter platform is not even read by the users (Gabielkov et al., 2016, p. 182). In other words, reading the headline is sufficient for deciding whether to believe or share the information or not. This means that the problem of fake news is becoming more complex and requires more investigation from different perspectives that were neglected before.

Considering these findings, alongside other features of fake news, understanding the cognitive factors of believing and sharing fake news becomes more important. It is widely known that repetition is an important factor in believing fake news (Dechêne et al., 2010; Fazio et al., 2015; Hasher et al., 1977), but recent findings show that reading the headline even one time might increase the perception of credibility and truth (Pennycook et al., 2018).

Pennycook and Rand (2018) contend that type of belief in fake news is influenced by low-level cognitive processing. Tseng and Fogg (1999) categorise this low level of thinking as gullibility errors, where users are not interested in processing the information thoroughly. Therefore, cognition and the level of thinking becomes crucial to believing in fake news. Hence, this factor might explain why particular individuals believe in fake news and others do not.

Dual-process theory asserts that human cognition is divided into two processes, intuitive process and analytic process (De Neys, 2012; Evans & Stanovich, 2013; Kahneman, 2011; Pennycook et al., 2015). In this study, this theory is used to understand individual differences in the thinking process, whether they use intuitions or analytic thinking and analysing the outcome of these thinking proves of believing and sharing fake news.

The cognitive reflection test (CRT) (Frederick, 2005) is one of the most useful measures in studies of differences in thinking (Baron et al., 2015). This test has been used in different areas and, for a three-item test, is highly reliable with reliability (Cronbach’s alpha) being around 0.6 for most samples (Baron et al., 2015). CRT is a small set of mathematical questions for which at first glance the answer is obvious. The result is an accumulative composite measure of three items. This obvious answer is an intuitive or lure answer, and providing the correct answer requires respondents to overcome the temptation to give the initial intuitive answer. Since CRT correlates with many other measures (Toplak & Stanovich, 2002), it is used in many studies in different strings of research such as moral judgments and values (Baron et al., 2015; Paxton et al., 2012; Royzman et al., 2015), attitudes towards science (Kahan et al., 2017; Shtulman & McCallum, 2014), altruism (Arechar et al., 2017) and smartphone use (Barr et al., 2015), to name a few.

With this simple measure, researchers might evaluate the willingness to think analytically (Pennycook et al., 2015), which might be a factor in rejecting fake news. Humans usually try to avoid analytical thinking (Fiske & Taylor, 2013; Stanovich & West, 2000). which in today’s over-saturated information ecosystem, might lead to falling for fake news. The problem of fake news becomes crucial when people or media users do not recognise some particular news as fabricated news and respond to that news as truth. Since this research aims to investigate the correlation between analytical thinking and acceptance of fake news, this measure is used to evaluate the individuals’ cognitive process in fake news accuracy perception. This intuition response also might indicate another trait, overclaiming, which is closely tied with the willingness to elicit fast non-analytical responses (Pennycook et al., 2017). The following hypotheses might be put forward:

H₃: The ability to think analytically on social media content correlates negatively with susceptibility to fake news.

H₄: The ability to think analytically on social media content is negatively related to trust in real news.

Overclaiming

Overclaiming is a measure developed by Paulhus et al. (2013) that refers to the general tendency of individuals to ‘self-enhance’ while being asked about their familiarity with different topics. For example, a respondent might assert familiarity with a person or historical incident that does not exist at all. Thus, we assume that overclaiming might be correlated with believing in fake news. Initially, his concept originated by Phillips and Clancy (1972) who developed an index for consumer behaviour studies in which respondents rated their familiarity with some consumer items that were actually non-existent. In another study, Stanovich and Cunningham (1992) used 50% non-existent authors to control their responses, and, surprisingly, many respondents claimed familiarity with authors that do not exist at all. Gradually, overclaiming became a variable in studies. For example, in a study on unethical behaviours, Randall and Fernandes (1991) used overclaiming as a control bias. Paulhus et al. (2013) argue that the findings of this research suggest that evaluating overclaiming in a study might be a suitable index for bias in self-report studies. Therefore, considering the correlation of overclaiming with a lack of analytical thinking and indexing bias, this measure is used in this study. Based on the previous literature, this article tries to shed some light on the relation between overclaiming and believing in fake news. Therefore, the following hypothesis might be put forward:

H₅: Overclaiming correlates negatively with the ability of analytical thinking.

Experiment Design

Many studies have examined how psychological factors affect susceptibility to fake news. But we aim to take other factors into account to enrich the previous findings. This article tries to investigate the relationships between CRT, acceptance of fake/real news, social approval and sharing the news. The theoretical framework is presented in Figure 1.

To randomise each respondent to a group and treatment, a two-by-two factorial design was used to permit the simultaneous test of several hypotheses. In this type of design, the interaction between treatments is generally an advantage rather than a limitation since this design enables researchers to measure an effect and relation which otherwise was not possible.

To this end, and to investigate the relationship between analytical thinking, overclaiming and social approval in sharing fake news, an experiment with four groups were designed. The Instagram platform was chosen due to its popularity in Iran and due to the fact that it is not filtered and all users have access to this social network. In general, it is hypothesised that there is a negative correlation between CRT and fake news acceptance. Also, overclaiming manifests itself as a positive correlation between believing in fake news and gaining a high rank in overclaiming test scores. We hypothesise that lower CRT and higher overclaiming lead to sharing fake news with others more. Additionally, since the researchers did not intend for participants to perceive the variables, no manipulation check was used.

In the first group (hereafter referred to as FNLSA), the potential correlation between fake news, low social approval, CRT and overclaiming were investigated. Also, the possibility of sharing information was assessed as well. The second group, (hereafter referred to as FNHSA), is a replication of group 1, but in order to assess the effects of social approval, high social approval is manifested in the number of likes and mentions of the designed post.

To evaluate the effect of social approval on believing in real news, an Instagram post with factful information was designed with fewer likes and mentions to be used in the third group (hereafter referred to as RNLSA). It is hypothesised that those with higher CRT tend to share real news with others. Following the third group, in the fourth group, the real news was used in the group with some modifications to show a high level of social approval (hereafter referred to as RNHSA). To illustrate the experimental design, Figure 2 is presented:

Considering the time of the research and the important topics in society, real and fake information about online banking was chosen. During the research time, the Iranian banking system was rolling out the static pin codes for online banking or mobile banking applications, and users had to obtain dynamic pin codes through an application or SMS; therefore, the topic was a national headline affecting almost all people’s daily lives. Besides these factors, respondents were asked if they share this information with their respective peers or not. Information about sharing behaviour of respondents provides a vehicle for comparing sharing online information (fake or real) to other factors such as overclaiming and CRT.

To summarise, in the first group, fake news with low social approval was designed to evaluate the responses in sharing and to study the relation between CRT and overclaiming. In the second group, we investigate the relation between fake news and high social approval regarding the aforementioned variables. The third and fourth groups are a replication of the first and second groups with real news. The Instagram posts and headline used in the questionnaire are presented in the Appendix.

Participant and Questionnaire

Due to the behaviour habits of social media users and the ease of distribution, online questionnaires were used to gather data. Four Instagram posts with a headline, both fake news and real news, were created using Photoshop software. The news was presented in a common format of the Instagram post—a headline over a photo with the information presented in the caption of the post. Considering the fact that people judge the credibility of news by the headline (Gabielkov et al., 2016), only the headline was intended to be judged.

A target of 250 respondents was set for each group, which was exceeded. Four questionnaires were formulated and each respondent could complete only one questionnaire just once. The questionnaires’ links were distributed through social networks and online groups, and complete virality was aimed for. The first wave of distribution was among the University of Tehran students. They were asked to post the link in other groups and social media. Hence, the researchers had no control over the age or distribution of the respondents. Online users shared the link through their circles and to their peers creating a sample of social network users. The total number of participants is 1,160. The demography of respondents is presented in Table 1.

Table 1.

Descriptive Analysis.

		Group 1 (N = 270)	Group 2 (N = 280)	Group 3 (N = 360)	Group 4 (N = 250)	Total (N = 1,160)
Male %		44	57	50	56	52.0
Female %		56	43	50	44	48.0
Age	>20	11	0	0	0	2.6
	20–24	41	39	58	60	50.0
	25–29	33	46	19.5	32	31.9
	30–35	11	11	19.5	8	12.9
	36<	4	4	3	0	2.6
Education	BSc	74	46	64	72	63.8
	MSc	22	54	25	24	31.0
	PhD	4	0	11	4	5.2
Online time (minutes)	~30	0	0	3	0	0.9
	30–60	18.5	7	11	4	10.3
	~90	41	25	28	32	31.0
	~120	22	29	22	24	24.1
	120<	18.5	40	36	40	33.6

Source: The authors.

Overclaiming was measured by the scale introduced by Paulhus et al. (2013). The cultural modification was carried out on the topics and contemporary and historical incidents, buildings, and incidents such as the president of Iran, were chosen to adapt the questionnaire to Iranian context, contemporary topics and historical incidents were replaced. A Likert 5 scale was used to assess the overclaiming. Respondents were asked if they know, for example, President Rouhani, and answers were ranging from ‘1 = not familiar at all’ to ‘5 = completely familiar with’. Five out of 15 questions were a non-existence item to measure the overclaiming in respondents. These questions are culture-specific. For example, respondents were asked if they know a non-existence poet named Kamalodin Najafi or a non-existence historical building named Moala Dome.

Originally, Paulhus et al. (2003) evaluated the overclaiming score by subtracting an indication of familiarity with non-existence items from real ones. To ease the numerical computation, any indication of familiarity scored 1 and non-familiarity scored 0. In this experiment, the sum of the scores of non-existence answers and some of the genuine answers were subtracted. For a better presentation of the results, the overclaiming score was subtracted from 10. For example, if a respondent indicates familiarity with all items (high overclaiming), their overclaiming score is 5,¹ and if indicates no familiarity with non-existence items (low overclaiming), the score is 0.²

As stated previously, the CRT was measured by a test originally developed by Frederick (2005). In this experiment, like Toplak et al. (2011), a composite measure of three items was used. Every correct answer scored 1 and an incorrect answer scored 0. Hence, a participant score might be ranging from 0–3. Finally, to assess low social approval, the number of likes and mentions on the post were designed as low. Also, respondents were asked if they share this piece of information with their peers.

Findings

The correlation among variables in group 1 (FNLSA) is shown in Table 2. As expected, the correlation between CRT results and overclaiming is negative, which means that the higher the CRT score, the lower the overclaiming in the respondent. Overclaiming is presented in the format of answering the correct answers and non-existence answers. Not surprisingly, overclaiming in non-existence items correlates negatively with the CRT score. Since in calculating the overall OC score, the final score was subtracted from 10, the correlation must be interpreted inversely, which correlate positively. CRT correlates negatively with the acceptance of fake news and sharing. This means analytical thinking reveals the fake news and users tend to share less.

As assumed, there is a significant negative correlation between overclaiming and analytical thinking. To have more insight, exploratory factor analysis was done. Scree plot analysis and retention of factors with an eigenvalue greater than 1 indicated there is a single factor (eigenvalue = 1.643, 42% of variance explained). All the variables loaded on this factor in expected directions: overclaiming = 0.885, CRT = −0.853, fake news accuracy = 0.294 and social ties = 0.213. These findings suggest that tendency to believe and sharing fake news both share a low CRT score with higher overclaiming.

Table 2.

Descriptive and Correlation of Group 1.

	Scale	Mean	SD	Skew^a	Kurtosis^b	1	2	3	4	5	6
1. OC score	0–1	2.190	1.495	1.376	3.158	1
2. CRT score	0–1	1.629	1.043	0.448	−.152	−.596**	1
2. CRT score	0–1	1.629	1.043	0.448	−.152	.001
3. Fake news accuracy	1–5	1.910	0.652	0.846	0.154	.190	–.001	1
3. Fake news accuracy	1–5	1.910	0.652	0.846	0.154	.343	.997
4. Overclaiming right	0–1	0.848	0.152	−1.742	5.193	−.697**	.309	−.075	1
4. Overclaiming right	0–1	0.848	0.152	−1.742	5.193	.000	.117	.711
5. Overclaiming wrong	0–1	0.133	0.235	2.397	6.529	.365	−.355	.144	.413*	1
5. Overclaiming wrong	0–1	0.133	0.235	2.397	6.529	.062	.069	.475	.032
6. Social ties	1–5	2.098	0.852	0568	0.195	.025	−.101	.096	.031	.072	1
6. Social ties	1–5	2.098	0.852	0568	0.195	.900	.615	.635	.878	.720

Source: The authors.

Notes: N = 270.

^aSD = 0.488.

^bSD = 0.872.

**Correlation is significant at the 0.01 level (2-tailed)

*Correlation is significant at the 0.05 level (2-tailed).

The second group, actually a replication of group 1 with high social approval, extends the findings of a previous group considering the effects of high social approval. Also, we might investigate the correlation between the social ties and fake news accuracy based on the social approval effect.

To compare the effects of low social approval in group 1 and high social approval in group 2, the ANOVA test was used, and since the significance is less than 5% (p-value = .005), the difference between sharing fake news and real news was not significant. In the Scheffe post hoc test, the mean difference is −0.121 and p-value is .970. Also, the Tukey test’s result was the same. Additionally, social approval on acceptance of fake news might be measured. Findings show that low social approval and high social approval yield no significant difference in believing fake news (p-value = .082). A descriptive analysis is presented is available in Table 3.

Table 3.

Descriptive and Correlation of Group 2.

	Scale	Mean	SD	Skew^a	Kurtosis^b	1	2	3	4	5	6
1. OC score	0–1	1.928	1.184	.580	.206	1
2. CRT score	0–1	1.428	0.997	−0.029	−0.993	−.067	1
2. CRT score	0–1	1.428	0.997	−0.029	−0.993	.734
3. Fake news accuracy	1–5	2.114	0.614	−0.487	−0.737	.073	−.131	1
3. Fake news accuracy	1–5	2.114	0.614	−0.487	−0.737	.713	.506
4. Overclaiming right	0–1	0.892	0.097	−0.359	−1.003	−.324	.033	.174	1
4. Overclaiming right	0–1	0.892	0.097	−0.359	−1.003	.092	.869	.375
5. Overclaiming wrong	0–1	0.164	0.255	2.078	4.233	.676**	−.083	.244	.463*	1
5. Overclaiming wrong	0–1	0.164	0.255	2.078	4.233	.000	.675	.211	.013
6. Social ties	1–5	2.220	1.033	0.471	−0.761	.311	−.173	.371	.114	.409*	1
6. Social ties	1–5	2.220	1.033	0.471	−0.761	.107	.379	.052	.564	.031

Source: The authors.

Notes: N = 280.

^aSD = 0.441.

^bSD = 0.858.

**Correlation is significant at the 0.01 level (2-tailed).

*Correlation is significant at the 0.05 level (2-tailed).

In this group, correlations suggest that there is a positive correlation between overclaiming and sharing on social media, which reveals that those who claim more knowledge, share more information with their peers. Exploratory factor analysis shows that one factor explains 40% of the variances (eigenvalue = 1.604). Other variables load as follows: overclaiming = 0.550, CRT = −0.439, fake news accuracy = 0.664 and social ties = 0.817.

Since real news with low social approval was used in the third group, the effect of low social approval on accepting real news and fake news might be investigated as well. ANOVA was used to test the correlation. Since the significance is low (Sig. 0.082) Post hoc analysis (Tukey, Scheffe and LSD) was done as well confirming the insignificance correlation. To have an overview of the variables, descriptive analysis of this group is presented in Table 4.

Table 4.

Descriptive and Correlation of Group 3.

	Scale	Mean	SD	Skew^a	Kurtosis^b	1	2	3	4	5	6
1. OC score	0–1	2.305	2.039	.992	.726	1
2. CRTsScore	0–1	1.611	1.128	−.166	−1.340	−.108	1
2. CRTsScore	0–1	1.611	1.128	−.166	−1.340	.530
3. Real news accuracy	1–5	2.400	.902	.047	−1.134	−.031	−.129	1
3. Real news accuracy	1–5	2.400	.902	.047	−1.134	.857	.453
4. Overclaiming right	0–1	.863	.164	−1.431	2.492	−.777**	−.047	.081	1
4. Overclaiming right	0–1	.863	.164	−1.431	2.492	.000	.785	.638
5. Overclaiming wrong	0–1	.188	.257	1.730	2.676	.595**	−.232	.054	.044	1
5. Overclaiming wrong	0–1	.188	.257	1.730	2.676	.000	.174	.754	.797
6. Social ties	1–5	2.115	1.012	.479	−1.010	−.004	−.135	.687**	.052	.060	1
6. Social ties	1–5	2.115	1.012	.479	−1.010	.983	.434	.000	.765	.728

Source: The authors.

Notes: N = 360.

^aSD = 0.393.

^bSD = 0.768.

**Correlation is significant at the 0.01 level (2-tailed).

*Correlation is significant at the 0.05 level (2-tailed).

Not surprisingly, the overclaiming score correlates positively with sharing the information with acquaintances. Results of exploratory factor analysis revealed that in this group, a single factor explains 44% of variance with an eigenvalue of 1.735. Other factors’ loading is as follows: overclaiming = 0.005, CRT = 0.635, real news accuracy = 0.902 and social ties = 0.903.

Following the third group, the real news was used in the fourth group (RNHSA). Hence, the effect of high social approval on believing real news and information-sharing compared to real news with low social approval and believing and sharing fake news with high social might be examined comparatively. Descriptive and correlation are illustrated in Table 5.

Table 5.

Descriptive and Correlation of Group 4.

	Scale	Mean	SD	Skew^a	Kurtosis^b	1	2	3	4	5	6
1. OC score	0–1	2.560	1.635	.040	−1.076	1
2. CRTsScore	0–1	1.400	1.258	.109	−1.681	−.316	1
2. CRTsScore	0–1	1.400	1.258	.109	−1.681	.124
3. Real news accuracy	1–5	2.152	.666	.110	−.594	−.196	.093	1
3. Real news accuracy	1–5	2.152	.666	.110	−.594	.348	.657
4. Overclaiming right	0–1	.896	.130	−1.505	2.185	−.262	−.041	.467*	1
4. Overclaiming right	0–1	.896	.130	−1.505	2.185	.205	.847	.019
5. Overclaiming wrong	0–1	.304	.361	.842	−.853	.716**	−.315	.160	.486*	1
5. Overclaiming wrong	0–1	.304	.361	.842	−.853	.000	.125	.445	.014
6. Social ties	1–5	2.040	1.064	.902	−.147	.262	−.225	.369	.221	.397*	1
6. Social ties	1–5	2.040	1.064	.902	−.147	.206	.279	.070	.288	.049

Source: The authors.

Notes: N = 250.

^aSD = 0.464.

^bSD = 0.902.

**Correlation is significant at the 0.01 level (2-tailed).

*Correlation is significant at the 0.05 level (2-tailed).

ANOVA post hoc analysis showed no significance of the effect of low and high social approval in the interpretation of real information. Post hoc analysis of intent to share showed different results. The mean difference between sharing real news with low social approval and with high social approval is significant at 0.05 level (0.581).

Finally, the last group revealed some insight on true information combined with a high degree of social approval. Again, social ties and sharing information on social media correlates positively with overclaiming. Exploratory factor analysis indicates that a single factor explains 40% of variances (eigenvalue = 1.539). Variables load onto factor as follows: CRT = 0.274, overclaiming = −0.389, real news accuracy = 0.907 and social ties = 0.716.

CRT

To analyse the CRT score for all groups of respondents, one-way ANOVA test was used and the results showed no significance among groups (F(3,112) = 0.291, p-value = .832). These findings might suggest that analytical thinking and the urge to elicit the first intuitive response is not related to the traits of the respondents in different groups. Multiple regression was used to predict CRT from other variables. These variables statically did not predict CRT (F(3,109) = 2.524, p-value = .061, R² = 0.065). This results also reinforce the previous finding that CRT is rather a trait. The results are presented in Table 6.

Table 6.

Multiple Regression Results for CRT.

	B	SE	Beta	t	Sig.
(Constant)	2.102	.327		6.420	.000
Normalised Overclaiming Score	−.146	.059	−.232	−2.486	.014
Normalised Fake news	−.019	.130	−.014	−.144	.886
Normalised Social Ties	−.089	.102	−.085	−.868	.387

Source: The authors.

Notes: Dependent variable: NormalCRTScore.

Overclaiming

Investigating overclaiming as a variable in four groups was in line with the previously mentioned findings that overclaiming is a personal trait and is not significantly related to the group’s variables rather individuals’ traits (F(3,111) = 0.669, p-value = .573). Again, multiple regression was not predicted by other variables (F(3,109) = 2.539, p-value = .060, R² = 0.256). This provides a support for the notion that overclaiming is a personal trait. This findings are presented in the Table 7.

Table 7.

Multiple Regression Results for Overclaiming.

	B	SE	Beta	t	Sig.
(Constant)	2.896	.545		5.318	.000
Normalised fake news	−.157	.207	−.074	−.758	.450
Normalised social ties	.138	.162	.083	.848	.398
Normalised CRT score	−.368	.148	−.232	−2.486	.014

Source: The authors.

Note: Dependent variable: NormalOCScore.

Social Ties

To investigate the probability of sharing information between all four groups, one-way ANOVA test was used. In each group, information accuracy and social approval were independent variables, and social ties (the degree each respondent share the information with their peers) was the dependent variable. Results reveal that the probability of sharing in each group of respondents is significantly different (F(3,111) = 6.536, p-value = .000). The post depicting real news headline with low social approval (1.418 ± 0.887) is shared significantly less than real news with high social approval (2.016 ± 0.824, p-value = .041), fake news with low social approval (2.168 ± 0.762, p-value = .004) and fake news with high social approval (2.250, 0.883, p-value = .001).

Unlike CRT and overclaiming, social ties can be predicted with other factors. These variables statistically and significantly predicted social ties (F(3,109) = 4.644, p-value = .004, R² = 0.113). Table 8 shows the results.

Table 8.

Multiple Regression Results for Social Ties.

	B	SE	Beta	t	Sig.
(Constant)	1.067	.345		3.094	.003
Normalised overclaiming score	.048	.056	.079	.848	.398
Normalised CRT score	−.078	.089	−.081	−.868	.387
Normalised fake news	.401	.116	.313	3.465	.001

Source: The authors.

Note: Dependent variable: NormalSocialTie.

Information Accuracy Perception

Respondents answered questions about whether they find the presented information credible or not. Analysing the information accuracy perception between four groups revealed that there are no statistically significant differences among the four groups of respondents (F(3,111) = 1.640, p-value = .184). Multiple regression reveals that this variable significantly predicts information accuracy perception. Results are as follows: F(3,109) = 4.141, p-value = .008 and R² = 0.102.

Table 9.

Multiple Regression Results for Information Accuracy Perception.

	B	SE	Beta	t	Sig.
(Constant)	1.773	.226		7.853	.000
Normalised overclaiming score	−.033	.044	−.071	−.758	.450
Normalised CRT score	−.010	.070	−.014	−.144	.886
Normalised social ties	.247	.071	.317	3.465	.001

Source: The authors.

Note: Dependent variable: Information Accuracy Perception.

Discussions and Conclusion

This article contributes to the existing literature on fake news and tries to extend what we know about the psychological aspect of fake news. The experiment provided different insights into the questions, and some consistent results were found across the groups. It was found that the tendency to share information (both fake news and credible news) is associated and correlates positively with the overclaiming score of respondents across three groups (FNHSA, RNLSA and RNHSA). Also, as predicted, one group proved that overclaiming is negatively correlated with CRT in detecting fake news. This finding supports that analytical thinking might affect and reduce believing in fake news.

Additionally, about sharing the news on social media, findings show that social approval plays an important positive role. Users share real news with high social approval significantly more than that real news with low social approval. Also, compared to fake news with high and low social approval, real news with low social approval is shared less. This explains that in real news, the number of likes and shares is an important factor that encourages users to share information with their peers, but in sharing fake news, the fake news itself captures the attention of users, enticing them to share the provocative news. Also, as collateral finding that is not directly related to the purpose of the study but noteworthy to mention, age proved to be another variable affecting sharing false information. Young respondents share significantly more fake news with others.

Although all variables across four groups were measured with different methods, exploratory factor analysis showed that a single factor explains a reasonable amount of variance (about 45%). This evidence provides the support existence of a common factor consistent with the lack of analytical thinking over overclaiming in knowledge, which helps to explain the intention to share information on social media or believing the fake news.

This article provides an understanding of why some people believe fake news in social media and share that fake news with others. In this experiment, four different (but similar) groups were designed, and CRT and overclaiming were analysed as variables across different social approval and news. Users who share more information are also those who overclaim knowledge on topics that do not exist at all, and they are sensitive to social approval methods of social media in formats of number of likes and shares. All variables and measures loaded heavily on one single factor, which we assume is a tendency to overclaim on topics and news and consequently sharing them to claim knowledge and expertise on different topics. This overclaim is heavily associated with a lack of analytical thinking leading to believe fake news.

Implication

This article, along with many other papers (e.g., Pennycook et al., 2018; Rochlin, 2017), denotes that in our age truth is becoming less important and provides evidence for this claim that might help marketing practitioners. Based on the findings, it might be suggested that truth and honesty, unfortunately, do not suffice for marketing communication. Marketing practitioners must bear in mind that in today’s media and modern communications, social approval is more important than the truthfulness of the information. Thus, in order to spread their information on social media, they should not solely rely on the information and its truthfulness. Influence marketing, social media viral campaigns and digital marketing are new tools in modern marketing that should not be neglected. Also, findings on the effects of the education on sharing fake news point out that marketer practitioners must formulate their information, marketing communication and campaigns based on the education of the targeted segments. It can be suggested that a truthful information combined with high social approval through influence marketing and viral campaigns might work well for segments with lower education.

Finally, the proven relation between sharing more and overclaiming more reveals a marketing opportunity. Gamification of communication that provides a vehicle for users to overclaim their knowledge to their peers on social media might be a suitable strategy on social media to spread the message.

Limitations and Future Researches

In this article, a set of fake and real news in an Instagram format (a headline with a picture and a caption) were presented to the participants, and they were asked to answer several questions about their judgment and indicate their perceived credibility. We acknowledge the limits to this method. First, respondents were only given a headline, which is based on the assumption that the majority of the social media users only read the headline (Gabielkov et al., 2016). Second, they were not able to click on the caption and expand the presented information to gain more insight on the topic; hence, their judgment is solely based on the headline and visible part of the caption. These factors might limit the generalisability of the findings.

The second major limitation is the nature of the data-gathering tool. In this case, self-explanatory online questionnaires were used, which might affect the real response. Respondents simply might react to headlines and news differently than what they normally do. Another factor that should be taken into account is the general information literacy of social media users, which might lead to the inability to detect real and fake news.

Considering the limitations of this study, future researchers might use the full information or article to investigate the news more comprehensively. Also, the difference in judging headlines and full news might be analysed in future researches. Additionally, future researches might use other methods such as observation or in-depth interviews to investigate users’ judgments in formats other than self-explanatory. Also, further researches are needed to take into account the personal traits and the individual and contextual differences.

Supplemental Material

Supplemental material for this article is available online.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Notes

Four Instagram Posts with a Headline Used in the Questionnaire

References

Ahlers

(2006). News consumption and the new electronic media. Harvard International Journal of Press/Politics, 11(1), 29–52.

Allcott

, & Gentzkow

(2017). Social media and fake news in the 2016 election. Journal of Economic Perspectives, 31(2), 211–236.

Andrews

, Preece

, & Turoff

(2002). A conceptual framework for demographic groups resistant to on-line community interaction. International Journal of Electronic Commerce, 6(3), 9–24.

Arechar

A. A.

, Kraft-Todd

G. T.

, & Rand

D. G.

(2017). Turking overtime: How participant characteristics and behavior vary over time and day on Amazon Mechanical Turk. Journal of the Economic Science Association, 3(1), 1–11.

Baron

, Scott

, Fincher

, & Metz

S. E.

(2015). Why does the Cognitive Reflection Test (sometimes) predict utilitarian moral judgment (and other things)? Journal of Applied Research in Memory and Cognition, 4(3), 265–284.

Barr

, Pennycook

, Stolz

J. A.

, & Fugelsang

J. A.

(2015). The brain in your pocket: Evidence that Smartphones are used to supplant thinking. Computers in Human Behavior, 48, 473–480.

Brennan

, & Schafer

(2010). Branded!: How retailers engage consumers with social media and mobility. John Wiley & Sons.

Browning

(1962). [Review of the book Procopius: Secret history, translated by R. Atwater]. The Classical Review, 12(3), 309–310.

Constantinides

(2009). Social media/Web 2.0 as marketing parameter: An introduction. In Proceedings of 8th International Congress Marketing Trends (pp. 15–17). https://research.utwente.nl/en/publications/social-media-web-20-as-marketing-parameter-an-introduction

10.

Dechêne

, Stahl

, Hansen

, & Wänke

(2010). The truth about the truth: A meta-analytic review of the truth effect. Personality and Social Psychology Review, 14(2), 238–257.

11.

De Neys

(2012). Bias and conflict: A case for logical intuitions. Perspectives on Psychological Science, 7(1), 28–38.

12.

Evans

J. S. B. T.

, & Stanovich

K. E.

(2013). Dual-process theories of higher cognition: Advancing the debate. Perspectives on Psychological Science, 8(3), 223–241.

13.

Fallis

(2015). What is disinformation? Library Trends, 63(3), 401–426.

14.

Fallis

, & Mathiesen

(2019). Fake news is counterfeit news. Inquiry, 1–20.

15.

Fazio

L. K.

, Brashier

N. M.

, Payne

B. K.

, & Marsh

E. J.

(2015). Knowledge does not protect against illusory truth. Journal of Experimental Psychology: General, 144(5), 993.

16.

Ferrara

, Varol

, Davis

, Menczer

, & Flammini

(2016). The rise of social bots. Communications of the ACM, 59(7), 96–104.

17.

Fernández-Luque

, & Bau

(2015). Health and social media: Perfect storm of information. Healthcare Informatics Research, 21(2), 67.

18.

Finneman

, & Thomas

R. J.

(2018). A family of falsehoods: Deception, media hoaxes and fake news. Newspaper Research Journal, 39(3), 350–361.

19.

Fiske

S. T.

, & Taylor

S. E.

(2013). Social cognition: From brains to culture. SAGE Publications.

20.

Frederick

(2005). Cognitive reflection and decision making. Journal of Economic Perspectives, 19(4), 25–42.

21.

Gabielkov

, Ramachandran

, Chaintreau

, & Legout

(2016, June). Social clicks: What and who gets read on Twitter? In Proceedings of the 2016 ACM sigmetrics international conference on measurement and modeling of computer science (pp. 179–192). Association for Computing Machinery.

22.

Goyanes

(2014). An empirical study of factors that influence the willingness to pay for online news. Journalism Practice, 8(6), 742–757.

23.

Goswami

M. P.

, Anjali

, Raj

, Puthiyakath

H. H.

, & Thanvi

(2022). Media during pandemic crisis in India: An analysis of people’s perception during the first wave of COVID-19. Journal of Creative Communications. 10.1177/09732586221090750

24.

, Kropotov

, & Yarochkin

(2017). The fake news machine: How propagandists abuse the internet and manipulate the public. Trend Micro, 5, 1–85.

25.

Gupta

, Lamba

, Kumaraguru

, & Joshi

(2013, May). Faking Sandy: Characterizing and identifying fake images on Twitter during Hurricane Sandy. In Proceedings of the 22nd International Conference on World Wide Web (pp. 729–736). Association for Computing Machinery.

26.

Hasher

, Goldstein

, & Toppino

(1977). Frequency and the conference of referential validity. Journal of Verbal Learning and Verbal Behavior, 16(1), 107–112.

27.

Hong

, Chen

Z. F.

, & Li

(2017). ‘Liking’ and being ‘liked’: How are personality traits and demographics associated with giving and receiving ‘likes’ on Facebook? Computers in Human Behavior, 68, 292–299.

28.

Kahan

D. M.

, Peters

, Dawson

E. C.

, & Slovic

(2017). Motivated numeracy and enlightened self-government. Behavioural Public Policy, 1(1), 54–86.

29.

Kahneman

(2011). Thinking, fast and slow. Macmillan.

30.

Kaplan

A. M.

, & Haenlein

(2010). Users of the world, unite! The challenges and opportunities of social media. Business Horizons, 53(1), 59–68.

31.

Lai

L. S.

, & Turban

(2008). Groups formation and operations in the Web 2.0 environment and social networks. Group Decision and Negotiation, 17(5), 387–402.

32.

Lai

, & Wong

(2002). The tie effect on information dissemination: The spread of a commercial rumor in Hong Kong. Social Networks, 24(1), 49–75.

33.

Law

, Lau

, & Wong

Y. H.

(2003). From customer relationship management to customer-managed relationship: Unraveling the paradox with a co-creative perspective. Marketing Intelligence & Planning, 21(1), 51–60.

34.

Leeder

(2019). How college students evaluate and share ‘fake news’ stories. Library & Information Science Research, 41(3). 10.1016/j.lisr.2019.100967

35.

Lee-Won

R. J.

, Shim

, Joo

Y. K.

, & Park

S. G.

(2014). Who puts the best ‘face’ forward on Facebook? Positive self-presentation in online social networking and the role of self-consciousness, actual-to-total Friends ratio, and culture. Computers in Human Behavior, 39, 413–423.

36.

Leonardi

P. M.

, Huysman

, & Steinfield

2013, Enterprise social media: Definition, history, and prospects for the study of social technologies in organizations. Journal of Computer-Mediated Communication, 19(1), 1–19.

37.

Lewandowsky

, Ecker

U. K.

, Seifert

C. M.

, Schwarz

, & Cook

(2012). Misinformation and its correction: Continued influence and successful debiasing. Psychological Science in the Public Interest, 13(3), 106–131.

38.

Marwick

, & Lewis

(2017). Media manipulation and disinformation online. Data & Society Research Institute.

39.

Mayfield

(2008). What is social media. National Defense University Press.

40.

Meyer

(2004). Saving journalism. Columbia Journalism Review, 43(4), 55.

41.

Molina

M. D.

, Sundar

S. S.

, Le

, & Lee

(2019). ‘Fake news’ is not simply false information: A concept explication and taxonomy of online content. American Behavioral Scientist, 65(10). 10.1177/0002764219878224

42.

Mudgal

(2019). A vital report on trends in media freedom and its link with the future of democracy. Journal of Creative Communications, 14(2), 160–170. 10.1177/0973258619852004

43.

Newman

, Fletcher

, Kalogeropoulos

, Levy

D A. L.

, & Nielsen

R. K.

(2017). Reuters Institute digital news report. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3026082

44.

Paxton

J. M.

, Ungar

, & Greene

J. D.

(2012). Reflection and reasoning in moral judgment. Cognitive Science, 36(1), 163–177.

45.

Paulhus

D. L.

, Harms

P. D.

, Bruce

M. N.

, & Lysy

D. C.

(2003). The over-claiming technique: measuring self-enhancement independent of ability. Journal of Personality and Social Psychology, 84(4), 890.

46.

Paulhus

D. L.

, Westlake

B. G.

, Calvez

S. S.

, & Harms

P. D.

(2013). Self-presentation style in job interviews: The role of personality and culture. Journal of Applied Social Psychology, 43(10), 2042–2059.

47.

Pennycook

, Fugelsang

J. A.

, & Koehler

D. J.

(2015). What makes us think? A three-stage dual-process model of analytic engagement. Cognitive Psychology, 80, 34–72.

48.

Pennycook

, & Rand

D. G.

(2020). Who falls for fake news? The roles of bullshit receptivity, overclaiming, familiarity, and analytic thinking. Journal of Personality, 88(2), 185–200.

49.

Pennycook

, Ross

R. M.

, Koehler

D. J.

, & Fugelsang

J. A.

(2017). Dunning–Kruger effects in reasoning: Theoretical implications of the failure to recognize incompetence. Psychonomic Bulletin & Review, 24(6), 1774–1784.

50.

Pennycook

, Cannon

T. D.

, & Rand

D. G.

(2018). Prior exposure increases perceived accuracy of fake news. Journal of Experimental Psychology: General, 147(12), 1865.

51.

Phillips

D. L.

, & Clancy

K. J.

(1972). Some effects of ‘social desirability’ in survey studies. American Journal of Sociology, 77(5), 921–940.

52.

Randall

D. M.

, & Fernandes

M. F.

(1991). The social desirability response bias in ethics research. Journal of Business Ethics, 10(11), 805–817.

53.

Rochlin

(2017). Fake news: Belief in post-truth. Library Hi Tech, 35(3).

54.

Ridings

C. M.

, & Gefen

(2004). Virtual community attraction: Why people hang out online. Journal of Computer-mediated Communication, 10(1), JCMC10110.

55.

Ridings

C. M.

, Gefen

, & Arinze

(2002). Some antecedents and effects of trust in virtual communities. The Journal of Strategic Information Systems, 11(3–4), 271–295.

56.

Royzman

E. B.

, Landy

J. F.

, & Leeman

R. F.

(2015). Are thoughtful people more utilitarian? CRT as a unique predictor of moral minimalism in the dilemmatic context. Cognitive Science, 39(2), 325–352.

57.

Segev

, Shoham

, & Ruvio

(2013). Gift-giving among adolescents: Exploring motives, the effects of givers’ personal characteristics and the use of impression management tactics. Journal of Consumer Marketing, 30(9).

58.

Shao

, Ciampaglia

G. L.

, Varol

, Yang

K. C.

, Flammini

, & Menczer

(2018). The spread of low-credibility content by social bots. Nature Communications, 9(1), 1–9.

59.

Shen

, & Gong

(2019). Personal ties, group ties and latent ties: Connecting network size to diversity and trust in the mobile social network WeChat. Asian Journal of Communication, 29(1), 18–34.

60.

Shu

, Sliva

, Wang

, Tang

, & Liu

(2017). Fake news detection on social media: A data mining perspective. ACM SIGKDD Explorations Newsletter, 19(1), 22–36.

61.

Shtulman

, & McCallum

(2014). Cognitive reflection predicts science understanding. In Proceedings of the annual meeting of the cognitive science society (Vol. 36, No. 36). https://escholarship.org/uc/item/4t79p8pj#author

62.

Silverman

(2016). This analysis shows how viral fake election news stories outperformed real news on Facebook. BuzzFeed News.

63.

Skågeby

(2009). Exploring qualitative sharing practices of social metadata: Expanding the attention economy. The Information Society, 25(1), 60–72.

64.

Skågeby

(2010). Gift-giving as a conceptual framework: Framing social behavior in online networks. Journal of Information Technology, 25(2), 170–177.

65.

Stanovich

K. E.

, & Cunningham

A. E.

(1992). Studying the consequences of literacy within a literate society: The cognitive correlates of print exposure. Memory & Cognition, 20(1), 51–68.

66.

Stanovich

K. E.

, & West

R. F.

(2000). Individual differences in reasoning: Implications for the rationality debate? Behavioral and Brain Sciences, 23(5), 645–665.

67.

Stewart

D. W.

, & Pavlou

P. A.

(2002). From consumer response to active consumer: Measuring the effectiveness of interactive media. Journal of the Academy of Marketing Science, 30(4), 376–396.

68.

Subramanian

(2017). Inside the Macedonian fake-news complex. Wired Magazine, 15.

69.

Tandoc Jr

E. C.

, Lim

Z. W.

, & Ling

(2018). Defining ‘fake news;: A typology of scholarly definitions. Digital Journalism, 6(2), 137–153.

70.

Toplak

M. E.

, & Stanovich

K. E.

(2002). The domain specificity and generality of disjunctive reasoning: Searching for a generalizable critical thinking skill. Journal of Educational Psychology, 94(1), 197.

71.

Tseng

, & Fogg

B. J.

(1999). Credibility and computing technology. Communications of the ACM, 42(5), 39–44.

72.

Wang

, Li

, & Tang

(2015, April). Dwelling and fleeting encounters: Exploring why people use WeChat-A mobile instant messenger. In Proceedings of the 33rd annual ACM Conference extended abstracts on human factors in computing systems (pp. 1543–1548). Association for Computing Machinery.

73.

Wardle

(2017). Fake news. It’s complicated. First Draft, 16, 1–11. https://firstdraftnews.com:443/fake-news-complicated/

74.

Wardle

, & Derakhshan

(2017). Information disorder: Toward an interdisciplinary framework for research and policy making. Council of Europe Report, 27, 1–107.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB