Skip to Content
MDPIMDPI

Advanced

Education SciencesEducation Sciences
PDF
  • Article
  • Open Access

28 October 2026

18 Pages

The Subjective Experience of Reading: State Anxiety Measured After a Reading Task, General Self-Evaluation of Reading Skills and Reading Skill Measurement

,
and
German Institute for Adult Education—Leibniz Centre for Lifelong Learning, 53175 Bonn, Germany
*
Author to whom correspondence should be addressed.
Educ. Sci.2026, 16(11), 1811;https://doi.org/10.3390/educsci16111811
This article belongs to the Special Issue The Many Faces of Adult Literacy Across Contexts, Cultures, and Modalities
Version Notes
Order Reprints

Abstract

This study investigated reading task-related state anxiety and general self-evaluation of reading skills among adults with low reading comprehension, comparing the results with those of participants with higher reading comprehension. All participants attended the lower-level educational system in Germany (N = 131). Standardized measures of reading comprehension, reading speed, and reading task-related state anxiety were collected. Additionally, participants provided a general evaluation of their level of reading comprehension and reading speed. After screening and due to missing data, the sample sizes varied across analyses (112 to 123). Participants with low reading comprehension showed somewhat lower general self-evaluations of reading skills when compared to typical readers. However, linear regression analysis indicated no group differences in reading task-related state anxiety, even after controlling for age and gender. Finally, we found moderate positive correlations between reading skills and general self-evaluations of these skills. However, reading task-related state anxiety was not significantly correlated with either. Taken together, in this sample, individuals with low reading comprehension did not show overly negative general self-evaluations of reading skills or associated reading task-related state anxiety. This suggests heterogeneity in affective and emotional dimensions related to reading that may be influenced by factors other than mere reading level.

1. Introduction

Literacy education remains a critical topic across the lifespan, from early childhood development into adulthood (Mullis et al., 2023; OECD, 2024, 2026). This was recently confirmed by the high proportion of adults with low literacy skills reported among the working-age population of the OECD countries participating in the Program for the International Assessment of Adult Competencies survey (PIAAC, OECD, 2024). Proficiency in reading not only underpins academic and professional success but can also influence self-concept and emotional well-being (Aro et al., 2025). One relevant emotional aspect is anxiety, characterized by physical and cognitive symptoms of heightened distress and arousal (Francis et al., 2019). While many studies support the assumption that general anxiety and/or specific types of anxiety are prevalent among struggling readers (e.g., the meta-analysis by Francis et al., 2019), others challenge this claim (e.g., Bonifacci et al., 2008; Lindeblad et al., 2016). These inconsistencies suggest complex relations between reading skills and anxiety, with self-concept proposed as a mediator in this relationship (G. McArthur, 2022).

1.1. Conceptualizing the Relationship Between Self-Concept, Anxiety, and Reading Skills

Self-concept is defined as entailing all perceptions an individual has formed about themselves (Marsh & Shavelson, 1985). Beyond this general self-concept, reading self-concept specifies how individuals perceive their own reading skills (Chapman & Tunmer, 1997). Accordingly, we can tap into reading self-concept by asking individuals to self-evaluate their reading competence (Edwards et al., 2023). There are several types of anxiety. One of these is state anxiety, which is defined as a temporary negative emotional, cognitive, physiological and behavioral reaction associated with a specific situation (Hutchins & Young, 2018). Therefore, this type of anxiety is of particular interest when relating to a specific context, such as an academic one. Accordingly, state anxiety may be elicited during reading-related situations. The term reading anxiety has been used in research to define anxiety that is specific to the reading domain and involves state anxiety responses when engaging in reading (Jalongo & Hirsh, 2010; Piccolo et al., 2017).
As proposed by the Interpretation Account (Ramirez et al., 2018, originally developed in the context of math anxiety, often applied to reading) and the Poor Reading-Anxiety (PRAX) model (G. McArthur, 2022), the relationship between skills and anxiety is mediated by self-concept. Both theoretical frameworks assume that poor reading might lead to negative experiences related to reading, which become attached to an individual’s reading self-concept. A negative reading self-concept may in turn lead to anxiety, particularly reading anxiety. The mediating role of reading self-concept, as described by these frameworks, has been supported by a study on Finnish adolescents representing a general school sample: Torppa et al. (2024) found indirect longitudinal associations between reading skills and well-being measures via reading self-concept. Both the Interpretation Account (Ramirez et al., 2018) and the PRAX model (G. McArthur, 2022) leave room for the possibility of protective factors that prevent the development of a negative reading self-concept and consequently anxiety in individuals with reading difficulties. It is important to note that research on second language anxiety has also emphasized the potential role of underlying language aptitude and prior achievement, in addition to current language achievement, in the development of later anxiety (Sparks, 2025). In the present study, reading skills are conceptualized as an indicator of current language achievement in the reading domain.

1.2. Empirical Evidence for the Relationship Between Self-Concept, Anxiety, and Reading Skills

Studies exploring the relation between state anxiety in the context of reading and reading achievement have shown a negative association. For example, Macdonald et al. (2021) and Fishstrom et al. (2024) have demonstrated relations between state anxiety measured in the domain of reading and reading achievement among children with reading difficulties. Orbach et al. (2026) have also shown such negative relations among children with a broad range of skills; however, these were restricted to timed reading tasks. A recent meta-analysis by Johnson et al. (2026) including 64 studies reported a significant moderate correlation between reading anxiety and reading achievement (r = −0.30, p < 0.001), albeit with considerable heterogeneity across effect sizes. While most studies relating to reading achievement traditionally examine children, the age range included in the meta-analysis was 6 to 28 years, and age was not a significant moderator. A study by Wang et al. (2023) on Chinese-speaking students from elementary school to 9th grade revealed similar associations between state anxiety in relation to a reading task and reading comprehension. Interestingly, however, Meer et al. (2016) have reported increased physical arousal measured in relation to a reading task among adults with dyslexia compared to typical readers, but with no differences between the groups in self-reported anxiety symptoms.
Notably, the relationship between another type of anxiety, termed second language anxiety, and language achievement has been extensively studied. Findings suggest significant links between language achievement, self-perceived competence, and this form of anxiety that is specific to the second language domain (Teimouri et al., 2019). Additionally, causal investigations have shown that second language achievement is likely to precede and predict this type of anxiety (e.g., Alamer et al., 2025; Alamer & Lee, 2024). In this context, Saito et al. (1999) proposed that reading anxiety is distinguishable from more general second language anxiety, which has primarily been linked to the oral domain. However, both second language anxiety and reading anxiety in a second language are correlated and the associations between second language anxiety and achievement seem to extend to the reading contexts (Li, 2022; Teimouri et al., 2019).
When it comes to reading skills and reading self-concept, a meta-analysis by G. M. McArthur et al. (2020) provided support for the association between the two. About 85% of the studies were based on English-speaking children and adolescents, with the rest also including adults. The same results were found for German-speaking (Roeschl-Heils et al., 2003) and Greek-speaking children (Polychroni et al., 2024).
Studies investigating how reading self-concept relates to anxiety indicate that lower reading self-concept predicts both higher reading anxiety and general anxiety levels in children (e.g., Katzir et al., 2018; Polychroni et al., 2024). Similar results were found in a meta-analysis by Brumariu et al. (2023) for the relationship between academic self-concept and anxiety across the lifespan, from elementary school into adulthood, with only a few studies focusing on the latter. In their analysis, they did not discriminate between different subdomains of academic self-concept, such as language and reading, and types of anxiety due to limited numbers of studies per category.

1.3. The Current Study

The aim of this study is to extend the findings on the relations between certain aspects of self-concept, anxiety, and reading skills to a variety of populations and educational contexts. Most previous studies have focused on younger participants. However, the age of participants could influence the extent to which self-concept, anxiety, and skills relate to one another (Brumariu et al., 2023). In the present study, we focus on an understudied population in the context of literacy education: young adults who are in a critical transition phase from formal schooling into adulthood and the job market, where they are potentially vulnerable to long-lasting adverse effects of poor reading proficiency (Torppa et al., 2024). Research has traditionally focused on the academic skills of adults with low literacy skills (e.g., Perin, 2020). However, it is necessary to also understand the associated affective and emotional dimensions. Examining these dimensions in relation to measurable reading skills in the German educational context is of particular interest. The German school system separates students with higher and lower academic potential into different secondary education tracks as early as 5th grade, earlier than in many other countries. This results in more homogeneous learning groups and the adjustment of academic goals, which may have consequences for academic self-expectations. Additionally, findings from English-language research may not be directly transferable due to systematic differences in orthography. German orthography is more transparent than English, making reading acquisition easier in terms of reading accuracy (Share, 2008). These factors may have distinct long-term effects on how individuals experience and cope with academic and reading difficulties. Hence, understanding emotional aspects associated with reading in a variety of populations of readers is important in order to identify potential psychological barriers that individuals may face in different language and educational contexts.
Notably, this study was part of a broader research project examining mainly the language skills of adults with low literacy (Bar-Kochva et al., 2026). As the project included a long testing battery which had to be administered within a limited testing time, we had to rely on quick questionnaires addressing emotional aspects. Consequently, this study does not capture reading self-concept and state anxiety in the domain of reading in a broad and nuanced sense and cannot directly test the PRAX model (G. McArthur, 2022) or the Interpretation Account (Ramirez et al., 2018). Instead, the study addresses certain aspects of these constructs that are considered important for the population examined. These are general self-evaluations of reading comprehension and reading speed, and state anxiety in relation to a multiple-choice reading comprehension task. The rationale for focusing on general self-evaluation specifically of reading comprehension and reading speed is that reading comprehension has been identified as a significant challenge for a large proportion of the working-age population, including young adults (OECD, 2024). Furthermore, reading speed is a measure that is more sensitive to variance between readers than reading accuracy, particularly beyond the early stages of reading acquisition and across different orthographies (Share, 2008). The rationale for focusing on state anxiety in relation to a multiple-choice text comprehension task is that this type of task is often used in adult education settings and can be employed to assess functional literacy (OECD, 2024), which is a crucial skill for adults’ daily lives.
We investigate the following research questions: (1) Do general self-evaluations of reading skills and state anxiety measured following a reading task differ across groups of German-speaking young adults representing varying levels of reading comprehension? (2) What is the relationship between reading skills, reading task-related state anxiety, and general self-evaluations of reading skills in a sample of German-speaking young adults representing a broad range of reading comprehension?
Based on the previous literature (e.g., Johnson et al., 2026; Katzir et al., 2018; G. M. McArthur et al., 2020) and while considering the theoretical frameworks (Interpretation Account, Ramirez et al., 2018, and PRAX model, G. McArthur, 2022), we expect to find a difference in general self-evaluations of reading skills and in reading task-related state anxiety between participants with low reading comprehension and typical readers. We also expect reading skills to be positively correlated with general self-evaluations of these skills, and both reading skills and general self-evaluations to be negatively correlated with reading task-related state anxiety.

2. Materials and Methods

2.1. Participants

Overall, 131 young adults were recruited from vocational schools and adult education centers. In Germany, vocational schools offer classroom instruction as part of the dual education and training system as well as preparatory classes before starting vocational training. The latter classes are offered to students who lack foundational skills that would allow them to successfully complete vocational training. Adult education centers provide basic education as well as courses to obtain a school-leaving certificate for adults who left school without such a certificate. As both settings offer a variety of programs, including those addressing young adults with low basic academic skills, we expected to find participants with a broad range of literacy skills.
Participants were included in the analysis if they had been enrolled in a German-speaking school since the first grade, did not report any special neurological conditions, and had an IQ estimation score (ZVT, Oswald, 2016) within the norm (cut-off below two standard deviations from the mean according to WHO, 2018). The ZVT was used to estimate IQ based on its validation against established full IQ measures (Oswald, 2016). The IQ estimation score can be extrapolated from the raw score according to conversion guidelines and norms provided in the test manual. As a full IQ screening was not possible within the restricted testing time, the screening of participants is based only on this rough estimation of IQ. These inclusion criteria were applied to verify that potential reading difficulties were not due to late acquisition of the German language or due to low general cognitive ability. During screening, eight participants were excluded from analyses: two of whom had joined a German-speaking school later than first grade, four reported autism diagnoses, one had an IQ estimation score below the norm, and one did not complete the ZVT; therefore, general cognitive ability could not be estimated for this participant. This left data from 123 participants (aged 15–31 years, M = 18.61 years, SD = 2.79, 90 male, 31 female, 2 gender not indicated, no cases of diverse) in the sample.
Notably, while all 123 participants completed the reading measures, only 121 completed the reading task-related state anxiety questionnaire, 119 participants responded to the reading comprehension general self-evaluation question, and 118 responded to the reading speed general self-evaluation question. Additionally, there was information missing regarding the gender of two participants and the age of two participants. The missing data were omitted in the respective analyses. All participants, as well as a legal guardian if the participant was underage, gave written informed consent prior to their participation. The study was approved by the institutional ethics committee (registered under DIE-LEK-2021–004).

2.2. Materials

To assess reading skills, we obtained measures of reading comprehension (RC) and reading speed (RS) from a standardized reading test designed for 5th to 12th graders (LGVT 5–12+, Schneider et al., 2017). In this pen-and-paper test, participants were presented with a cloze-format text and asked to choose one of three words to fill in the respective blanks. They had a time limit of six minutes to complete as many blanks as possible. The measure of RC represents comprehension accuracy. The measure of RS refers to the total number of words read within the time limit. Test–retest reliability coefficients range from 0.82 to 0.89. Although this test was not specifically designed for adults, it was chosen for this study because it allows capturing a broad range of reading levels using the same test material, and its content is not restricted to the world of children.
To measure reading task-related state anxiety, we used a standardized state anxiety questionnaire (STADI, Laux et al., 2013). The STADI consists of two forms, one for state and one for trait emotion. Only the state emotion form was administered in a pen-and-paper format. It is a 20-item questionnaire with each item relating either to symptoms of state anxiety or depression. We only used the subscore for anxiety for analyses. The test’s internal consistency for state anxiety is 0.90.
To capture participants’ immediate emotional response regarding a reading task, the questionnaire was administered directly after they completed a multiple-choice reading comprehension task (originally developed for the broader project, Bar-Kochva et al., 2026). Two parallel versions of tasks were presented to participants in a counterbalanced manner (i.e., each participant completed only one task). The tasks were matched in terms of content, length, and readability level. The content of both texts was informative, relating to real-life contexts of young adults, specifically to terms and conditions of service providers; one text was about a gym membership and the other about a streaming service. The texts were created with the help of Artificial Intelligence (ChatGPT-4, OpenAI, 2024) and further adapted for clarity by the authors. The prompt presented to the AI tool appears in the Supplementary Material. The resulting texts were one A4 page long (253 and 256 words respectively). Readability, represented by language complexity, was at an intermediate level according to the readability LIX (https://www.psychometrica.de/lix.html, accessed on 8 October 2024) and Flesch indices (https://www.fleschindex.de/, accessed on 8 October 2024, http://www.leichtlesbar.ch/html/fleschformel.html, accessed on 8 October 2024). Each text also included 11 multiple-choice questions examining different levels of reading comprehension (the surface level representation, the propositional representation, and the mental representation level of understanding; see Richter & Schnotz, 2018). The questions of the two versions of tasks were matched in terms of difficulty level according to the criteria suggested by Hutzler et al. (2014) to determine the number of steps needed to answer a text comprehension question. In addition, results of these texts loaded on the same factor as the results of the standardized LGVT reading test (0.837), thereby suggesting that the multiple-choice task addressed the same construct of reading comprehension. The texts and multiple-choice questions can be accessed online (https://alpha-material.de/Record/1981366180, accessed on 10 August 2026). The multiple-choice reading comprehension task was not analyzed in the present study and its results are therefore not presented here. It was only used to create the context of a reading task in relation to which state anxiety could be measured. This approach is in line with many other studies investigating the relationship between reading and anxiety (e.g., Carroll & Iles, 2006; Chow et al., 2021; Orbach et al., 2026; Wang et al., 2023). It is based on the assumption that state anxiety closely reflects the emotions experienced during a specific task that involves reading (Orbach et al., 2026; Robinson & Clore, 2002). However, it cannot be completely ruled out that a state anxiety score after a reading task may also be influenced by anxiety related to the testing situation.
Finally, participants were asked to fill out a general self-evaluation questionnaire. The full questionnaire in the related study (Bar-Kochva et al., 2026) consisted of six questions evaluating reading comprehension, reading speed, spelling, writing, listening comprehension and speaking. Only the reading-related questions were analyzed in this study, in line with the study’s focus on reading skills. Notably, within the scope of our larger study (Bar-Kochva et al., 2026) and in order to minimize fatigue in our participants, we needed a short questionnaire. Accordingly, in designing the questionnaire, we drew on measures of self-perceived reading competence used in large-scale OECD surveys such as PISA 2018 (OECD, 2019). Participants were asked to evaluate their overall reading comprehension on a five-point response scale (How well do you understand what you read?; 1 = Not well at all; 2 = Not well; 3 = Well; 4 = Very well; 5 = I don’t know). They were also asked to evaluate their overall reading speed on a six-point response scale (How quickly can you read?; 1 = Very slowly; 2 = Slowly; 3 = At a normal pace; 4 = Quickly; 5 = Very quickly; 6 = I don’t know). Consequently, we hereby relate to a restricted scope of reading self-concept rather than the broader multidimensional construct by assessing participants’ general perception of their reading comprehension and reading speed.

2.3. Procedure

Participants were individually tested in quiet rooms in their vocational schools or adult education centers. The total testing time of all tasks included in the broader project (Bar-Kochva et al., 2026) was between two and three hours for each participant. This testing time also included tasks not analyzed in the present study. The tests were administered by trained research assistants in the fields of psychology and education and the authors of this study. Each session began with an oral interview in which participants were asked for background information (age, gender, language, and educational background, as well as about previous diagnoses that might have affected learning). The reading comprehension task was administered next (LGVT 5–12+, Schneider et al., 2017), followed by six language and cognitive tasks (not analyzed here) administered in random order. After a break (lasting about 20 minutes), the second part of the testing session began, which included the multiple-choice reading task followed immediately by the STADI questionnaire (Laux et al., 2013). Four additional language and cognitive tests (not analyzed here) were administered thereafter in random order. The general self-evaluation of reading skills was administered last. To minimize fatigue and to ensure that participants were comfortable, breaks were offered at regular intervals (in addition to the long break between the first and second part of the testing session). All test administrators were instructed to pay close attention to signs of stress and fatigue and to offer breaks accordingly.

3. Results

3.1. Data Preparation and Analysis

Given our motivation to better understand emotional aspects of reading among adults with low reading comprehension, we divided the participants into three groups defined by their RC level instead of treating reading skills as a continuous variable. To this end, the standardized LGVT 5–12+ (Schneider et al., 2017) was used, relating to pre-defined cut-off points of RC level with the following rationale: the first group consisted of participants who performed at or below 8th-grade RC level (n = 46). This cut-off is often used to define adults who require adult basic education in the domain of literacy (e.g., Greenberg & Feinberg, 2019). The second group consisted of participants who reached 9th-grade RC level but were below 10th-grade level (n = 18). 10th-grade level RC was considered a threshold of typical RC, considering that the lower tracks of the German educational system continue until 9th or 10th grade. Norms for 11th and 12th graders only relate to students who attend the higher level of the German educational system (high school and matriculation). Therefore, the third group consisted of those individuals whose RC was at 10th-grade level or above (n = 59) and were considered typical readers. Table 1 shows the mean percentile ranks for RC and RS by group, using the 10th-grade RC and RS levels, respectively, as reference. Previous diagnoses of reading and spelling difficulties were self-reported by 26.1% of the individuals in Group 1, 33.3% in Group 2, and 18.6% in Group 3.
Table 1. Mean percentile ranks (10th-grade level) for reading skills, RC and RS, with standard deviations in brackets by group.
For both general self-evaluations of RC and RS, there were no cases where participants indicated “I don’t know.” This answer option was therefore disregarded in the analysis. This left a four-point scale for RC general self-evaluation (How well do you understand what you read?; 1 = Not well at all; 2 = Not well; 3 = Well; 4 = Very well) and a five-point scale for RS general self-evaluation (How quickly can you read?; 1 = Very slowly; 2 = Slowly; 3 = At a normal pace; 4 = Quickly; 5 = Very quickly). To facilitate interpretation of both general self-evaluations alongside one another, the general self-evaluation of RS was linearly rescaled from a five-point response scale to a scale from 1 to 4 while retaining all five response categories.
The data were analyzed using the statistical computing software R (version 4.5.1, R Core Team, 2025). To answer the research questions, we chose analyses that would allow controlling for age- and gender-related differences. As indicated previously, the classes from which participants were recruited included students with a broad age range. Additionally, the representation of male participants was higher than that of females, which is in line with national statistics relating to vocational educational settings (German Federal Statistical Office, n.d., retrieved 4 August 2026). Accordingly, to assess group differences in general self-evaluations of reading skills (ordinal variables), cumulative ordinal logistic regression models with a logit link function were estimated using the ordinal package (Christensen, 2025). RC and RS general self-evaluations were specified as the respective ordinal outcome variables and group, age, and gender as predictors. The proportional-odds assumption was evaluated using the nominal-effects test. As the test indicated a violation for age, sensitivity analyses were conducted by fitting partial proportional-odds models, allowing the effect of age to vary across outcome thresholds (VGAM package, Yee, 2015). The original proportional-odds models were retained for the primary analyses, and the robustness of the group effect was evaluated by comparing estimates across models. Model adequacy was assessed by comparing each full model with the corresponding intercept-only model using likelihood-ratio tests. Model fit was assessed using McFadden’s pseudo-R2. Estimated threshold parameters and coefficients were obtained from the full models. Overall group differences were evaluated by comparing the full model including group, age, and gender to a reduced model without group. If significant, pairwise comparisons between groups were conducted using the emmeans package (Lenth & Piaskowski, 2026) with Tukey adjustment. Pairwise comparisons were expressed as odds ratios. To examine group differences in reading task-related state anxiety levels (continuous variable), a multiple linear regression model was fitted using the lm() function. Reading task-related state anxiety was specified as the continuous outcome variable and group, age, and gender as predictors. Model assumptions were assessed using regression diagnostic plots and the Breusch-Pagan test (lmtest package, Zeileis & Hothorn, 2002). Because the Breusch-Pagan test indicated heteroscedasticity, HC3 robust standard errors, 95% confidence intervals, and p-values were used for inference (sandwich package, Zeileis, 2004; Zeileis et al., 2020). The overall effect of group was assessed using a robust Wald test (car package, Fox & Weisberg, 2019). To further understand the extent to which participants reported elevated reading task-related anxiety, frequency analysis of anxiety level according to the norms by group was carried out. Finally, partial Spearman’s rank correlations from the ppcor package (Kim, 2015) were used to test the second research question, i.e., the associations between reading task-related state anxiety, reading skills, and general self-evaluations of RC and RS while considering the entire sample, and while controlling for age and gender.

3.2. Group Differences in General Self-Evaluation of Reading Skills and Reading Task-Related State Anxiety

The descriptive statistics, including median values for RC and RS general self-evaluations as well as the mean values for reading task-related state anxiety, for all three groups are shown in Table 2. For the ordinal regression analysis, four participants had missing values for RC general self-evaluation, and five participants had missing values for RS general self-evaluation. Additionally, three participants had missing age and/or gender information and were therefore excluded from both analyses. This resulted in effective sample sizes of N = 116 for RC general self-evaluation and N = 115 for RS general self-evaluation. For both RC and RS, the full model was significant (RC: χ2(4) = 14.26, p = 0.007; RS: χ2(4) = 20.69, p < 0.001) with McFadden’s pseudo-R2 = 0.07 for RC and 0.08 for RS. For the RC model, the nominal-effects test indicated evidence of a violation of the proportional-odds assumption for age (χ2(2) = 13.83, p = 0.001). For the RS model, the nominal-effects test did not yield an estimable test statistic for age, while the proportional-odds model itself converged successfully. For both models, no evidence of violations was found for group and gender. Sensitivity analyses using partial proportional-odds models, which allowed the effect of age to vary across thresholds, yielded absolute group-effect estimates and p-values similar to those of the original models (RC: Group 2 vs. 1: β = −0.49, SE = 0.59, 95% CI [−1.64, 0.67], p = 0.41; Group 3 vs. 1: β = −1.57, SE = 0.46, 95% CI [−2.48, −0.67], p = 0.001; RS: Group 2 vs. 1: β = −0.91, SE = 0.55, 95% CI [−1.98, 0.16], p = 0.10; Group 3 vs. 1: β = −1.73, SE = 0.44, 95% CI [−2.58, −0.88], p < 0.001). The original model estimates, including the covariates age and gender and estimated thresholds, are presented in Table 3. The analyses revealed an overall effect of group for RC and RS general self-evaluations after controlling for age and gender (RC original model: χ2(2) = 13.31, p = 0.001; RC partial model: χ2(2) = 13.20, p = 0.001; RS original model: χ2(2) = 16.96, p < 0.001; RS partial model: χ2(2) = 17.27, p < 0.001). Pairwise comparisons indicated that Group 1 (low reading comprehension skills) and Group 3 (typical reading skills) differed in their general self-evaluations (RC: β = −1.57, SE = 0.46, OR = 0.21, 95% CI [0.07, 0.60], p = 0.002; RS: β = −1.71, SE = 0.43, OR = 0.18, 95% CI [0.07, 0.50], p < 0.001). Adults with typical reading skills had higher odds of reporting higher reading general self-evaluations compared to individuals with low reading comprehension skills. Group 2 neither differed from Group 1 nor from Group 3 in both RC and RS general self-evaluations. Due to the small number of participants in Group 2, estimates involving this group may be associated with greater uncertainty. Interestingly, participants in Group 1, whose reading level was significantly below that expected for their educational level, still evaluated their skills as rather positive on average.
Table 2. Descriptive statistics including median of RC and RS general self-evaluations with interquartile range in brackets, and mean reading task-related state anxiety scores with standard deviations in brackets by group.
Table 3. Results of the ordinal regression models for RC and RS general self-evaluations, including regression coefficients and estimated threshold parameters, and coefficients of the linear regression model for reading task-related state anxiety.
For the linear regression analysis, two participants with missing values for the state anxiety measure and three participants with missing age and/or gender specifications resulted in an effective sample size of N = 118. The model explained 9% of the variance in reading task-related state anxiety (R2 = 0.09, adjusted R2 = 0.06). Inspection of the regression diagnostics indicated minor deviation from normality of the residuals, while the assumption of linearity was met. As the Breusch–Pagan test indicated significant heteroscedasticity (p < 0.001), HC3 robust standard errors, 95% confidence intervals, and p-values were used. The model estimates are presented in Table 3. The analysis revealed no group differences for reading task-related state anxiety (F(2, 113) = 0.03, p = 0.971). All groups’ mean anxiety scores were within the normal range by age and across genders. However, we observed a noteworthy variance within the reading task-related state anxiety score across the sample and within the groups.

3.3. Frequency Analysis of Reading Task-Related State Anxiety Compared to the Norms by Group

To further understand the results concerning reading task-related state anxiety, we carried out a frequency analysis of the proportion of participants within each group by level of reading task-related state anxiety compared to the norms. According to the norms, t-values < 30 are considered well below average, t-values between 30 and 39 are below average, and t-values between 40 and 60 indicate an average state anxiety level. Particularly relevant for this study are t-values between 61 and 70 (above average), and t-values above 70 (well above average), indicating high or very high reading task-related state anxiety. Results of the frequency analysis are shown in Table 4. Due to some cases with null to very low frequencies, the data were only descriptively inspected without further statistical analysis. In all groups, most participants showed an average reading task-related state anxiety level. In all groups, smaller proportions indicated low, high, and very high levels of reading task-related state anxiety. In neither group were values well below average. It appears, then, that the means represent well the level of reading task-related state anxiety of most participants within each group.
Table 4. Absolute number and percentage of participants per reading task-related state anxiety level according to the norms by group.

3.4. Relationship Between Reading Skills, Reading Task-Related State Anxiety and General Self-Evaluations in the Large Sample

Finally, we were interested in possible correlations between reading general self-evaluations and the corresponding reading scores, reading task-related state anxiety and reading skills (RC and RS), and reading task-related state anxiety and general self-evaluations (RC and RS). Eight participants did not complete the STADI and/or the self-evaluation questionnaire and three were missing age and/or gender specifications. These eleven cases with missing values were excluded listwise for partial correlations. The effective sample size was therefore N = 112. The correlations are shown in Table 5. Moderate correlations emerged between reading skills and the respective general self-evaluations (RC: rS = 0.46, p < 0.001; RS: rS = 0.47, p < 0.001). They indicate that higher reading skills are generally associated with higher general self-evaluations of these skills. However, in this sample, reading task-related state anxiety was not significantly correlated with reading skills or general self-evaluations. These results are convergent with the findings of group differences in general self-evaluations of reading skills but not in reading task-related state anxiety and add information on the relationship between reading task-related state anxiety and general self-evaluations. In the present sample, neither analysis provided evidence of an association between reading task-related state anxiety and reading level, regardless of whether it was analyzed as a continuous variable across the entire sample or categorized according to critical thresholds of reading level.
Table 5. Correlation coefficients for correlations between reading skills, general self-evaluations of reading skills, and reading task-related state anxiety for the large sample. *** p < 0.001.

4. Discussion

This study investigated reading task-related state anxiety and general self-evaluations of reading skills among young adults with varying levels of reading skills. To explore the first research question, we compared reading general self-evaluations for RC and RS between participants grouped according to their reading comprehension level. We found that the low reading comprehension group presented lower general self-evaluations of reading skills compared to the group with a typical reading level. This finding aligns with previous research demonstrating a more negative reading self-concept among individuals with reading difficulties across countries and a broad age range: in primarily English-speaking children, adolescents, and adults (G. M. McArthur et al., 2020), as well as in German-speaking (Roeschl-Heils et al., 2003) and Greek-speaking children (Polychroni et al., 2024). However, we noticed that participants in the low reading comprehension group still evaluated their reading skills as rather positive on average despite low performance in measures of RC and RS. Similar results have been found by Lindeblad et al. (2016) and Kasperski et al. (2016), among others. This observation might suggest a gap between skills and general self-evaluations of these skills, which has previously been interpreted as a systematic overestimation of skills among individuals with difficulties in reading (e.g., Kwon & Linderholm, 2014). However, based on the current results, it is impossible to conclude whether the participants overestimated their reading skills, as they were not specifically asked about their evaluation of performance in the specific reading comprehension task administered in this study. Instead, they were asked to provide a general evaluation of their reading comprehension and reading speed without any specific task reference. The general reading self-evaluations of Group 2, in accordance with their reading level being right in between the individuals with low reading comprehension and the typical readers, did not differ from the other two groups.
Our analysis revealed no group differences in reading task-related state anxiety, which is in line with other studies indicating that individuals with reading difficulties do not exhibit higher symptoms of anxiety compared to those with typical reading skills (e.g., Bonifacci et al., 2008; Lindeblad et al., 2016; Polychroni et al., 2024). However, our findings conflict with meta-analyses by Francis et al. (2019) on general anxiety and Johnson et al. (2026) on reading anxiety, as well as studies showing higher general anxiety in German-speaking children with reading difficulties (Novita et al., 2019; Visser et al., 2020). There are several possible explanations for these diverging findings. The first one being methodological differences across studies, such as variations in the classification of the samples. While the above-mentioned studies, including the ones reviewed in the meta-analyses, relied on previous diagnoses of reading difficulties or measures of word reading, participants in our study were divided according to critical thresholds of text reading competence guided by expected norms, with all groups including individuals diagnosed with reading difficulties. Text reading, as opposed to single-word reading, more closely mirrors everyday demands in school and private life, where compensation strategies such as using context clues can be applied. Perceived competence and reading-related emotions may be shaped by these situated experiences.
Methodological differences between studies resulting in varying effect sizes can also be found regarding measures of anxiety, as discussed by Johnson et al. (2026). While the authors note that most studies they reviewed used self-report measures of anxiety, others assessed anxiety through parent-reported scales (e.g., Visser et al., 2020). Another difference is in the actual measure used, whether these are specific reading anxiety scales, other anxiety scales (e.g., general or math anxiety) adapted to reading, or self-developed measures, which affected effect sizes as analyzed by Johnson et al. (2026). There are some studies that used a state anxiety scale in relation to a reading task (e.g., Wang et al., 2023), similar to what was done in our study, whose results parallel the findings of an association between anxiety and reading comprehension established in the meta-analyses. It therefore seems unlikely that our divergent findings can be attributed to the choice of anxiety measure alone. Findings on the relationship between reading skills and anxiety might also differ depending on whether general anxiety or reading-specific anxiety is assessed (Macdonald et al., 2021).
Further longitudinal research may clarify whether the relationship between anxiety and reading difficulties might change with age, i.e., be detectable at younger ages but not as children grow older. In their meta-analysis, Brumariu et al. (2023) found that effect sizes were in fact stronger for samples of younger children as compared to adolescents or adults. As noted earlier, our sample differed significantly from the populations studied in the reviewed literature, not only in terms of age but also regarding their language and educational background. Specifically, participants in the current study were considerably older than those included in the majority of previous studies. Additionally, as compared to the predominantly English-language literature, we investigated German speakers whose reading development, including affective and emotional dimensions, might have been influenced by the greater orthographic transparency of German and the tracking system in school. This, in addition to the fact that we used a measure of silent reading, may have reduced any reading-related anxiety symptoms.
Overall, the results of the present study provide support for our hypothesis that participants with low reading comprehension differ from typical readers in regard to their general self-evaluation of skills. The low reading comprehension group had lower general self-evaluations of both RC and RS skills on average. However, we do not find support for our hypothesis that participants with low reading comprehension would show higher state anxiety following a reading task. Instead, all groups in our sample showed mean state anxiety t-values within the norm. The frequency analyses suggested that the means represented well the level of reading task-related state anxiety of most participants within each group. Considering the previously reported heterogeneity in effect sizes relating to the relations between anxiety and reading skills (Francis et al., 2019; Johnson et al., 2026), the state of knowledge suggests that these relations are sensitive to specific testing conditions. Therefore, adults with low literacy skills may not experience symptoms of anxiety in any given reading situation. As the current study examined reading task-related anxiety under limited conditions (silent reading of an informative text in an individual testing of an experimental session), the exact conditions in which anxiety symptoms may be elicited in this group remain to be explored.
Finally, to answer the second research question, we examined the relationships between the three variables in focus. A correlation analysis including the entire sample, i.e., the sample representing a broad range of reading skills, revealed moderate correlations between reading skills and general self-evaluations of these skills. These results are in line with the meta-analysis by G. M. McArthur et al. (2020) and studies by Roeschl-Heils et al. (2003) and Polychroni et al. (2024). However, for the present sample, our findings do not support our hypothesis of an association between reading general self-evaluations and state anxiety following a reading task. One possible explanation for these results is that clearer correlation patterns to reading task-related state anxiety may only emerge when general self-evaluations of reading skills fall below a certain threshold. Our hypotheses regarding associations in the present sample can therefore only be confirmed for the positive correlation between reading skills and general self-evaluations of these skills, but not for the negative correlations between reading skills and reading task-related state anxiety, or between reading general self-evaluations and reading task-related state anxiety.

5. Limitations and Future Directions

There are limitations of our study that need to be considered when interpreting the results. The self-evaluation questions relate to a restricted scope of reading self-concept in general without paying attention to specific contexts. However, how individuals rate their own skills may vary depending on the context of reading. A more fine-grained, context-dependent self-evaluation of reading skills might reveal more nuanced insights. This could include reading in specific settings (test situation vs. regular classroom vs. at home), reading of specific text types (children’s book vs. manual vs. newspaper), modalities of reading (aloud vs. silently), as well as different components of reading (speed vs. comprehension). By doing so, we might be able to provide a clearer picture of subjective evaluations of strengths or weaknesses and set them in relation to potential corresponding fears and measurable skills.
Two aspects regarding the anxiety measures also warrant further consideration. First, as the measure relied on self-report, it may be susceptible to social desirability bias (a limitation that also applies to the general self-evaluation measure). Future studies might therefore want to incorporate more objective measures of anxiety such as physiological ones (e.g., Meer et al., 2016) alongside self-reports. Secondly, our measure of state anxiety does not allow separating anxiety that is evoked by the requirement to read from anxiety evoked by the testing situation. Although participants were instructed to complete the state anxiety questionnaire in relation to and following a reading task, the possibility that other state anxiety factors may have been involved cannot be ruled out. That said, reading anxiety and test anxiety have repeatedly been found to be highly correlated with one another (e.g., Fishstrom et al., 2024). Additionally, Carroll and Iles (2006) found that differences in state anxiety in relation to a reading task between dyslexic and non-dyslexic university students were no longer significant once reading ability was controlled for. State anxiety in this case might therefore also be attributed to reading skills rather than the testing situation alone. Nonetheless, as Chow et al. (2021) already suggested, future studies might want to integrate general test anxiety measures to control for its influence on outcomes of state anxiety in relation to a reading task.
The findings should also be interpreted in light of additional factors that might influence the development of anxiety, namely language aptitude and prior language achievement (Sparks, 2025). Although these factors were not directly measured in the present study, participants’ reading general self-evaluations may partly reflect achievement history. Dolean et al. (2026) have shown that adding measures of initial reading skills and reading development in a longitudinal study (elementary through end of middle school) of Romanian students seems to clarify the pathways leading to first language reading anxiety. Future research might want to apply a similar design to evaluate whether results can be replicated for German adults with low literacy skills.
To be able to investigate the targeted population, young adults with low literacy skills, participants were recruited from schools. As study participation was voluntary, the possibility of self-selection bias should be considered. Participants received reimbursement and were assured their performance on the tests would remain confidential and would not affect their grades. Their participation was further supported by the teachers. Nonetheless, it is still possible that students with higher anxiety, particularly those with co-occurring reading difficulties, might have been less likely to participate, meaning they may be underrepresented in the given sample. Group differences in anxiety levels, as well as associations of both reading skills and general self-evaluations of reading skills with reading task-related state anxiety, may have been affected if self-selection bias occurred. The results should therefore be interpreted while keeping this in mind.
Finally, the present investigation aimed to learn about the emotional aspects of adults with low literacy skills in relation to reading tasks. Thus, we divided the sample into groups based on reading comprehension level. However, as the LGVT uses school-grade norms and was not specifically designed for adults, results should be interpreted while keeping in mind that our grouping reflects expected school-grade reading levels rather than adult-specific norms. Additionally, this approach may have overlooked individual differences within the groups. Treating reading skills as a continuous variable may provide additional valuable information in future analyses.

6. Conclusions

This study contributes to the understanding of the subjective experience of reading by focusing specifically on young adults with low reading comprehension who read German and were educated in the German school system, and by comparing them to young adults with higher reading levels. The results suggest that individuals with low reading comprehension proficiency tend to have lower general self-evaluations of reading skills as compared to those with typical reading skills. However, for this population, we did not find an association between reading task-related state anxiety and both measurable reading skills and general self-evaluations of these skills. Thus, contrary to previous findings that were largely based on younger children (e.g., Francis et al., 2019; Johnson et al., 2026), in the present study, low reading comprehension level alone did not account for elevated reading task-related state anxiety. The current work focuses on an understudied population with various background characteristics that may interact with the results, such as low literacy in young adulthood, the German education system, and the German language. More research is hence needed to understand whether our findings can be replicated in other populations of adults with low literacy skills in different languages and educational contexts.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/educsci16111811/s1, Prompt presented to ChatGPT-4 (OpenAI, 2024) for the creation of two texts for a multiple-choice reading comprehension task.

Author Contributions

Conceptualization, I.B.-K. and E.H.; methodology, I.B.-K.; formal analysis, E.H.; investigation, I.B.-K., L.G., and E.H.; resources, I.B.-K., and L.G.; data curation, E.H.; writing—original draft preparation, E.H.; writing—review and editing, I.B.-K., L.G., and E.H.; supervision, I.B.-K.; project administration, I.B.-K., L.G., and E.H.; funding acquisition, I.B.-K. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—Project number 512028503.

Institutional Review Board Statement

The study was approved by the institutional ethics committee of the German Institute for Adult Education—Leibniz Centre for Lifelong Learning (registered under DIE-LEK-2021-004).

Data Availability Statement

The authors are currently processing the data and can make it available upon request. Once the project and the writing of the planned articles are completed, data will be archived for Good Scientific Practice on secure servers of the German Institute for Adult Education—Leibniz Centre for Lifelong Learning. In addition, long-term archiving for at least 10 years will be done externally, through the German “Verbund Forschungsdaten Bildung” in a certified data repository (regarding the FAIR principles) for researchers for secondary analysis, provided they adhere to the acceptable ethics and data protection guidelines. The study will be referenced by the “Verbund Forschungsdaten Bildung”, the “DIE-directory for Research data,” and is planned to be accessed through classification of these institutes. A DOI will then be available.

Acknowledgments

We would like to thank the schools, adult education centers, and the participants for taking part in our study. We also want to thank our research assistants who helped collect and code the data, as well as the Research Infrastructures Department at the German Institute for Adult Education—Leibniz Centre for Lifelong Learning for their support in the implementation of the project.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Abbreviations

The following abbreviations are used in this manuscript:
PRAXPoor Reading-Anxiety model
RCReading comprehension
RSReading speed

References

  1. Alamer, A., Alrabai, F., & Sparks, R. (2025). Reducing language anxiety by increasing language achievement: A new experimental study. Language Teaching Research. OnlineFirst. [Google Scholar] [CrossRef] [Scilit]
  2. Alamer, A., & Lee, J. (2024). Language achievement predicts anxiety and not the other way around: A cross-lagged panel analysis approach. Language Teaching Research, 28(4), 1572–1593, (Erratum in Language Teaching Research, 28, https://doi.org/10.1177/13621688231155440). [Google Scholar] [CrossRef] [Scilit]
  3. Aro, T., Rossi, M.-L., Paakkari, L., & Torppa, M. (2025). Experienced protective factors while growing up with reading disability. Learning Disability Quarterly, 48(3), 173–187. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  4. Bar-Kochva, I., Gerkens, L., & Herbig, E. (2026). The effect of digital read-aloud on text comprehension and its relations with decoding and listening comprehension skills among students at lower levels of the educational system [Manuscript submitted for publication].
  5. Bonifacci, P., Candria, L., & Contento, S. (2008). Reading and writing: What is the relationship with anxiety and depression? Reading and Writing, 21(6), 609–625. [Google Scholar] [CrossRef] [Scilit]
  6. Brumariu, L. E., Waslin, S. M., Gastelle, M., Kochendorfer, L. B., & Kerns, K. A. (2023). Anxiety, academic achievement, and academic self-concept: Meta-analytic syntheses of their relations across developmental periods. Development and Psychopathology, 35(4), 1597–1613. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  7. Carroll, J. M., & Iles, J. E. (2006). An assessment of anxiety levels in dyslexic students in higher education. British Journal of Educational Psychology, 76(3), 651–662. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  8. Chapman, J. W., & Tunmer, W. E. (1997). A longitudinal study of beginning reading achievement and reading self-concept. British Journal of Educational Psychology, 67(3), 279–291. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  9. Chow, B. W.-Y., Mo, J., & Dong, Y. (2021). Roles of reading anxiety and working memory in reading comprehension in English as a second language. Learning and Individual Differences, 92, 102092. [Google Scholar] [CrossRef] [Scilit]
  10. Christensen, R. H. B. (2025). Ordinal-regression models for ordinal data [R package version 2025.12-29]. Available online: https://CRAN.R-project.org/package=ordinal (accessed on 3 August 2026).
  11. Dolean, D. D., Khanolainen, D., & Torppa, M. (2026). Reading and cognitive predictors of reading anxiety: An 8-year longitudinal study. Journal of Educational Psychology. Advance online publication. [Google Scholar] [CrossRef] [Scilit]
  12. Edwards, A. A., Daucourt, M. C., Hart, S. A., & Schatschneider, C. (2023). Measuring reading anxiety in college students. Reading and Writing, 36(5), 1145–1180. [Google Scholar] [CrossRef] [Scilit]
  13. Fishstrom, S., Capin, P., Fall, A.-M., Roberts, G., Grills, A. E., & Vaughn, S. (2024). Understanding the relation between reading and anxiety among upper elementary students with reading difficulties. Annals of Dyslexia, 74(1), 123–141. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  14. Fox, J., & Weisberg, S. (2019). An R companion to applied regression (3rd ed.). SAGE Publications. Available online: https://www.john-fox.ca/Companion/ (accessed on 2 October 2026).
  15. Francis, D. A., Caruana, N., Hudson, J. L., & McArthur, G. M. (2019). The association between poor reading and internalising problems: A systematic review and meta-analysis. Clinical Psychology Review, 67, 45–60. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  16. German Federal Statistical Office. (n.d.). Vocational training. Available online: https://www.destatis.de/EN/Themes/Society-Environment/Education-Research-Culture/Vocational-Training/_node.html (accessed on 4 August 2026).
  17. Greenberg, D., & Feinberg, I. Z. (2019). Adult literacy: A perspective from the United States. Zeitschrift Für Erziehungswissenschaft, 22(1), 105–121. [Google Scholar] [CrossRef] [Scilit]
  18. Hutchins, B. E., & Young, S. G. (2018). State anxiety. In V. Zeigler-Hill, & T. Shackelford (Eds.), Encyclopedia of personality and individual differences. Springer. [Google Scholar] [CrossRef] [Scilit]
  19. Hutzler, D., David, E., Avigal, M., & Azoulay, R. (2014). Learning methods for rating the difficulty of reading comprehension questions. In 2014 IEEE international conference on software science, technology and engineering (pp. 54–62). IEEE. [Google Scholar]
  20. Jalongo, M. R., & Hirsh, R. A. (2010). Understanding reading anxiety: New insights from neuroscience. Early Childhood Education Journal, 37, 431–435. [Google Scholar] [CrossRef] [Scilit]
  21. Johnson, R. M., Schaefer, M., Norris, C. U., Wagner, R. K., & Hart, S. A. (2026). The association between reading anxiety and reading achievement: A meta-analysis and systematic review. Psychological Bulletin, 152(3), 288–325. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  22. Kasperski, R., Shany, M., & Katzir, T. (2016). The role of RAN and reading rate in predicting reading self-concept. Reading and Writing, 29(1), 117–136. [Google Scholar] [CrossRef] [Scilit]
  23. Katzir, T., Kim, Y.-S. G., & Dotan, S. (2018). Reading self-concept and reading anxiety in second grade children: The roles of word reading, emergent literacy skills, working memory and gender. Frontiers in Psychology, 9, 1180. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  24. Kim, S. (2015). ppcor: Partial and semi-partial (part) correlation [R package version 1.1]. Available online: https://CRAN.R-project.org/package=ppcor (accessed on 3 August 2026).
  25. Kwon, H., & Linderholm, T. (2014). Effects of self-perception of reading skill on absolute accuracy of metacomprehension judgements. Current Psychology, 33(1), 73–88. [Google Scholar] [CrossRef] [Scilit]
  26. Laux, L., Hock, M., Bergner-Köther, R., Hodapp, V., & Renner, K.-H. (2013). Das state-trait-angst-depressions-inventar: STADI; manual. Hogrefe. [Google Scholar]
  27. Lenth, R., & Piaskowski, J. (2026). emmeans: Estimated marginal means, aka least-squares means [R package version 2.0.3]. Available online: https://CRAN.R-project.org/package=emmeans (accessed on 3 August 2026).
  28. Li, R. (2022). Foreign language reading anxiety and its correlates: A meta-analysis. Reading and Writing, 35(4), 995–1018. [Google Scholar] [CrossRef] [Scilit]
  29. Lindeblad, E., Svensson, I., & Gustafson, S. (2016). Self-concepts and psychological well-being assessed by beck youth inventory among pupils with reading difficulties. Reading Psychology, 37(3), 449–469. [Google Scholar] [CrossRef] [Scilit]
  30. Macdonald, K. T., Cirino, P. T., Miciak, J., & Grills, A. E. (2021). The role of reading anxiety among struggling readers in fourth and fifth grade. Reading & Writing Quarterly, 37(4), 382–394. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  31. Marsh, H. W., & Shavelson, R. (1985). Self-concept: Its multifaceted, hierarchical structure. Educational Psychologist, 20(3), 107–123. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  32. McArthur, G. (2022). Poor reading and anxiety (PRAX): Building a theory and practice. Australian Journal of Learning Difficulties, 27(1), 169–180. [Google Scholar] [CrossRef] [Scilit]
  33. McArthur, G. M., Filardi, N., Francis, D. A., Boyes, M. E., & Badcock, N. A. (2020). Self-concept in poor readers: A systematic review and meta-analysis. PeerJ, 8, e8772. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  34. Meer, Y., Breznitz, Z., & Katzir, T. (2016). Calibration of self-reports of anxiety and physiological measures of anxiety while reading in adults with and without reading disability. Dyslexia, 22(3), 267–284. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  35. Mullis, I., Von Davier, M., Foy, P., Fishbein, B., Reynolds, K., & Wry, E. (2023). PIRLS 2021 international results in reading. TIMSS & PIRLS International Study Center. [Google Scholar] [CrossRef] [Scilit]
  36. Novita, S., Uyun, Q., Witruk, E., & Siregar, J. R. (2019). Children with dyslexia in different cultures: Investigation of anxiety and coping strategies of children with dyslexia in Indonesia and Germany. Annals of Dyslexia, 69(2), 204–218. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  37. OECD. (2019). PISA 2018 results (volume II): Where all students can succeed. OECD Publishing. [Google Scholar] [CrossRef] [Scilit]
  38. OECD. (2024). Do adults have the skills they need to thrive in a changing world?: Survey of adult skills 2023. OECD. [Google Scholar] [CrossRef] [Scilit]
  39. OECD. (2026). PISA 2025 results (volume I): Future ready students. OECD Publishing. [Google Scholar] [CrossRef] [Scilit]
  40. OpenAI. (2024). ChatGPT-4 [Large language model]. Available online: https://chat.openai.com (accessed on 24 August 2024).
  41. Orbach, L., Schmitz, F., Fritz, A., & Willmes, K. (2026). Reading and math anxiety in children: Differential roles of state and trait components in academic performance, and the moderating effects of intelligence and time pressure. Frontiers in Child and Adolescent Psychiatry, 5, 1778068. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  42. Oswald, W. D. (2016). Zahlen-verbindungs-test (3rd rev. and renormed ed.). Hogrefe Verlag. [Google Scholar]
  43. Perin, D. (Ed.). (2020). The Wiley handbook of adult literacy (1st ed.). Wiley. [Google Scholar] [CrossRef] [Scilit]
  44. Piccolo, L. R., Giacomoni, C. H., Julio-Costa, A., Oliveira, S., Zbornik, J., Haase, V. G., & Salles, J. F. (2017). Reading anxiety in L1: Reviewing the concept. Early Childhood Education Journal, 45(4), 537–543. [Google Scholar] [CrossRef] [Scilit]
  45. Polychroni, F., Antoniou, A.-S., Kofa, O., & Charitaki, G. (2024). Reading self-concept, trait emotional intelligence and anxiety of primary school children with dyslexia. Frontiers in Education, 9, 1371627. [Google Scholar] [CrossRef] [Scilit]
  46. Ramirez, G., Shaw, S. T., & Maloney, E. A. (2018). Math anxiety: Past research, promising interventions, and a new interpretation framework. Educational Psychologist, 53(3), 145–164. [Google Scholar] [CrossRef] [Scilit]
  47. R Core Team. (2025). R: A language and environment for statistical computing [Computer software]. R Foundation for Statistical Computing. Available online: https://www.R-project.org/ (accessed on 19 October 2026).
  48. Richter, T., & Schnotz, W. (2018). Textverstehen. In D. H. Rost, J. R. Sparfeldt, & S. R. Buch (Eds.), Handwörterbuch pädagogische psychologie (5th ed., pp. 826–837). Beltz. [Google Scholar]
  49. Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an accessibility model of emotional self-report. Psychological Bulletin, 128(6), 934–960. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  50. Roeschl-Heils, A., Schneider, W., & Van Kraayenoord, C. E. (2003). Reading, metacognition and motivation: A follow-up study of German students in grades 7 and 8. European Journal of Psychology of Education, 18(1), 75–86. [Google Scholar] [CrossRef] [Scilit]
  51. Saito, Y., Horwitz, E. K., & Garza, T. J. (1999). Foreign language reading anxiety. The Modern Language Journal, (83), 202–218. [Google Scholar] [CrossRef] [Scilit]
  52. Schneider, W., Schlagmüller, M., & Marco, E. (2017). Lesegeschwindigkeits-und-verständnistest für die Klassen 5-12: LGVT 5-12. Hogrefe. [Google Scholar]
  53. Share, D. L. (2008). On the Anglocentricities of current reading research and practice: The perils of overreliance on an “outlier” orthography. Psychological Bulletin, 134(4), 584–615. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  54. Sparks, R. L. (2025). Second language anxiety: Affective or linguistic variable? Multilingual Matters. [Google Scholar]
  55. Teimouri, Y., Goetze, J., & Plonsky, L. (2019). Second language anxiety and achievement: A meta-analysis. Studies in Second Language Acquisition, 41(2), 363–387, (Erratum in 2019, Studies in Second Language Acquisition, 41, 489). [Google Scholar] [CrossRef] [Scilit]
  56. Torppa, M., Aro, T., Eklund, K., Parrila, R., Eloranta, A.-K., & Ahonen, T. (2024). Adolescent reading and math skills and self-concept beliefs as predictors of age 20 emotional well-being. Reading and Writing, 37(8), 2075–2099. [Google Scholar] [CrossRef] [Scilit]
  57. Visser, L., Kalmar, J., Linkersdörfer, J., Görgen, R., Rothe, J., Hasselhorn, M., & Schulte-Körne, G. (2020). Comorbidities between specific learning disorders and psychopathology in elementary school children in Germany. Frontiers in Psychiatry, 11, 292. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  58. Wang, L.-C., Chen, J.-K., & Poon, K. (2023). Relationships between state anxiety and reading comprehension of Chinese students with and without dyslexia: A cross-sectional design. Learning Disability Quarterly, 46(4), 247–260. [Google Scholar] [CrossRef] [Scilit]
  59. World Health Organization. (2018). International statistical classification of diseases and related health problems [11th Revision]. World Health Organization. Available online: https://icd.who.int/browse/2026-01/mms/en#605267007 (accessed on 19 October 2026).
  60. Yee, T. W. (2015). Vector Generalized Linear and Additive Models: With an Implementation in R. Springer. [Google Scholar]
  61. Zeileis, A. (2004). Econometric computing with HC and HAC covariance matrix estimators. Journal of Statistical Software, 11(10), 1–17. [Google Scholar] [CrossRef] [Scilit]
  62. Zeileis, A., & Hothorn, T. (2002). Diagnostic checking in regression relationships. R News, 2(3), 7–10. Available online: https://CRAN.R-project.org/doc/Rnews/ (accessed on 19 October 2026).
  63. Zeileis, A., Köll, S., & Graham, N. (2020). Various versatile variances: An object-oriented implementation of clustered covariances in R. Journal of Statistical Software, 95(1), 1–36. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Undergraduate Research and Institutional Logics: Tensions and Conflicts at an Undergraduate-Serving Institution
Previous
Classroom Simulations for Social–Emotional Learning and Inclusive Teacher Education: A Four-Layer Conceptual Framework in the Age of AI
Next