1 Introduction

Scientific misinformation poses an existential threat to humanity, as it disrupts our collective ability to respond to major challenges such as pandemics, antibiotic resistance, climate change and food scarcity. In the interest of clarity, it is essential to highlight that although the terms “misinformation” and “disinformation” are sometimes used interchangeably, their meaning is profoundly different. While both words refer to information that is untrue or factually incorrect, they diverge with regards to the intent or scope with which they are produced. According to the American Psychological Association, misinformation can be concisely defined as “false or inaccurate information—getting the facts wrong”, whereas disinformation is “false information which is deliberately intended to mislead—intentionally misstating the facts” [1]. Therefore, a misinformed person is someone who holds factually incorrect information to be true, regardless of the intent with which the false information was originally created.

The intentional production and dissemination of scientific misinformation is not a recent phenomenon: human history is rife with fabricated information being presented as scientific facts. These fabrications range from relatively harmless hoaxes (such as the Piltdown man, Redheffer’s perpetual motion machine and the Roswell alien autopsy) to disinformation campaigns with catastrophic long-lasting impact, such as downplaying the harmful effects of tobacco smoke, denying the impact of fossil fuels on the climate, or suggesting links between vaccines and autism [2, 3].

The popularisation and accessibility of social media and digital technologies has democratised the dissemination of information, enabling virtually anyone with internet access to easily create and share written and audio-visual content. While this has undoubtedly facilitated knowledge exchange and scientific progress, it has also paved the way for the rapid spread of misinformation, including scientific fake news encompassing a wide range of factually incorrect or purposefully misleading information presented as genuine scientific findings, often designed to deceive or manipulate readers for financial gain, political agendas, or ideological motives [4]. Over the last few years, the widespread popularisation of generative artificial intelligence (AI) and large language models (LLM) has further facilitated the production and diffusion of media misinformation. Recent studies revealed that “individuals are largely incapable of distinguishing between AI- and human-generated text” and that “the ability of LLMs to rapidly produce vast amounts of text could leverage misinformation spread at an unprecedented scale, this could create an ‘AI-driven infodemic,’ a novel public health threat” [5, 6].

Tackling misinformation in what has been defined as a “post-truth era” is a seemingly unsurmountable yet crucial endeavour, requiring deep understanding not only of the factors underpinning its creation and diffusion, but also of those influencing people’s susceptibility to it [4, 7]. Enabling citizens to identify scientific misinformation requires a combination of short-term mitigation strategies (e.g. campaigns to raise public awareness or policies to make social media platforms accountable for fact-checking content shared on them) and longer-term educational initiatives to foster scientific and digital literacy in future citizens [8].

Secondary school years (typically corresponding to ages 11 to 18) are a fundamental period in the development of abstract reasoning, decision-making and critical thinking skills [9, 10]. Crucially, secondary school is also the highest educational stage in which science teaching is still compulsory in most educational systems, making it an ideal setting to provide all students (including those who choose not to pursue further scientific studies) with the cognitive and conceptual skills required to identify and avoid scientific misinformation [11].

Worryingly, school curricula often prioritise the transmission of factual knowledge over the development of critical thinking skills, leaving students ill-equipped to discern between credible and unsubstantiated scientific claims [12, 13]. As a result, students may uncritically accept false information (particularly if it contains scientific terminology or concepts that they are familiar with) that aligns with their preconceived beliefs or comes from a source they might erroneously consider as reliable. Several intrinsic and extrinsic psychological factors may underpin secondary school students' susceptibility to scientific fake news. For example, cognitive biases such as confirmation bias (interpreting new information to fit pre-existing beliefs) and the illusion of explanatory depth (overestimating one’s understanding of a topic) predispose individuals to seek out and endorse information that confirms their existing beliefs while overestimating their understanding of complex scientific concepts [14]. Moreover, adolescents’ vulnerability to (and sharing of) fake news might be mediated by fear of missing out (FoMO) and desire for belonging and popularity [15, 16].

The ubiquity of social media platforms and their increasing popularity among younger audiences are key contributors to the spread of scientific misinformation. Adolescents spend a substantial amount of time engaging with social media platforms such as Instagram, TikTok, X, Facebook and SnapChat, where misinformation can convincingly be presented as legitimate scientific content [17, 18]. Moreover, the algorithmic content curation on these platforms often reinforces users’ existing beliefs and preferences, creating echo chambers where false information is amplified and dissenting viewpoints are marginalised [19, 20]. It is worth noting that adolescents’ information-seeking behaviour is influenced by their relatively limited vocabulary, which might lead them to prefer sources where information is presented in a simple and accessible manner (e.g. a short video or post on social media) over more complex ones such as academic journal articles or textbooks [21]. Information literacy skills gaps are further compounded by mounting pressure on teachers to “teach to examinations” by prioritising the memorisation of subject content over the development of critical and analytical skills [22].

The primary aim of the study is to quantitatively investigate secondary school students’ susceptibility to scientific misinformation by measuring their belief in commonly held scientific conspiracy theories based on factually incorrect information. The secondary aim is to gain insight into students’ awareness and perception of the reliability of different sources of scientific information.

2 Methods

2.1 Ethical considerations and approval

This study was conducted in accordance with the University of Portsmouth research ethics policy and with the Declaration of Helsinki for research involving human participants. The research protocol, including the survey instrument and procedure for obtaining informed consent, was reviewed and approved prior to the start of the study (code BIOL-ETHICS #030-2023) to ensure that it adhered to ethical guidelines and protected the rights and welfare of the participants.

The online questionnaire used in the study was prefaced by a disclaimer informing participants of the aims and scope of the research, its voluntary and anonymous nature, and their right to leave any question unanswered or withdraw at any point prior to submitting their answers. Participants were informed that by clicking the “Submit” button they were providing informed consent to the use of their anonymous answers towards this study.

All data collected during the survey were stored securely in a password-protected drive, in compliance with GDPR and University of Portsmouth data protection policy.

2.2 Survey design and distribution

An anonymous, self-administered online questionnaire (Table 1) was created using Google Forms. The questionnaire was divided in three sections. Section one (“About you”) contained questions regarding participants’ demographic information, as well as their preference for science subjects and intentions to study them in the future. Section two (“True or false”) contained ten Likert-type questions where students were asked to score the veracity of each statement on a scale from 1 (false) to 5 (true). Five of the statements were factually incorrect (e.g. “The Earth is flat”) and based on common conspiracy theories, while the other five were factually correct statements relating to science topics appropriate to the students’ educational level (e.g. “The Earth has different layers, including the crust, mantle, and core”). The five factually incorrect statements were adapted from a previous YouGov poll [23] that investigated belief in fake news among British adults, and slightly modified to be accessible to an adolescent population sample. The five factually correct statements were based on the Science National Curriculum and drafted in collaboration with school teaching staff to ensure that they reflected concepts that students have already encountered in their studies. The order of the ten questions was scrambled to prevent students from guessing answers based on the position of the questions. The survey was administered between January and March 2024 in The Cowplain School, a coeducational secondary school located in Hampshire, UK. A convenience sampling strategy was chosen to ensure a high response rate while minimising disruption to participating students. The survey link was posted by teachers onto their Google Classroom stream, and students were allowed to choose whether to complete the survey during school hours or at home.

Table 1 Questionnaire used in the survey
Full size table

2.3 Scoring and statistical analysis

To assess students’ vulnerability to misinformation, two separate scores were developed: the Unintentional Misinformation Susceptibility Score (UMSS, being misinformed by holding a true statement as false) and the Disinformation Susceptibility Score (DSS, being misinformed by believing in a deliberately false piece of information). Each score was calculated by adding the responses (on a scale of 1 to 5) to five questions (indicated with asterisks in Table 1), meaning that both scores range from 5 (least misinformed) to 25 (most misinformed). It is important to point out that for the calculation of the UMSS the allocation of points was inverted to account for the fact that agreement with a true statement indicates lower misinformation. Therefore, points allocation was inverted to facilitate understanding, meaning that a higher UMSS indicates higher susceptibility to misinformation. The internal consistency of the questions was evaluated using Cronbach’s Alpha tests. The tests revealed an Alpha value of 0.717 for the five factually correct statements and 0.660 for the five factually incorrect statements, indicating good reliability of the questions used to calculate the misinformation scores. Cochran’s formula was used to calculate the minimum sample size required to achieve the desired confidence level. With an error margin of 5%, minimum sample sizes were estimated at 164 participants for a confidence level of 80%, and 664 participants for a confidence level of 99%. All statistical tests were performed using IBM SPSS Statistics version 28. Since the data collected in the survey were categorical in nature (therefore failing the assumptions for parametric testing), non-parametric tests were used for the statistical analysis. Kruskal–Wallis tests were used to compare medians between different groups of participants. Where statistically significant differences were observed, post-hoc analysis was carried out using Dunn’s pairwise tests. Kendall's tau-b (τb) tests were used to measure the strength and direction of correlations between study variables. A significance threshold of p = 0.05 was used for all tests.

3 Results

A total of 776 students completed the survey across all secondary school year groups (school years 7 to 11), exceeding the 664 participants required to achieve a confidence level of 99% with a margin of error of 5% in accordance to Cochran’s formula. The participants’ breakdown by year group, gender, religion, and ethnicity is shown in Table 2.

Table 2 Demographic features of the study participants
Full size table

The questionnaire responses indicated that the vast majority of students were able to correctly identify scientific fake news as not true. When asked to score statements between 1 (false) and 5 (true), only 3.6% of students responded “true” (score of 4 or 5) to “The Earth is flat”, 5.7% to “Global warming is a myth”, 10.8% to “Humans never landed on the moon”, 4.9% to “Vaccines are not safe”, and 13.4% to “the Earth is 6000 years old, so it’s impossible that humans descended from apes”. To capture the whole cohort’s susceptibility to disinformation, students’ responses to the fake statements were pooled into a single score (DSS). In order to better differentiate students’ misinformation between belief in deliberate disinformation and simple ignorance of the scientific facts, students’ responses to level-appropriate true statements were used to calculate an Unintentional Misinformation Susceptibility Score (UMSS), which was used as a measure of students’ understanding of key scientific concepts. A Kendall’s tau-b test revealed a statistically significant moderate positive correlation (τb = 0.317, p = 4.29 × 10−29) between the Unintentional Misinformation Susceptibility Score and the Disinformation Susceptibility Score (Fig. 1).

Fig. 1
figure 1

Correlation between Disinformation Susceptibility Score and Unintentional Misinformation Susceptibility Score. The black solid line represents the linear regression fit, the red dashed line the 95% confidence interval, and the grey dashed line the 95% prediction interval

Full size image

As shown in Fig. 2, a decline in disinformation susceptibility was observed between Key Stage 3 (KS3, Years 7–9) and Key Stage 4 (KS4, Years 10–11). Kruskal–Wallis test revealed that median DSS changed significantly across year groups (χ2 = 15.288; df = 4; p = 0.004). Year 10 students had the lowest DSS, which was significantly lower than Year 7 (p = 0.002), Year 8 (p = 0.024), Year 9 (p = 0.0001) and Year 11 (p = 0.044).

Fig. 2
figure 2

Students’ Disinformation Susceptibility Score (on a scale between 5 and 25) by year group. Red lines represent medians, circles represent outliers, stars represent extreme outliers

Full size image

Figure 3 shows a statistically significant (χ2 = 25.252; df = 4; p = 0.00004) decrease in Students’ Unintentional Misinformation Susceptibility Scores across the year groups, with median UMSS gradually declining from 7 (Year 7) to 6 (Years 8 and 9) to 5 (Years 10 and 11). Post-hoc tests highlighted that Year 11 students had significantly lower UMSS than Year 7 (p = 0.00013), Year 8 (0.00013) and Year 9 (p = 0.017) students. Similarly, Year 10 students had significantly lower UMSS than Year 7 (p = 0.000091), Year 8 (p = 0.0032) and Year 9 (p = 0.017) students.

Fig. 3
figure 3

Students’ Unintentional Misinformation Susceptibility Score (on a scale between 5 and 25) by year group. Red lines represent medians, circles represent outliers, stars represent extreme outliers

Full size image

Across the year groups, students showed a significant (χ2 = 67.346; df = 4; p = 8.23 × 10−14) decline in intentions to pursue scientific studies (Fig. 4). The decrease was statistically significant between Year 7 and Year 8 (p = 0.045), Year 7 and Year 9 (p = 2.03 × 10−9), Year 7 and Year 10 (p = 0.00011), Year 7 and Year 11 (p = 8.21 × 10−13), Year 8 and Year 9 (p = 0.002), Year 8 and Year 11 (p = 0.000008), Year 10 and Year 11 (p = 0.023).

Fig. 4
figure 4

Students’ intention to pursue further scientific studies (on a scale between 1 and 5) by year group. Red lines represent medians, circles represent outliers, stars represent extreme outliers

Full size image

No statistically significant association was found between students’ intentions to pursue further scientific studies and DSS (Fig. 5A) or UMSS (Fig. 5B). Likewise, no statistically significant association was found between students’ enjoyment of scientific studies and DSS (Fig. 5C) or UMSS (Fig. 5D).

Fig. 5
figure 5

Students’ Disinformation Susceptibility Score (A and C) and Unintentional Misinformation Susceptibility Score (B and D) as a function of their intention to pursue further scientific studies (A and B) and enjoyment of scientific studies (C and D). Red lines represent median UMSS, circles represent outliers, stars represent extreme outliers. NS none of the observed differences are statistically significant

Full size image

When asked to rank their source of scientific news from a list of options (Fig. 6), students’ first choice was school teachers (26%), followed by TikTok (13%) and news websites like BBC or Sky News (11%). The least popular sources were newspapers (2.5%), Facebook (1.9%) and magazines (1.5%). Students were then asked to provide examples of scientific news sources they considered trustworthy and untrustworthy. Analysis of their open answers revealed that BBC (including BBC News and BBC Bitesize) was the most frequent (22%) example of a trustworthy source, followed by teachers (11%) and books (3%). The most frequent examples of sources deemed untrustworthy were TikTok (21%), Wikipedia (5%) and YouTube (4%).

Fig. 6
figure 6

Students’ first choice of scientific news source by percentage of respondents

Full size image

4 Discussion

The present study revealed low levels of scientific misinformation among secondary school students, and highlighted that susceptibility to both disinformation and unintentional misinformation decreases over the course of secondary education. While it is widely acknowledged that adolescents are vulnerable to misinformation and that group dynamics (e.g. peer pressure, FoMO) might mediate their susceptibility, our findings support previous reports that within adolescent peer groups “truth-seeking is favored, even when contending with ingroup loyalty” [24] and that secondary students “were able to deal with fake news, identifying it as such” [25]. It is worth pointing out that that some respondents might have answered negatively to questions related to their belief in fake news due to social desirability bias, which would lead to an underestimation of the prevalence of misinformation. However, this risk was mitigated by the fact that the survey was online and anonymous, a setting that minimises social biases typically associated with non-anonymous surveys [26].

Remarkably, the secondary school students who responded to the present survey showed substantially lower belief in common science-based conspiracy theories than adults living in the same country. However, as no adults were surveyed in the present study, caution should be taken when comparing results obtained in different (albeit analogous) surveys at different times. According to a recent YouGov poll [23], between 9–32% of British adults believe in fake news on climate change (5.7% of the secondary students in our sample), 18–24% on vaccine risks (4.9% in our sample), and 13–21% of adults believe that the moon landing was staged (10.8% in our sample). On the other hand, the present study and the YouGov survey provided comparable results with regards to the Earth being flat (3% of British adults, 3.6% in our sample) and on evolution being false or a part of God’s plan (14% of British adults, 13.4% in our sample). While the observation of adolescents being less misinformed than adults might sound counterintuitive, this finding corroborates the trends reported in the YouGov poll that belief in scientific fake news (particularly those regarding climate change and vaccines) increases with age. A tentative explanation is that digital natives are more familiar with the online environment than previous generations and may therefore be less susceptible to trusting unreliable sources, however there is conflicting evidence as to whether that is the case. For example, while previous studies [27] reported a direct correlation between age and engagement with/sharing of fake news, a recent report [28] appears to indicate that belief in conspiracy theories was higher in teenagers than in adults. It should however be noted that these studies were not only focused on scientific fake news, but also on political/social conspiracy theories, e.g. “Deep State” and “Great Replacement”, so it is conceivable that different demographic groups might be more susceptible to different types of misinformation. For example, within our study population students were much better at identifying fake news on climate change and vaccines than they were with respect to the moon landing or human evolution. A plausible interpretation of this observation is that young people are more likely to seek information and consult multiple sources on topics (such as vaccine safety and climate change) that affect their present or future, and to take a less critical approach towards matters that they might perceive as not directly affecting their life, such as the moon landing or the theory of evolution [29, 30].

The two scores developed in this study allowed to quantify misinformation within the study population and—to an extent—discriminate based on whether it is caused by belief in deliberate disinformation (DSS) or poor knowledge of the scientific subject (UMSS). The observation of a positive correlation between DSS and UMSS should not come as a surprise, considering that individuals with stronger scientific and numerical literacy are likely to be less susceptible to trusting unreliable sources and believing in fake news [31, 32]. However, our results show that the correlation between DSS and UMSS is not a strong one, indicating that the scores successfully captured subtle differences in the way students become misinformed. Individual students might not necessarily believe in fake news because of mere ignorance of scientific facts, but rather due to the contribution of intrinsic and extrinsic psychological factors (as discussed in the introduction of this paper).

Both the DSS and UMSS decreased significantly across the year groups, indicating that students become progressively less misinformed across educational Key Stage 3 and Key Stage 4. On one hand, this should not come as a surprise considering that students become increasingly educated as they progress through their studies and are taught new concepts. On the other hand, this observation is in apparent conflict with previous reports that susceptibility to fake news increases with age within the adult population. Taken together, these findings indicate that the relationship between age and misinformation is not linear and might change direction between adolescence and early adulthood, reinforcing previous observations that the transition between compulsory and voluntary studies is a critical period in the development of scientific literacy [33].

Worryingly, our study shows that students’ intentions to pursue further scientific studies declined considerably between Year 7 and Year 11, which are respectively the first and last year of secondary education in England. Students undertake their General Certificate of Secondary Education (GCSE) exams at the end of Year 11, which for most students represents the completion of compulsory education. Therefore, it is concerning that students’ intentions to pursue scientific studies decline significantly in the period crucial towards their choice of voluntary further education. This observation aligns with previous reports that students’ propensity towards careers in science and maths declines over the course of secondary education, a phenomenon that appears mediated by a concomitant decline in self-confidence in both subjects [34, 35].

Students’ susceptibility to misinformation (as measured by their DSS and UMSS) was not significantly associated with their enjoyment of scientific studies nor with their intention to further pursue them after completing their GCSEs. This was an interesting and unexpected observation, as previous research indicates that students’ attitude towards and engagement with scientific studies might be a good predictor for their critical evaluation skills [36, 37].

Students were asked to rank their most frequently used sources of scientific news from a list. School teachers were indicated as the most popular source, receiving twice as many preferences as the second most popular option, TikTok. Interestingly, despite TikTok being their second most used source of scientific information, students named the same platform as the most common example of an untrustworthy source of scientific information. The uncontrolled use of TikTok and similar social media among children and adolescents has raised concerns over its detrimental impact on their mental health, cognitive development, critical thinking, and attention span [38, 39]. Reassuringly, the secondary students who took part in the present survey displayed evidence of critical judgement towards their use of TikTok and awareness of the high prevalence of fake news and misinformation that circulates on it, in a similar fashion to what was reported among university students in a recent study [40].

The findings of the present study can be framed within the growing body of literature investigating scientific misinformation in secondary schools across multiple countries. A recent survey carried out among secondary students in Spain revealed that while participants expressed substantial concerns regarding climate change, they harboured high levels of misinformation and misconceptions on the topic [41]. Likewise, a study involving 483 secondary students in Sweden highlighted that “youths often find it difficult to determine credibility when faced with different types of digital news and misinformation” [42]. Interestingly, a survey of high school students in Western Australia revealed that “although students often recognised legitimate news sources versus opinion, they often fail to recognise bias when it relates to political or organisation affiliation” [43].

While adding considerable insight to our understanding of adolescents’ susceptibility to misinformation and their perception of scientific sources’ reliability, this study is not without limitations. As the survey was carried out in a single secondary school, it reflects the opinions and mindsets of participants coming from a specific geographic area and socio-economic background. Moreover, the convenience sampling strategy utilised in this study may result in selection bias and low external validity, so these findings might not be fully transferable to populations with different demographic and cultural compositions. Finally, the survey was limited to students enrolled in school years 7 to 11 (educational Key Stages 3 and 4), and it therefore provides no information on misinformation susceptibility in primary (Key Stages 1 and 2) and post-secondary (Key Stage 5) school students.

5 Conclusion

This study provides valuable insight into the extent of scientific misinformation amongst secondary school students, and on their use and perception of science news sources. The 776 students who took part in the survey displayed on average considerably lower belief in scientific fake news compared to the levels reported in a recent poll of the adult British population. Susceptibility to both disinformation and unintentional misinformation declined significantly across the 5 years of secondary education, reinforcing the essential role of schools in fostering scientific literacy among the population.

Over the course of their secondary studies, students were progressively less likely to express an interest to pursue further scientific studies after completing compulsory education. This observation raises significant concerns, particularly considering that science graduates play an increasingly crucial role in addressing global challenges such as public health emergencies, food scarcity, climate change and energy sustainability. Therefore, allocating adequate funding and staffing to schools and universities should be prioritised by the government and considered not as an expense, but rather a long-term investment to ensure that the British scientific sector (and society at large) remains innovative, competitive and sustainable.

When asked to rank their sources of scientific news, the majority of the students indicated school teachers as their first choice. In an era of widespread online misinformation, this observation should be seen as a testament to the crucial role of educators and to the trust students put in them. This finding also reinforces the necessity to provide school teachers with training opportunities and continuing professional development to ensure they remain up to date with respect to their subject area. The survey revealed that social media are also a popular source of scientific information among secondary students. However, students showed critical awareness of the prevalence of online misinformation, and indicated TikTok as the most common example of an untrustworthy source for science news. On the other hand, BBC was named by students as the most frequent example of a trustworthy source.

Despite its limitations, this study provides novel understanding of adolescents’ vulnerability to misinformation over the course of their secondary studies and their approach to acquiring scientific information. The questionnaire and scoring system developed in this study might constitute a simple and effective tool that can be implemented or adapted by educators and researchers looking to gauge the susceptibility of a student group to scientific misinformation. Building upon the findings of this study, further investigations would be advisable in order to assess susceptibility to scientific misinformation in different demographic groups, or to evaluate the impact of initiatives designed to improve scientific and digital literacy.