Student argumentation on health issues dealt with in online sources

In today’s information society, the internet is the most used information source for finding answers to various important and current questions (e.g. Eurostat, 2022). Specifically, health- and nutrition-related information is largely available on the internet and commonly searched for (Chen et al., 2018; Eurostat, 2022). However, a vast amount of this information available both on the internet in general (Peng et al., 2023; Swire-Thompson and Lazer, 2020) and on social media in particular (Suarez-Lledo and Alvarez-Galvez, 2021) is unreliable, even including deliberately produced disinformation (Broniatowski et al., 2018; George, 2024).

Several previous studies have shown that the health literacy skills of adolescents and young adults are inadequate. In their survey (n = 14,916), Dadaczynski et al. (2021) found that a large fraction of German university students (42.3%) had great difficulty in assessing the reliability of health-related information and in determining whether the information was based on a commercial interest (38.9%). In the self-report study (n = 256) by Patil et al. (2021), only 49% of US college students had adequate health literacy skills. Kiili et al. (2018) showed that only about half (51.4%) of Finnish early adolescents questioned the credibility of a commercial health-related online source and only about 19% recognized commercial bias.

Poor health literacy skills and distrust of online health information, combined with its heavy use (Freeman et al., 2023; Wartella et al., 2016), may lead to illogical and unwanted health decisions. The ability to differentiate between reliable and unreliable online information and to analyze the argumentation of online sources is important for young people to be able to find and use relevant and trustworthy information from the internet. This study aims to examine upper secondary school students’ argumentation on health topics dealt with in online sources in a mother tongue and literature course named “Texts and influence”.

Toulmin’s (1958) argument analysis framework differentiates the core elements that comprise an argumentative text. In this framework, a claim refers to the argument’s conclusion, data to the facts cited as premises or support for the claim and a warrant to a general operating principle allowing the formation of a bridge between the data and the claim (Fulkerson, 1996, p. 18; Nussbaum, 2011). Some researchers also use the terms grounds (e.g. Osborne et al., 2004), evidence (Forzani, 2016) and reason (e.g. Means and Voss, 1996) to describe elements of arguments that support a claim or a conclusion. The reasons given in support of a claim or conclusion may vary in their standpoints, for example, finance, ethics or health. These different standpoints have been termed argumentation perspectives (see Xu and Yao, 2015).

The argumentative quality of texts can be quantitatively evaluated through the depth and breadth of the argumentation used (see van Amelsvoort et al., 2007). The depth refers to the number of reasons used to support a claim, while the breadth refers to the number of argumentation perspectives for the claim. In a good argumentative text, the claim is supported by several relevant reasons indicating the depth of argumentation (c.f. Britt et al., 2014 and Fulkerson, 1996 for relevance and sufficiency of argumentation). Further, the use of reasons representing many argumentation perspectives indicates that the writer has researched the topic extensively.

In an online inquiry, students are supposed to deal with and investigate internet sources to learn about an issue through the framework of new literacies of online research and comprehension (Leu et al., 2013). This learning process includes the phases of Information search, Evaluating information, Synthesizing information and Communicating results to others. To engage successfully in an online inquiry, students need both argumentation and sourcing skills, which are intertwined and embedded in all these phases (Barzilai et al., 2015; Bråten et al., 2018; Forzani et al., 2022; Wineburg, 1991). Sourcing refers here to the use of information related to the sources throughout the inquiry process (Kiili et al., 2021; McCrudden et al., 2018). From an information behavior perspective, students’ inquiry processes are shaped by their perceived information needs as well as intervening personal and environmental variables (e.g. Wilson, 1999). Factors such as prior knowledge, the prevalence of misinformation and the complexity of the online environment influence how students search for, evaluate, synthesize and communicate information.

In the searching phase, one should identify various argumentation perspectives for the topic in question to be able to formulate relevant search terms, which can then be used to reach a broad understanding of the topic and to avoid biased and one-sided information (Marttunen et al., 2021). By sourcing, students can, first, consider which sources provide the most reliable information; second, frame the search terms with references to reliable organizations or research-based information; and third, inspect source features, for example, the descriptions for the pages that appear below the titles (e.g. Hämäläinen et al., 2023; Kiili et al., 2021). However, from the perspective of everyday information behavior (e.g. Case and Given, 2016), students, however, often gravitate toward sources that are convenient and “good enough” rather than optimal, which may limit the breadth of perspectives they encounter.

When evaluating online information, one is expected to identify relevant and sufficient arguments (Fulkerson, 1996), detect information bias (e.g. economic or ideological) and determine whether argumentation is balanced (Kiili et al., 2018; Walton, 1991). By sourcing, students evaluate, for example, the expertise and intentions of the author when interpreting the quality of the content of the source (Hämäläinen et al., 2023).

Synthesizing information aims to form a coherent and integrated representation of multiple sources and perspectives. One is thus expected to identify various arguments in the source texts, differentiate between relevant and irrelevant arguments and compare and contrast their strengths and weaknesses across the sources, all essential elements of argumentative reading (Kiili and Leu, 2019; Newell et al., 2011). By sourcing, students make connections between sources and the content (source-to-content links) and connections between different sources (source-to-source links), for example, to contrast sources (Hämäläinen et al., 2023; Kiili et al., 2021). From an information practices perspective, sourcing is not merely a cognitive operation but a socially situated activity shaped by the norms and expectations of the learning environment, including which sources students are expected to connect with and interact with (McKenzie, 2003).

Finally, when communicating findings to other people convincingly, one should put forward clear and well-justified conclusions that take into account qualifiers, rebuttals and source information included in multiple sources (Anmarkrud et al., 2013; Barzilai et al., 2015).

Previous studies have described the ways argumentation and sourcing are intertwined and iterative practices in a successful process of an online inquiry. Kiili et al. (2021) found that the more frequently students engaged in sourcing when specifying their information needs and making relevant search queries, the more frequently they engaged in sourcing when judging the credibility of online texts and used sources in their argumentative writing. Forzani et al. (2022) found, in turn, that effective evaluators were more critical of the veracity of information and more capable of evaluating and justifying the credibility of online sources. In the same vein, Barzilai et al. (2015) found that higher levels of sourcing were related to more complex argumentation and better integration of sources in written argumentation.

Students’ argumentation and sourcing skills have proved insufficient when seeking relevant and reliable information on the internet. Hämäläinen et al. (2021) found that when upper secondary school students selected and evaluated health-related online texts, they frequently used author, venue and evidence but quite rarely author’s intention and corroboration as evaluation criteria for the credibility of the texts. Corroboration, in particular, is a cognitively demanding aspect of critical online reading, as it requires skills to analyze and compare the argumentation of different sources. Students have also been found to have difficulties in citing, using source features and comparing sources when composing a written product (e.g. Kiili et al., 2020; Pérez et al., 2018). Marttunen et al. (2021) found that students struggled with recognizing unbalanced argumentation and justifying the credibility of online sources on vaccination.

Positive results in overcoming the challenges related to argumentation and sourcing have been achieved through well-planned sourcing interventions. For example, in the study by Hämäläinen et al. (2023), upper secondary school students with low sourcing skills benefited from an intervention aimed at fostering their sourcing during a collaborative online inquiry task. Further, Bråten et al. (2019) found long-term effects of an argumentative sourcing intervention implemented in an upper secondary school. However, Anmarkrud et al. (2021) found that many contextual and individual difference factors are associated with successful sourcing and critical reading.

Learners’ prior topic knowledge has been found to enhance critical and argumentative evaluation of online sources (Forzani, 2016; Hämäläinen et al., 2021) and multiple text comprehension (e.g. Bråten et al., 2020; Stang Lund et al., 2019). However, this association seems to be connected to the method of measuring prior knowledge, as significant differences were found between results gained through multiple-choice or open-ended questions in contrast to true/false questions (see a review by Anmarkrud et al., 2021).

To summarize, much is known about how students evaluate the credibility of sources and the arguments presented in them, and how they use this information related to the sources during the online inquiry process. However, less is known about how adolescents form positions based on online information and to what extent do they justify their opinions. Such knowledge is important for educating young people and strengthening their health literacy, enabling them to make rational decisions about their own lives based on a critical evaluation of digital information.

In this study, upper secondary school students engaged in a mother tongue and literature course in an online inquiry process during which they formed a position on the health topics of Vaccination or Dietary Fats based on online information and justified their position. The study aimed to explore the following research questions:

The participants were 345 upper secondary school students (93.2% aged 17) from 8 schools in Finland. There were 58.6% female students and 41.4% male students. Most (73.4%) of the participants were students in five large urban upper secondary schools, and the rest (26.6%) were students in three small and mid-sized schools in semi-urban and rural areas. Consent was obtained from all students, and if a student was underage, permission was also sought from their guardian(s). The statement of the University Ethics Board was not required because the study included no risk to the participants.

This study was part of a larger research project investigating upper secondary school students’ learning of argumentative online inquiry skills (Kiili et al., 2022). The study focused on the phase of the project in which students completed an argumentative online inquiry task on the health-related topics of Vaccination and Dietary Fats during mother tongue and literature lessons.

Students’ prior topic knowledge for both Vaccination and Dietary Fats was assessed with 10 statements, 3 of which were correct and 7 false. The selection of the statements was based on current scientific knowledge and common misinformation related to both topics. To strengthen the content validity, two university lecturers in health sciences and a medical expert reviewed the statements. The students were asked to select the three statements they thought were correct, and their knowledge scores varied from 0 to 3 depending on the number of correct statements selected.

In this study, students were assigned an online inquiry task; for 170 students, the topic of the task was Vaccination, and for 175 students, the topic was Dietary Fats. In the task, the students were asked to search for information on Vaccination or Dietary Fats using a Google custom search engine, evaluate the trustworthiness of the information and write a recommendation letter on a health dilemma by presenting a position on the topic and justifying it by relying on the online sources they had read.

Health topics, Vaccination and Dietary Fats, were selected as there is a lot of online information available on health issues that people commonly use and adopt (Freeman et al., 2023). In addition, information available on Vaccination and Dietary Fats is controversial by nature as there are advocates and opponents for both the uptake of vaccination and the reduction of hard dietary fats. Finally, one side of both topics aligns with the current scientific view, that is, it is recommended to follow vaccination guidelines and to reduce hard dietary fats in one’s diet.

The online inquiry task followed the phases included in the online research and comprehension framework (Leu et al., 2013). In the phase of Information search, the students were tasked to ponder their information needs and search for three relevant online texts for both topics by using a Google custom search engine, which included a set of pre-selected online texts (35 texts per topic). The online texts represented texts aligning with the current scientific view on the topic, that is, vaccination is beneficial (6 texts) and that the intake of hard dietary fats should be reduced (6 texts); texts opposing vaccination (2 texts) and texts arguing against reducing hard dietary fats (2 texts); and texts not useful for completing the task (27 texts for both topics). The relevant texts aligning with the current scientific view on the topic were research-based and published either by researchers, governmental or medical institutions, or doctors or communicated by journalists, health associations or educational institutions (see Hämäläinen et al., 2021). The imbalance between the texts aligning with either for or against the current scientific view on the topic is supposed to mirror the real-world availability of scientific information on these topics in digital media.

Contrary to the instructions given, the students selected a total of 40 texts on vaccination and 22 texts on dietary facts outside the custom search engine. These texts were analyzed and categorized according to the same criteria as the pre-selected texts. In the phase of Evaluating, the students defined the three most important points from the texts they had selected and defined points for and against the credibility of the texts. Finally, in the phase of Communication, after reading and evaluating the selected texts, the students responded to a fictitious young expectant mother on whether she should vaccinate her child (Vaccination topic) or to a fictitious young student on whether to reduce using hard dietary fats (Dietary Fats topic).

For their response, the students were asked to indicate their position by selecting one of the given response alternatives. Alternatives for the scientific view depending on the topic of the task were “The mother should vaccinate her child” (1a) and “The student should avoid hard dietary fats” (1b); Alternatives against the scientific view were “The mother should not vaccinate her child” (2a) and “The student should not avoid hard dietary fats” (2b); and alternative 3 was “Some other” (followed by a clarifying question “What”). Next, the students were asked to write down their reasons for the position they had selected. The data for this study consist of 342 student responses (3 incomprehensible responses excluded): 167 (48.8%) responses for the Vaccination topic, and 175 (51.2%) responses for the Dietary Fats topic. Student responses to the expectant mother included 15,290 words (M = 92) and to the young student 14,143 words (M = 81) in total.

The students’ responses were first classified as follows: For the scientific view, Against the scientific view, or Some other. Students’ responses to the alternative “Some other” were included in the two other categories based on whether the presented reasons supported the scientific view on the topic or not. When students’ argumentation (position and related reasons) was qualified or conditional in nature, a fourth category Conditional was formed (see an example below).

Example: A conditional position and related reasons (Student 1031).

Position: A small amount of hard dietary fats is neither harmful nor beneficial.

Reasons: You should both avoid and not avoid. A plentiful amount of hard dietary fats may expose you to serious cardiovascular diseases (a reason for), but a small amount is neither harmful nor beneficial (a reason against).

The reasons presented in the students’ responses were identified and counted, resulting in 1424 reasons in total: 758 reasons for Vaccination and 666 reasons for Dietary Fats. A variable Depth of argumentation was formed based on the number of reasons in the individual responses. The reasons were then examined for the argumentation perspective they represented. Based on data-driven classification (Hsieh and Shannon, 2005), six argumentation perspectives for Vaccination (Societal, Economic, Security, Efficacy, Risk, Other) were named. See argumentation perspectives (AP) and examples of representative reasons (R) for the scientific view of vaccination below:

Vaccination

AP: Societal

R1. Vaccination of people has prevented spreading of many diseases and also defeated many other diseases (St. 3005).

R2. Wide vaccination protects the population, so when you vaccinate your own child, you also protect the children of others (St. 1250).

R3. The Finnish health care is One of the best in the world, so you can take the risk of having your child vaccinated (St. 1152).

AP: Economic

R1. Vaccination is much cheaper than to fall ill with a serious disease and treat it (St. 1031).

R2. Vaccination is One of the most cost-efficient ways for the health care system to maintain health (St. 1159).

AP: Security

R1. Vaccinations that have been given to children are safe (St. 1064).

R2. Based on research knowledge, vaccines prevent many diseases serious to people (St. 1055).

R3. Children belong to a risk group for most diseases, and therefore it is important to vaccinate them (St. 1093).

AP: Efficacy

R1. Children get sick much less, almost not at all when the children have been vaccinated at an early age (St. 1015).

R2. Vaccines give permanent protection against diseases (St. 1046).

R3. Vaccination is a faster way to protect oneself than going through the disease (St. 3010).

AP: Risk

R1. No vaccination increases autism, diabetes, or allergies (St. 1001).

R2. The side effects of vaccines are rare, and symptoms are mild and brief (St. 1018).

R3. Vaccinations do not contain any more harmful substances than a child already receives from his/her environment (St. 1021).

AP: Other

R1. No vaccination is useless (St. 1076).

R2. Vaccination does its job (St. 1215).

R3. The most reliable sources are in favor of vaccination (St. 1235).

Note(s): St. = student

Further, four argumentation perspectives for Dietary Fats (Disease and risk, Health, Dietary guideline, Other) were named. See argumentation perspectives (AP) and examples of representative reasons (R) for the scientific view of dietary fats below:

Dietary fats

AP: Disease and risk

R1. Hard dietary fats increase blood cholesterol and cause heart diseases (St. 1008).

R2. Soft dietary fats do not impair the function of arteries, as they reduce blood cholesterol (St. 1020).

R3. In Finland, most people have a gene that makes hard dietary fats harmful to them (St. 1181).

AP: Health

R1. Hard dietary fats are not good for health (St. 2148).

R2. Soft dietary fats contain fatty acids vital for the body (St. 2002).

R3. Losing weight and weight management are difficult for a person if s/he consumes hard dietary fats (St. 1012).

AP: Dietary guideline

R1. Based on research results and dietary guidelines, hard dietary fats should be avoided in a diet (St. 2011).

R2. The Finnish Heart Association, for example, recommends substituting soft dietary fats for hard dietary fats (St. 2004).

R3. The guideline is: at least Two thirds of soft dietary fats and at most One third of hard dietary fats (St. 3029).

AP: Other

R1. The human body does not necessarily need any hard dietary fats (St. 1043).

R2. Fat is really high in calories (St. 1045).

R3. Hard dietary fats are produced by the human body, but soft dietary fats can only be obtained from food (St. 1181).

Note(s): St. = student

A variable Breadth of argumentation was formed based on the number of argumentation perspectives.

The analysis of the students’ use of sources in their argumentation was based on the documents model framework (Britt et al., 2014). Intertext links indicating relations between source information and content (source-to-content links) and between sources (source-to-source links) (Britt and Rouet, 2012; Hämäläinen et al., 2023; Kiili et al., 2021; Perfetti et al., 1999) were identified. A source-to-content link shows how the source and its content have been connected (e.g. “according to author X …”). A source-to-source link, in turn, shows how the writer has combined the content of various sources on the same topic, which may support, complement or contradict each other (e.g. “Document X contradicts document Y…”). Using source-to-source links is demanding as the writer has to compare and integrate the content of multiple sources, potentially dealing with inconsistent and conflicting information. Britt et al. (2014) found that dealing with conflicting information is important in science as it is an essential part of how a scientific community tests its knowledge and arrives at established truth.

The student responses were classified into five categories (variable Use of sources) based on various combinations of source-to-content and source-to-source links: Poor (no sources referred to); Passable (source contents have been referred to at a general level; no source-to-content link identified); Moderate (one source-to-content link used); Good (from two to three source-to-content links used); Excellent (from two to three source-to-content links and at least one source-to-source link used). Examples of students’ use of different intertext links are presented below:

A source-to-content link (St. 1048):

Position: The student should avoid hard dietary fats

Justifications: If I were you, I would avoid the heavy use of hard fats in the diet, as several sources say they raise cholesterol levels, which predisposes to cardiovascular disease. The Terveyskirjasto [Health Library] website says that hard fats have no irreplaceable functions in the body, while soft fats have plenty of benefits for the body.

A source-to-source link (St. 1119):

Position: The mother should vaccinate her child

Justifications: I think the child should be vaccinated becauseaccording to Terveyskirjasto [health library] and Mehiläinen [a private provider of health services], vaccination is effective and safe.According to Mehiläinen, the immunity given by vaccination often lasts a lifetime [– –]. Vaccinations are effective as, according to the health library, previous difficult child diseases are now history [– –].

Note(s): Key points to indicate the link type are in bold; St. = student; [– –] = The response has been shortened.

The dependent variables used in the study are Position assumed, Depth of argumentation, Breadth of argumentation and Use of sources, and the independent variables are Task topic and Prior topic knowledge. Associations between the variables were tested with independent samples t-tests, analysis of variance and Pearson’s correlation tests (r).

For both topics, the majority of the students assumed a position aligned with the scientific view of the topic: For Vaccination, 95.8% of the students thought that the expectant mother should vaccinate her child, and for Dietary Fats, 85.1% of the students thought that the young student should avoid using hard dietary fats in their diet. Only a small minority of the students adopted a position against the scientific view of the topic: 3.0% for Vaccination and 3.4% for Dietary Fats. Additionally, just over 1 in 10 (11.4%) adopted a conditional position (i.e. the response included reasons both for and against the scientific view of the topic) for dietary fats, while the corresponding proportion for the vaccination topic was only 1.2%. The association between the variables “Task topic” and “Position assumed” was statistically significant [χ²(2) = 15.03; p < 0.001]. The effect size (V = 0.21; Ellis, 2010) was, however, weak.

The students presented an average of 4.17 reasons (SD = 2.35) for their position in their responses. For the Vaccination topic, the average was 4.54 (SD = 2.62) and for the Dietary Fats topic 3.81 (SD = 2.01). The difference between the topics was statistically significant [t(340) = 2.89, p = 0.004] with a weak effect size (d = 0.31; Cohen, 1988).

A total of 43% of the students (see the two lowest categories in Figure 1: 11.7% + 31.3%) presented three or fewer reasons for their position in total. The proportion for the Dietary Fats topic was 48% and for the Vaccination topic 37.7%. Several reasons (8–12, see the two highest categories) were presented by only 8.7% (5.8% + 2.9%) of the students in total; the proportion for the Vaccination topic was 13.8% and for the Dietary Fats topic 4%.

The average number of argumentation perspectives in the students’ responses in total was 2.35 (SD = 1.11). For the Vaccination topic, the average was 2.71 (SD = 1.22) and for the Dietary Fats topic 2.01 (SD = 0.87), indicating a significant difference between the topics [t(340) = 6.07, p < 0.001] with a medium-sized effect size (d = 0.66).

A majority (56.1%) of students put forward 0–2 argumentation perspectives in their responses, showing narrow argumentation for the claim (see the three lowest categories in total in Figure 2). Extensive argumentation (either 4 or 5 argumentation perspectives in Figure 2) was expressed by 15.5% of the students in total, almost exclusively for the Vaccination topic (31.1%) rather than Dietary Fats (0.6%).

The average points students received in the use of sources in their argumentation was 1.78 (SD = 1.28) (min. 0; max. 4). For the Vaccination topic, students earned 1.784 points on average (SD = 1.28) and for the Dietary Fats topic, the average was 1.781 (SD = 1.28). The difference was not significant [t(343) = 0.026, p = 0.98, d = 0.00].

Students’ use of sources can be characterized as twofold (Figure 3): On the one hand, in total 39% of students used sources in a sophisticated way (either good or excellent performance), while on the other hand, 45% of the students performed quite weakly (either poor or passable performance). Moderate performance was demonstrated by 16% of the students.

An association was found between the students’ prior topic knowledge on dietary fats and the position they assumed on the topic [F(2, 172) = 6.96; p = 0.001)]. The mean of prior topic knowledge for students who assumed a position aligned with the scientific view (i.e. hard dietary fats in one’s diet should be avoided) was higher (M = 2.07, ranging from 0 to 3) than that of the students who supported either the opposite view (M = 1.17) or who assumed a conditional position (M = 2.00). The analysis showed a medium-sized effect size [(η²) = 0.075; Ellis, 2010]. For the Vaccination topic, the respective association was non-significant [F(2, 164) = 0.652, p = 0.522; η² = 0.008].

Intercorrelations among students’ prior topic knowledge, depth and breadth of argumentation and use of sources were tested with Pearson’s correlation (r) test, separately for both topics (Tables 1 and 2).

Students’ prior topic knowledge was somewhat associated with the depth of their argumentation, with a slightly stronger association for the Dietary Fats topic (r = 0.20, Table 2) than for the vaccination topic (r = 0.12, Table 1). In addition, for the Dietary Fats topic (Table 2), a slight association (r = 0.13) was found between prior topic knowledge and breadth of argumentation. Some associations were also observed between prior topic knowledge and students’ use of sources for both topics (Vaccination r = 0.16, Table 1; Dietary Fats r = 0.21, Table 2).

Depth and breadth of students’ argumentation were strongly associated with each other for both topics (Vaccination r = 0.81, Table 1; Dietary Fats r = 0.72, Table 2) and quite strongly with the use of sources, correlations varying from 0.37 to 0.44 (Tables 1 and 2).

This study examined Finnish upper secondary school students’ argumentation on two health topics, Vaccination and Dietary Fats, dealt with in online sources. During an online inquiry task, the students were asked to take a position on the topic and justify it by relying on the sources they had previously searched for and read. The analysis focused on the depth and breadth of students’ argumentation and how they used the online sources in their argumentation.

Almost half (43%) of the students presented few or no reasons (0–3) for their position, showing that these students’ argumentation was not very deep. Further, the majority (56%) of students put forward at most two argumentation perspectives in their responses, showing narrow argumentation for the claim (Figure 3). For the Vaccination topic, the students presented on average more reasons and more argumentative perspectives than for the Dietary Fats topic. These results are in line with previous studies showing that students at various educational levels have insufficient argumentation skills when dealing with health issues. For example, Puig et al. (2021) found that secondary school students had challenges in evaluating the credibility of COVID-19 claims and related scientific evidence presented in social media. Marttunen et al. (2021) reported on upper secondary school students’ difficulties in recognizing unbalanced argumentation in health-related online sources and, similarly, university students were found to have difficulties in assessing the reliability of digital health information and whether the information was commercially biased (Dadaczynski et al., 2021) when presented in social media.

The results indicated that the majority of the students assumed a position aligned with the current scientific view for both health topics. This is understandable since in Finland a positive attitude toward vaccination is very common (The Finnish Institute for Health and Welfare, 2024), and there is also a broad awareness of the official nutrition and food recommendations published by the National Nutrition Council (The Finnish Food Authority, 2024). However, conditional positions also occurred, interestingly 10 times more often for the Dietary Fats topic than for the Vaccination topic (11.4% vs 1.2%). This result may be a sign of students’ critical thinking, showing that they recognized the complexity of the topic of Dietary Fats, requiring weighing multiple factors. The question of using dietary fats is somewhat more of a multidimensional health issue than the question of the usefulness of vaccination, for which scientific research gives clearer evidence. Even if conditional arguments are typical in informal reasoning that often aims to persuade a person to a certain point of view (Thompson et al., 2005), these two health topics differed in nature. Vaccination as a topic offered a clear for vs against situation for the use of vaccines, whereas Dietary Fats was not that polarized a topic but was both conceptually and argumentatively more complex. This may explain the greater number of students’ conditional positions for dietary fats. In addition, in discussions on dietary fats, such concepts as hard and soft fats as well as saturated and unsaturated fats are commonly used (The Finnish Food Authority, 2024), often leading the discussion around to the question of how much of these different fats should be included in the diet. When favoring the position that hard dietary fats should be avoided, the students, in fact, had to consider arguments not only against hard dietary fats but also arguments supporting the use of soft dietary fats. By contrast, when presenting arguments on vaccination, the students could concentrate on only one dilemma: whether vaccination should be supported or not; this may have affected their more extensive argumentation for vaccination. The students would probably have needed more time for their argumentation on dietary fats than they were allocated when writing their responses for the task.

Students’ use of sources in their argumentation was, on average, moderate. However, the proportion of students possessing weak sourcing skills (either poor or passable) was high (45%), while the proportion of students possessing excellent skills was remarkably low (only 6%). This result is in line with previous studies showing that students struggled with sourcing when producing written products (e.g. Anmarkrud et al., 2021; Kiili et al., 2020; Pérez et al., 2018). Making relations between source information and content (source-to-content links) is somewhat easier than making links between sources (source-to-source links), which is cognitively a demanding task as it requires comparing different sources and building a coherent representation of the examined issue through multiple sources (Hämäläinen et al., 2023; Kiili et al., 2021; Perfetti et al., 1999). This, at least partly, explains the small number of source-to-source links in the students’ responses. In successful online inquiry, both argumentation and sourcing skills are needed throughout the process (Barzilai et al., 2015; Forzani et al., 2022; Kiili et al., 2021). Students seem to need more teaching and practice on how to use sources and the information related to them in an argumentative and synthesizing way when engaging in online inquiry tasks. Hämäläinen et al. (2023), for example, reported positive results of a teaching intervention in which different online inquiry phases were supported from the viewpoint of sourcing.

The results showed, concerning both Vaccination and Dietary Fats, correlations between students’ prior topic knowledge and both their argumentation (from 0.12 to 0.20) and their use of sources (0.16 and 0.21). Although low, the correlations were higher for Dietary Fats. Further, for Dietary Fats, the level of prior topic knowledge was lower among those students who assumed a position not aligned with the scientific view of the topic. These results suggest that better prior topic knowledge on health issues seems not only to predict better argumentation but also opinion formation aligned with scientific knowledge. Thus, it would be important to spread objective and research-based knowledge on health issues through reliable channels to promote opinion formation and related life decisions based on correct and reliable information (see, e.g. Freeman et al., 2023; George, 2024). This is particularly important when aiming at supporting broader public health issues, such as increased vaccination coverage and healthy eating habits in society. The results are consistent with findings by Baytelman et al. (2020), whose study on undergraduate students showed that prior knowledge had a significant predictive power for the quantity, quality and diversity of arguments students constructed on controversial and complex socioscientific issues, such as the use of dietary fats. Kiili et al. (2024) also found that prior knowledge and reading comprehension were statistically significant predictors of adolescents’ justification performance.

There are some limitations in this study that are important to note. First, although the study was carried out in an authentic environment as a part of students’ normal schoolwork and was developed in collaboration with their teachers, the online inquiry task assignment used was not necessarily a typical task for them (see Kiili et al., 2022). This may have affected the students’ performance. Another limitation is that although students’ prior topic knowledge was measured, we did not collect information on students’ epistemic beliefs, the role of which is highlighted by Baytelman et al. (2020). They emphasize that epistemic beliefs, that is, individuals’ beliefs about the nature of knowledge and knowing, may either promote or limit conceptual change and understanding of a topic under consideration, particularly when arguing about socioscientific issues. If students view knowledge as a complex system of theoretical concepts rather than discrete pieces of factual information, they tend to recognize and consider different viewpoints to develop their conceptual and argumentative understanding of controversial socioscientific topics. The third limitation is that we did not measure students’ engagement or interest in the topic, both important aspects to motivate building sound arguments justified by relying on multiple sources and reliable knowledge (Brandmo and Bråten, 2018). Nevertheless, in the present study, the task assignments in both topics were designed to support the students’ motivation to engage in argumentation as well as their health information behavior. In the Dietary Fats topic, they were asked to respond to a young student and in the Vaccination topic to a young expectant mother, individuals whose life situations the participants of this study were expected to identify with easily. Finally, it is important to note that the data were collected from Finnish students within the context of the Finnish educational system, which limits the generalizability of the findings to other cultural settings.

In future studies, measuring students’ epistemic beliefs alongside their prior knowledge and topic interest would give important information to increase our understanding of the nature and quality of students’ argumentation and sourcing on health issues in online sources. In addition, more detailed qualitative analyses of students’ argumentation are needed to yield deeper insights into the nuances of students’ argumentation.