Access to healthy ınformatıon: the ınteractıon of medıa lıteracy and health lıteracy

Abstract

Objective

Grounded in the Health Belief Model (HBM) and Media System Dependency Theory (MSDT), this study examines the relationship between media literacy and health literacy in the Turkish context, where digital media plays an increasingly dominant role in disseminating health-related information. It focuses on how individuals access, evaluate, and utilize health information, analyzing the effects of media literacy sub-dimensions—access, analysis, evaluation, and communication—on health literacy levels. The study aims to determine how media-based health information influences individuals’ decision-making processes and their ability to make informed health choices.

Methods

A survey-based quantitative research approach was employed with 485 participants from Turkey, of which 477 valid responses were analyzed. The Turkish Health Literacy Scale (TSOY-32) measured health literacy, while the Media Literacy Skills Scale assessed media literacy. The Generalized Ordered Logit Model (GOLM) was applied to examine the effects of media literacy on health literacy.

Results

Descriptive statistics, correlation analysis, and hypothesis testing were conducted. Media literacy levels significantly influence health literacy levels (p < 0.05). The access sub-dimension increased the likelihood of improving health literacy by 1.90 times (p < 0.05). The analysis sub-dimension significantly enhanced health literacy by 2.23 times (p < 0.05). The evaluation sub-dimension had a statistically significant effect on health literacy (p < 0.05). The communication sub-dimension supports individuals in sharing health information and making informed decisions (p < 0.05). Health recommendations disseminated through media significantly impact health literacy levels (p < 0.05). Trust in media-based health information had a weak but positive correlation with health literacy (r = 0.2097, p < 0.05). Education level was positively correlated with health literacy, while income level showed no significant effect. These findings suggest that enhancing media literacy skills, particularly in access and analysis, can meaningfully improve individuals’ capacity to navigate complex health information landscapes and foster more informed health behaviors.

Conclusion

The study highlights the crucial role of media literacy in improving health literacy and preventing misinformation. Access and analysis skills are particularly vital in enabling individuals to obtain accurate health information and avoid misleading content. Conducted within the Turkish sociocultural context, where digital media usage is high and misinformation about health is prevalent, the findings offer valuable insights for national health promotion strategies. Policymakers and educators should integrate media literacy into educational curricula, launch public awareness campaigns to combat misinformation, and develop strategies to enhance individuals’ critical thinking skills regarding health information. This study is limited by its cross-sectional design and reliance on self-reported data. Future research should employ longitudinal methods and explore the long-term impact of media and health literacy across different demographic groups and contexts.

Background

In the digital era, individuals’ ability to make informed health decisions is increasingly shaped by their access to and understanding of health information. Health literacy—defined as the cognitive and social ability to obtain, process, and apply health-related knowledge—is considered a cornerstone of public health [1, 2]. As technological tools rapidly evolve, the importance of health literacy has become more pronounced, especially in societies with high digital penetration. According to the World Health Organization (WHO), health literacy is not only a determinant of personal well-being but also a vital mechanism for reducing health disparities across populations [3].

Empirical evidence from various countries, including Turkey, Iran, Denmark, and the Netherlands, has consistently shown a strong relationship between education and health literacy [4]. For instance, Holt et al. [5] identified a significant association between digital literacy and health-related decision-making in nursing students in Denmark. Friis et al. [6] further demonstrated that health literacy mediates the relationship between education and health behaviors, suggesting that improving health literacy can serve as a pathway for enhancing public health outcomes. Such findings emphasize the strategic role of health literacy in addressing access inequalities and promoting equitable health services [7].

In today’s highly connected world, the internet serves as the primary medium through which individuals access health information. For instance, in 2024, Turkey reported over 74.4 million internet users and an internet penetration rate of 86.5%; approximately 66.8% of the population actively uses social media platforms [8]. While this digital access facilitates the rapid dissemination of health knowledge, it also amplifies the risk of exposure to misinformation and disinformation, which pose critical challenges to public health [9, 10].

A 2024 OECD report revealed that 59% of adults in member countries are concerned about misinformation encountered online, especially when it pertains to health topics [11]. A telling example is found in a Turkish study, which observed that pregnant women misunderstood the safety of glucose tolerance tests during pregnancy due to misleading media narratives, thereby avoiding a critical diagnostic tool for gestational diabetes [12]. This highlights a worrying dynamic: digital misinformation has the potential not only to shape individual health behaviors but also to exacerbate inequalities in the utilization of health services.

These issues bring media literacy into sharp focus. Media literacy refers to the set of competencies that enables individuals to access, analyze, evaluate, and create media content with critical awareness [13]. Understanding the political, economic, and ideological contexts in which media operate enhances individuals'capacity to distinguish credible health information from biased or misleading narratives [14]. Paek et al. [15] argue that media environments heavily influence youth health literacy, as adolescents frequently turn to online platforms for guidance on sensitive health matters.

In the current digital landscape, individuals who lack media literacy are more likely to misinterpret or be misled by health content, potentially leading to adverse health outcomes or distrust in medical institutions. As Nutbeam posited, health literacy should not be limited to understanding information but must also encompass the ability to act on it to improve health outcomes [16]. This assertion is especially relevant today, where a large share of health content is produced and disseminated through non-traditional sources such as influencers, peer networks, and unregulated platforms [17].

Media literacy supports the critical analysis of health content, allowing users to interpret underlying motives and distinguish scientifically sound messages from harmful misinformation [18]. Community-based training programs that integrate both health and media literacy have shown promise in promoting informed health decisions and improving preventive care behaviors [19].

This study is conceptually anchored in two major theoretical frameworks: the Health Belief Model (HBM) and the Media System Dependency Theory (MSDT). HBM, one of the most widely used frameworks in health psychology, suggests that health behaviors are driven by perceived susceptibility, severity, benefits, and barriers [20, 21]. Fleary et al. [20] expanded the scope of HBM by highlighting that it also shapes how individuals engage with and comprehend health information, positioning health literacy as both a predictor and an outcome of health behaviors.

Incorporating media literacy into this framework helps us understand how individuals assess the risks and benefits presented in media messages. For example, a study by Alqarni et al. [22] demonstrated that misinformation regarding the COVID-19 vaccine diminished individuals’ willingness to receive booster doses—a phenomenon that HBM explains through altered perceptions of vaccine risks and benefits. Similarly, Njoku and Ohiagu [23] found that awareness of counterfeit medicines was significantly linked to media literacy, as individuals capable of evaluating and verifying health information were more likely to avoid unsafe pharmaceutical products.

The explanatory potential of these findings is further enriched by the Media System Dependency Theory (MSDT), which argues that individuals increasingly depend on media systems to understand their environment, particularly during crises [24, 25]. Originally developed by Ball-Rokeach and DeFleur [25], MSDT provides a onceptual framework for understanding how individuals’ reliance on media for goal attainment—such as health protection—intensifies their susceptibility to both credible information and misinformation. This theory has become particularly relevant in the context of social media, where content algorithms often reinforce user biases and undermine trust in science-based health recommendations.

Given this, the current study examines the intersection of media literacy and health literacy through the lenses of HBM and MSDT, analyzing how sub-dimensions of media literacy—access, analysis, evaluation, and communication—affect individuals’ ability to comprehend, trust, and apply health information [26, 27].

Research hypotheses

In the study, basic hypotheses and sub-hypotheses were tested in order to understand the effects of media literacy sub-dimensions on health literacy levels.

Main Hypothesis:

• : The level of media literacy has a statistically significant effect on the level of health literacy.

Sub Hypotheses:

• : The “access” sub-dimension of media literacy has a statistically significant effect on health literacy levels.

• The “analysis” sub-dimension of media literacy has a statistically significant effect on health literacy levels.

• : The “evaluation” sub-dimension of media literacy has a statistically significant effect on health literacy levels.

• : The “communication” sub-dimension of media literacy has a statistically significant effect on health literacy levels.

• : Health recommendations published through the media have a statistically significant effect on health literacy levels.

Research gap and study contribution

Despite the growing interest in the relationship between media literacy and health literacy, there remains a significant methodological gap in the econometric modeling of this relationship. Most existing studies in this field utilize linear or standard ordered logit models, which assume that the effect of explanatory variables is constant across all thresholds of the ordinal dependent variable. However, the proportional odds assumption is frequently violated in social science data, particularly in areas like health literacy where transitions between categories (e.g., inadequate to adequate, or adequate to excellent) may be influenced differently by each independent variable.

This study contributes to the literature by applying the Generalized Ordered Logit Model (GOLM), which relaxes this assumption and allows coefficient estimates to vary across categories. From an econometric perspective, this approach provides a more robust and flexible modeling framework, enabling a more accurate understanding of how different dimensions of media literacy -such as access, analysis, and recommendation trust-affect the likelihood of upward transitions in health literacy levels. This methodological contribution is particularly important for researchers and policymakers who rely on nuanced, category-specific insights to design effective public health interventions and literacy programs.

By addressing the limitations of traditional ordinal models, the present study enhances the methodological rigor in this research domain and provides a replicable econometric framework for future studies involving ordinal outcomes in media and health behavior research.

Methodology

SPSS 25 and Stata 15 package programmes were used in the analysis of the data obtained within the scope of the research. The variables used in the study were structured according to the scope of the data collection tools and by utilising the literature in accordance with the purpose of the study. The total scores obtained from the TSOY-32 scale, which was used to measure the participants’ levels of understanding, evaluating and using health information, were defined as the dependent variable to determine their health literacy levels. Media literacy was assessed through four main sub-dimensions: access, analysis, evaluation and communication. The access dimension aims to measure the participants’ capacity to access information and their ability to use media sources effectively. The analysis dimension assesses the ability to critically analyze media content. The evaluation dimension measures the ability to evaluate the reliability and accuracy of media content. The communication dimension aims to assess the skills to share information and communicate using media tools. In addition, a scoring system between 1 and 10 was used to measure the participants’ trust in health advice published through the media and the effects of these advice on health.

Measurement tools

In this study, self-administered online questionnaire method was used in line with the findings in the literature and the aim of the research. The study was conducted in accordance with the principles defined by the Declaration of Helsinki, and informed consent was obtained from the participants. Ethical approval (date: 08 July 2024, no: 586) was obtained from Istanbul Beykent University Scientific Research and Publication Ethics Board for Social Sciences and Humanities.

The main population of the study consisted of adult individuals residing in Turkey who had access to the internet and were reachable through digital platforms such as university e-mails, online education portals, and social media networks. A sample group consisting of individuals selected by convenience sampling method was determined to represent the main population.

The survey was conducted online between August and September 2024, and voluntary participation was encouraged through anonymous distribution links. Individuals aged 18 and above, from various educational and socio-economic backgrounds, were included in the target group. In this context, a total of 485 people were surveyed, and 477 usable questionnaires were included in the analysis [28].

During the sampling process, the recommendations in the literature were taken into consideration in order to ensure that the sample is representative of the main population and to minimise sampling errors.

In determining the adequacy of the sample size, it was ensured that the number of observations was sufficient to provide reliable results. According to Young [29] a sufficient sample is one that includes enough units to yield reliable findings. In addition, Meyers et al. [30] recommend at least 10 observations per variable in multivariate analyses, and Velicer and Fava [31] support this rule of thumb. Considering the number of variables in this study, the sample size of 477 was deemed sufficient for conducting reliable statistical analyses such as factor analysis and the Generalized Ordered Logit Model.

Although the sampling method was non-probabilistic, efforts were made to ensure demographic diversity in terms of age, gender, education level, and regional distribution.

In order to measure health literacy, the Turkish Health Literacy Scale (TSOY-32) [32], consisting of 32 items and developed by the Ministry of Health, was used. The five-point Likert-type scale includes options from “Very Easy” to “Very Difficult”. The scale results were converted to a standard index in the range of 0–50 and the following formula was applied for the conversion: Index Score = (Arithmetic Mean—1) × (50/3). The obtained scores were classified according to the categories defined in the literature: 0–25 points were categorised as “Inadequate Health Literacy”, > 25–33 points as “Problematic—Limited Health Literacy”, > 33–42 points as “Adequate Health Literacy” and > 42–50 points as “Excellent Health Literacy”. High reliability values were obtained in the total and sub-dimensions of the scale. Cronbach’s alpha coefficient was calculated as 0.88 for the overall scale, 0.94 for the “Treatment and Service” sub-dimension and 0.90 for the “Disease Prevention and Health Promotion” sub-dimension. When the distribution of Turkey Health Literacy Scale scores by categories was analysed, 46.86% of the participants were in the “Problematic-Limited Health Literacy” category, 21.97% in the “Inadequate Health Literacy” category, 20.92% in the “Adequate Health Literacy” category and 10.25% in the “Excellent Health Literacy” category. These results show that most of the participants have a problematic-limited level of health literacy. The mean total score of the scale was calculated as 30.41 ± 8.37 and it was observed that this value was close to the “Problematic-Limited Health Literacy” category. As a result of the Kaiser–Meyer–Olkin (KMO) test, the sampling adequacy index was found to be 0.94. As a result of this value, it was decided that it was perfectly suitable for factor analysis. Therefore, it was supported that the data were suitable for factor analysis.

The 45-item Media Literacy Skills Scale developed by Erişti and Erdem [33] was used to assess media literacy. The scale consists of four sub-dimensions: Access (items 1–11), Analysis (items 12–26), Evaluation (items 27–33) and Communication (items 34–45).

The selection of these four media literacy sub-dimensions—access, analysis, evaluation, and communication—is theoretically grounded in prior research emphasizing that these skills are central to individuals’ ability to navigate and critically engage with health information [13, 14, 18]. Each dimension reflects a distinct cognitive or behavioral process that influences how media content is interpreted and used for health-related decision-making.

The overall reliability coefficient of the scale was calculated as 0.9747. Cronbach alpha values ​​for the sub-dimensions were determined as 0.9194, 0.9559, 0.9181 and 0.9038, respectively. As a result of factor analysis, it was determined that the first factor, which explained 73.18%, had the highest eigenvalue (21.52699). We also developed an additional scoring section ranging from 1 (lowest) to 10 (highest) to assess the impact of health advice provided through the media on individuals’ health perception, confidence, and decision-making processes. This section was created specifically for this study to assess how participants evaluated the health information they obtained from media sources and its subsequent impact on both individual and societal health. In developing this variable, content validity was prioritized, as existing trust scales in the literature did not sufficiently reflect the multidimensional nature of media-based health communication in the digital era. Internal consistency analysis showed high reliability, with a Cronbach’s alpha coefficient of 0.89. Moreover, the inclusion of a media trust variable is theoretically supported by prior research demonstrating that trust in media-based health information significantly influences individuals’ health behaviors and literacy outcomes. This justifies its integration into the current model to better capture the mediating role of trust in the relationship between media literacy and health literacy. This trust construct was used as an independent variable in the regression model. Theoretical and empirical support for including such a variable stems from prior research that links media trust to individuals’ health behaviors and attitudes [34, 35].

Sample structure

The demographic characteristics of the participants are presented in Table 1. As shown, the gender distribution indicates that 60% of the respondents were female, 35% were male, and 5% preferred not to disclose their gender. This result suggests a higher representation of female participants in the sample. This distribution is considered normal and contextually relevant, as the proportion of female employees in the healthcare sector in Turkey is considerably higher than that of males, particularly in nursing, public health, and caregiving roles.

Table 1.

Demographic characteristics of the sample

Variable	Categories	Frequency (n)	Percentage (%)
Gender	Female	286	60
	Male	167	35
	Prefer not to say	24	5
Age Group	18–29	119	25
	30–39	167	35
	40–49	96	20
	50 and above	95	20
Education Level	Primary	29	6
	Secondary	72	15
	Associate Degree	96	20
	Bachelor’s Degree	167	35
	Master’s Degree	57	12
	Doctorate	48	10

The participants’ age ranged from 18 to 73 years, with a mean of 35 and a median of 33. In terms of age groups, 25% were between 18–29 years, 35% between 30–39 years, 20% between 40–49 years, and another 20% were aged 50 and above.

Regarding education level, 35% of the participants held a bachelor’s degree, 20% had an associate degree, 15% completed secondary education, 12% held a master’s degree, 10% held a doctorate, and 6% had only primary education. Additionally, 2% of the respondents were literate but had not completed formal schooling. The data suggest that individuals with postgraduate degrees (master’s and doctorate) generally had income levels above the sample average, indicating a correlation between educational attainment and economic status.

In terms of income distribution, the average income was 68,000 TL, while the median was 50,000 TL. The income range spanned from 0 TL to 800,000 TL. Notably, 10% of the participants reported no income, and 5% earned 200,000 TL or more annually. Furthermore, 20% reported income between 40,000 and 60,000 TL. These findings highlight considerable income inequality within the sample. It was also observed that the average income of female participants was 12% lower than that of males. Younger participants (ages 18–29) tended to fall into the lower income brackets, while participants aged 50 and above were more likely to be in higher income groups.

In the research, various findings were obtained through the questions asked about the media use and digital access habits of the participants. In the participants’ preference of television programmes, news programmes were the most watched content type with a total of 297 preferences. These programmes were followed by films with 226 preferences, documentaries with 199 preferences, TV series with 187 preferences and information-culture competitions with 161 preferences. These results show that the participants attach high importance to informative and educational content. In particular, the fact that news programmes are the most preferred genre indicates the sensitivity of individuals to access current events. When the internet usage habits of the participants are analysed, the most common reason for using the internet for research and information purposes was stated by 157 people (32.9%). This category includes activities such as searching for information and doing homework. This was followed by communication use (e-mail, chat, etc.) with 72 respondents (15.1%) and entertainment use (games, surfing, etc.) with 50 respondents (10.5%). Internet use for commercial purposes (shopping, banking/investment transactions, etc.) was preferred by only 28 people (5.9%). The research findings also provide some important points about media literacy and digital access levels. Regarding media literacy education, the majority of the participants (84.3%) stated that they had not taken a media literacy course before. This situation indicates that media literacy trainings should be made widespread. In terms of computer access, 4.6% of the participants (22 people) stated that they do not have a computer that they can use whenever they want. Although internet access is quite common, only 1.3 per cent of the participants (6 people) stated that they cannot access the internet whenever they want. Finally, when newspaper reading habits are analysed, 39.2% (187 people) of the participants stated that they do not have a newspaper that they follow constantly.

Results

Descriptive statistics for the basic variables included in the study are presented below. These statistics allow the evaluation of the variables in terms of mean, median, standard deviation, skewness, kurtosis and normality distribution.

Table 2 shows the descriptive statistics of health literacy (HL) and media literacy sub-dimensions. The mean value of the “Health Literacy Level (HLL)” variable was 2.194 and the standard deviation was 0.896, and the distribution was right skewed and not suitable for normal distribution (Jarque–Bera, p < 0.05). The mean values of the variables Evaluation, Communicating, Accessing and Analysing, which are sub-dimensions of media literacy, were calculated as 3.978, 3.823, 3.869 and 3.969, respectively. These variables do not meet the assumption of normal distribution to a great extent (p < 0.05). The mean of the recommendation variable was 5.159, the standard deviation was 1.966, and it conformed to the normal distribution (p = 0.319). When evaluated within the scope of the research questions, the average value of the variable related to the health literacy level being 2.19 indicates that the individuals participating in the study have a low level of health literacy in general. This situation reveals that there are limitations in the capacity of individuals to understand, evaluate and use health information and therefore this competence needs to be developed. When the sub-dimensions of media literacy are examined, the means of access (3.869), analysis (3.969), evaluation (3.978) and communication (3.823) are close to each other but are at a medium level. These findings show that individuals have a certain awareness of accessing health information through the media and analyzing the content, but these competences do not develop equally in all individuals. In particular, the negative skewness of the analysis and evaluation variables suggests that some individuals have high competence in these areas but a large portion concentrates on lower scores. This emphasizes the importance of media literacy-based interventions aimed at increasing health literacy. On the other hand, the average level of trust in health advice provided through the media is 5.159, which shows a higher and more balanced distribution compared to other variables. This finding makes it meaningful to examine that individuals have a medium level of trust in media-sourced health information and the effect of this trust on health literacy. In general, descriptive statistics show that there are different levels of development between media literacy and health literacy, and confirm the necessity of analyses that examine this relationship.

Table 2.

Descriptive statistics

Variable Statistics	HLL*	Evaluation	Communicating	Access	Analysis	Recommendation**
Average	2.194	3.978	3.823	3.869	3.969	5.159
Median	2.000	4.000	3.923	4.000	4.000	5.000
Standard Deviation	0.896	0.658	0.689	0.672	0.633	1.966
Skewness	0.465	−0.774	−0.550	−0.862	−0.910	0.103
Kurtosis	2.527	5.008	4.021	4.859	5.717	2.731
Jarque–Bera	21.640	127.841	44.844	127.846	212.774	2.284
Probability	0.000	0.000	0.000	0.000	0.000	0.319

^*The HLL (Health Literacy Level) variable categorically represents the health literacy level of individuals (1 = Insufficient, 2 = Problematic–Limited, 3 = Adequate, 4 = Excellent). The “Evaluation” variable expresses the ability of individuals to critically evaluate health information they access through the media; while the “Communicating” variable shows their skills in sharing this information with others or discussing it in media environments. The “Access” variable represents the ability to access and reach health-related media content; while the “Analysis” variable expresses the ability to analyze and make sense of this content

^**The Recommendation variable is an independent variable created to measure participants’ levels of trust in health recommendations offered through the media and their perceptions of the effects of these recommendations on individual health and is rated between 1 and 10. In addition, while the skewness and kurtosis values ​​show the distribution properties of the variables, the Jarque–Bera test tests whether there is a deviation from the normal distribution (if p < 0.05, the variable is considered to be not normally distributed)

When the skewness and kurtosis values were analysed, it was observed that most of the variables were not within the range of −1.5 and + 1.5 suggested by Tabachnick and Fidell [36]. Since this situation shows that the data are not suitable for normal distribution, nonparametric analysis approaches were used in the study.

Spearman correlation analysis, which was conducted to determine the relationship between the basic variables in the study, revealed the relationship between the health literacy level (HLL) and the media literacy sub-dimensions and the level of trust in health advice obtained through the media. The findings show that there are positive and statistically significant relationships between the HLL variable and the media literacy sub-dimensions of evaluation (r = 0.3174), communication (r = 0.2638), access (r = 0.3201) and analysis (r = 0.3121) (p < 0.05). This situation indicates that the increase in the media literacy skills of individuals can increase their health literacy levels. However, the relationship between the “Recommendation” variable, which measures trust in health advice, and the health literacy level was found to be weak (r = 0.2097). This finding suggests that the trust in health advice by individuals may have a limited effect on health literacy. In addition, the correlations between the trust variable and the media literacy sub-dimensions are mostly weak and in some cases not significant (for example, r = 0.0652, p > 0.05 with the analysis). This result reveals that there is no strong relationship between the participants' capacity to analyze media content and their confidence in these contents. The income variable exhibited a negative and statistically insignificant relationship with HLL (r = −0.0601, p > 0.05). Similarly, the correlations between income status and other media literacy variables were also low and most were not found significant. Therefore, the income variable was not included in the model. This analysis shows the direction and strength of the relationships between the variables; however, it does not allow causal inference. Therefore, in order to evaluate the factors affecting the level of health literacy in more depth, the generalized ordered logit regression model was used for the ordinal dependent variables in the study.

The findings obtained as a result of testing the parallel regression assumption are given in Table 3. When the results in Table 4 are examined; since the assumption was rejected according to all test findings, it was decided to use the generalized ordered logit model instead of the ordered logit model in the study. Since the generalized ordered logit model allows the effects of the independent variables to vary according to the categories, it provides more accurate estimates when the parallel regression assumption is not met. The basic assumption of the model is that the dependent variable Y has K + 1 ordered categories of 0,1,2,…,K. The cumulative probability for each category is calculated with the logistic cumulative distribution function (G). In this case, α_j is the constant term for each category, is the coefficient vector for each category, X_i is the vector of independent variables and the model is expressed as The generalized ordered logit model removes the parallel regression assumption by assuming that the effects of the independent variables are not constant across categories, i.e. ​ This makes it possible to model category-specific effects; probabilities for each category are calculated by taking the difference of cumulative probabilities [37].

Table 3.

Spearman correlation analysis

Variables	HLL	Evaluation	Communicating	Access	Analysis	Recommendation	Income status
HLL	1
Evaluation	0.3174*	1
Communicating	0.2638*	0.7018*	1
Access	0.3201*	0.6468*	0.6569*	1
Analysis	0.3121*	0.7666*	0.6810*	0.7336*	1
Recommendation	0.2097	0.0566	0.1222*	0.0998*	0.0652	1
Income Status	−0.0601	0.0775	0.0152	0.0549	0.1319*	0.0506	1

^*indicates a statistically significant relationship at 5% significance level

Table 4.

Parallel lines assumption tests

Statistics test	χ2	P > χ 2
Wolfe-Gould	28.52	0.001
Score	42.28	0.000
Likelihood Ratio	33.1	0.000
Wald	336.5	0.000

Table 5 presents the estimation results of the generalized ordered logit model. The Health Literacy Level (HLL), which is used as the dependent variable in this model, was analyzed in ordered categories. The highest category, “Excellent Health Literacy” level (4), was accepted as the reference category. This choice reflects a modeling strategy aimed at understanding how the transition probabilities from subcategories to the reference category change. Only the recommendation, access and analysis variables that were found to be significant were included in the model; the evaluation and communication sub-dimensions were excluded because they did not contribute significantly to the model in the preliminary analyses and disrupted the model fit.

Table 5.

Generalised ordered logit regression estimation results

HLL	Variable	Coefficient	Odds ratio	P >|z|
1	Recommendation	0.2398	1.2710	0.000
	Access	0.3767	0.4575	0.137
	Analysis	0.3177	1.3740	0.258
	Fixed Term	2.5899	0.0750	0.002
2	Recommendation	0.1482	1.1598	0.011
	Access	0.6409	1.8982	0.018
	Analysis	0.8001	2.2259	0.005
	Fixed Term	7.3743	0.0006	0.000
3	Recommendation	0.2014	1.2232	0.013
	Access	1.0666	2.9055	0.008
	Analysis	1.3043	3.6851	0.002
	Fixed Term	13.2430	0.0003	0.000

Recommendation variable refers to trust in media-based health information (here after referred to as the Recommendation variable in tables) measured on a 1–10 scale

Waldχ² (9): 70.36

p > χ²: 0.000

Pseudo R² = 0.2880

The model results show that the recommendation variables, which measure the trust in health recommendations disseminated through the media, access, which represents the capacity to access information, and analysis, which represents the ability to interpret media content, have significant effects on the health literacy levels of individuals. In the transition from the “insufficient” health literacy level to higher categories, each unit increase in the recommendation variable statistically significantly increases the transition probability (OR = 1.27; p < 0.01); However, the effects of access and analysis variables were not found to be significant at this stage. In the transition from “Problematic—Limited” level to higher categories, the effects of all three variables were significant; the effects of access (OR = 1.90) and analysis (OR = 2.23) variables were more pronounced. The strongest effects were observed in the transition from “Sufficient” level to “Excellent” level; recommendation (OR = 1.22), access (OR = 2.91) and analysis (OR = 3.69) variables were significantly associated with the probability of transition.

Within the scope of the research design, the main hypothesis () and five subhypotheses (–, ) were developed to test the effects of the sub-dimensions of media literacy on health literacy. As a result of the preliminary analyses, the variables access (), analysis () and health advice provided through the media () were included in the generalized ordered logit model because they showed statistically significant relationships with the level of health literacy. On the other hand, the sub-dimensions evaluation () and communication () did not show significant relationships in the correlation analyses and were not included in the final model because they were insufficient in terms of model fit.

Therefore, the hypotheses , and were empirically supported. Although hypotheses and were theoretically predicted, they were excluded from the model because they were not sufficient in terms of statistical validity and this situation was explained and reported. This approach was preferred to preserve the simplicity of the model and to avoid multicollinearity problems.

These findings show that the level of health literacy is closely related not only to access to information, but also to the evaluation of this information and the trust in media recommendations. The high effect of the analysis variable, especially in the transition from “Sufficient” level to “Excellent” level, emphasizes the critical role of information evaluation skills. Therefore, establishing trust in media recommendations, strengthening access to information and developing analysis skills should be among the main priorities in strategies to increase health literacy.

Discussion

This study examined how subdimensions of media literacy—access, analysis, evaluation, and communication—affect health literacy within the Turkish population, framed by the Health Belief Model (HBM) and Media System Dependency Theory (MSDT). The findings contribute to a growing body of literature on how individuals process health information in a digital environment marked by both unprecedented access and high exposure to misinformation [38, 39].

Among the subdimensions, evaluation and analysis had the strongest associations with health literacy. These findings reinforce the idea that interpreting and assessing media content critically is essential for health decision-making [40, 41]. The HBM suggests that individuals’ perceptions of health threats and benefits influence behavior, and these perceptions are often shaped by the ability to evaluate media messages [42, 43]. Those who are capable of assessing credibility are more likely to trust reliable sources and act on accurate health advice [44].

Access, while foundational, showed a weaker predictive relationship, supporting findings that availability alone does not guarantee comprehension or appropriate application [45, 46]. According to MSDT, individuals increasingly rely on media systems when faced with complex or uncertain situations, such as health crises. This dependency strengthens the need for interpretive media skills to navigate overwhelming streams of content [47, 48].

The communication subdimension, though less emphasized in prior studies, emerged here as particularly relevant. Sharing or forwarding health content plays a powerful role in peer influence and the spread of public health messages [49, 50]. Informed sharing enhances community-level awareness and supports collective health actions. However, careless dissemination of unverified content can spread misinformation, reinforcing the critical importance of responsible communication behaviors. As highlighted by Karadag and Yigit [51], digital misinformation often circulates more rapidly through interpersonal networks than institutional channels. Similarly, Ucar et al. [52] stress the pivotal role of peer sharing in amplifying both accurate and inaccurate health content. In this regard, Eren and Yılmaz [53] emphasize that communicative behaviors in online settings directly influence public trust and health decision-making.

Importantly, the study validates the theoretical use of HBM and MSDT in interpreting digital health behaviors. HBM offers insight into how media literacy affects perception of susceptibility, severity, and benefits, which are central to health-related decision-making [54]. MSDT complements this by explaining why individuals turn to media systems to fulfill goals such as social understanding, orientation, and information-seeking—especially in contexts of uncertainty like a pandemic [55, 56]. Together, the two theories provide a robust lens through which the findings can be understood.

Nevertheless, the study has certain limitations. The convenience sampling method restricts the generalizability of results, as the sample may not reflect national diversity in education, socioeconomic status, or digital access [57, 58]. Second, the use of self-report measures invites the possibility of social desirability bias, potentially inflating positive behaviors or underreporting negative ones [59]. While validated scales were used, the subjective nature of responses remains a concern. Third, the study is based on a cross-sectional design, which limits causal inferences. Although significant associations were found, the temporal ordering of behavior and belief cannot be determined [60, 61]. Fourth, cultural context plays a role: the Turkish media and health systems may not be representative of global conditions, and caution is required in generalizing the results internationally [62].

These limitations point to several future research directions. Longitudinal or experimental designs would help determine causal relationships and explore how media literacy interventions impact health literacy over time [63, 64]. It would also be useful to examine how peer-to-peer communication evolves in health information ecosystems and whether trained individuals influence group behaviors. Mixed-methods research could add qualitative depth to understand motivational drivers of media engagement [65].

Policy and educational implications are substantial. Strengthening media literacy—particularly the evaluation and communication dimensions—can empower individuals to make healthier decisions and resist misinformation [66]. Public health strategies should go beyond simply improving access and focus on interpretative and participatory competencies. Training should emphasize the implications of sharing unverified content and promote digital citizenship in health communication [67, 68]. In this regard, integrating digital health literacy frameworks into national curricula could help develop resilient information consumers, especially in vulnerable populations [69, 70]. Such frameworks not only improve individual outcomes but also contribute to community-level preparedness against health misinformation [71].

The results also highlight the potential of using digital platforms for peer education. Individuals with strong media communication skills can act as health ambassadors within their networks, shaping norms and encouraging evidence-based behavior. This"social contagion"effect offers an opportunity to magnify the reach of health promotion efforts [72, 73].

Finally, health literacy must be viewed not as a static skill but as a dynamic competency shaped by evolving media environments. As platforms become more algorithm-driven and content more personalized, media literacy education must adapt to emphasize awareness of biases, platform mechanics, and source verification [74]. The intersection of media and health will only become more significant in the coming years, necessitating proactive, research-informed strategies [75, 76]. Institutions must remain agile, continuously updating both theory and practice to meet the demands of a rapidly shifting digital health landscape [64, 77].

Conclusions

This study provides empirical evidence that media literacy, particularly the dimensions of analysis and evaluation, plays a crucial role in shaping individuals’ health literacy levels. By applying the Generalized Ordered Logit Model, we were able to assess how these media competencies influence health literacy across varying outcome categories. The findings indicate that individuals with stronger analytical and evaluative skills are significantly more likely to attain higher levels of health literacy. While the access and communication dimensions also show associations, their predictive strength was comparatively modest. Additionally, trust in media-based health information was positively correlated with health literacy, underscoring the importance of media credibility in public health communication.

From a theoretical standpoint, the study extends the Health Belief Model (HBM) by demonstrating that media literacy influences not only health behavior but also the cognitive precursors of such behavior—namely, health literacy itself. The Media System Dependency Theory (MSDT) is similarly supported, as the results show that reliance on media—especially during periods of uncertainty like health crises—affects how individuals process and internalize health-related content. The integration of both theories offers a more comprehensive lens through which media-health interactions can be understood.

The primary contribution of this study lies in its multidimensional analysis of media literacy and its methodological innovation using GOLM to capture the graded effects of media skills on health literacy levels. This disaggregated approach reveals the unequal influence of subdimensions and supports targeted intervention design, moving beyond one-size-fits-all strategies commonly found in digital health education.

Regarding practical implications, these findings point to the urgent need for integrating media literacy into public health and educational programs in Turkey. Given the country’s high digital penetration, training should particularly focus on strengthening individuals'critical analysis and evaluation skills. Policymakers should support curriculum reforms at both school and community levels, embedding media and health literacy education within lifelong learning frameworks. Simultaneously, public health campaigns must strategically utilize highly trusted media—such as television news programs and authoritative digital platforms—to deliver accurate, accessible, and verifiable health content.

Future research should explore how demographic variables—such as age, education level, and digital access—influence the relationship between media literacy and health literacy. Longitudinal and experimental designs are especially needed to establish causal mechanisms and assess the long-term impact of media literacy interventions. Comparative cross-cultural studies could further illuminate how socio-cultural and technological contexts mediate these dynamics, offering a richer global understanding of media-health interrelations.

Limitations

This study presents several limitations that warrant consideration. First, the use of convenience sampling via voluntary participation in an online survey introduces a degree of selection bias. Although the sample size was statistically adequate, the non-probability sampling design limits the generalizability of the findings, particularly across socio-demographically diverse populations. Future studies should adopt stratified or probability-based sampling methods to enhance external validity and representativeness.

Second, while socio-demographic variables such as age and education level were initially included based on prior research linking them to health literacy, they did not yield statistically significant results in this sample (p > 0.05). These variables were therefore excluded from the final model to maintain parsimony. However, given their established relevance, future research should re-examine these factors using larger or more heterogeneous samples, or alternative modeling approaches, to fully explore their predictive potential.

Third, the reliance on self-reported data introduces possible response bias, especially concerning subjective constructs like media literacy. Participants may have unintentionally over- or underreported their competencies, affecting the accuracy of the measurements. While validated scales were used, the limitations of self-assessment in behavioral and cognitive domains remain a concern.

Fourth, the cross-sectional design limits the ability to draw causal inferences. Although associations between media literacy subdimensions and health literacy were established, the directionality of these relationships remains unclear. Longitudinal studies are recommended to determine whether improvements in media literacy lead to measurable gains in health literacy over time.

Finally, the sociocultural context of Turkey, with its unique media ecosystem and health communication patterns, may limit the transferability of results to other settings. Cultural norms, institutional trust, and media access vary widely across countries and may influence both media use and health literacy in distinct ways. Future comparative research across diverse cultural contexts is necessary to validate and extend the current findings.

Authors’ contributions

Conception and design of the work: H.M., G.B., G.O., and Y.G. Survey design and implementation: H.M., G.B., G.O., and Y.G. Data analysis: G.B. Interpretation of data for the work: G.B. and H.M. Drafting the work: H.M., G.O. and Y.G. Revising it critically: all authors. Final approval of the version to be published: all authors.

Funding

There is no financial support from any institution.

Data availability

The survey data can be shared on reasonable request to the corresponding author.

Declarations

Ethics approval and consent to participate

Ethical approval for this study was obtained from the Scientific Research and Publication Ethics Board for Social Sciences and Humanities at Istanbul Beykent University (approval date: July 8, 2024; approval number: 586). Informed consent was obtained from all participants prior to data collection. All procedures were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.