Methods

This section includes information on the participants’ characteristics, recruitment process, descriptions of the instruments used, and data analysis procedures.

3.1. Participants

In order to determine the required sample size for this thesis, a priori power analysis was conducted using the open-source G*Power application (version 3.1.9.6) (Faul et al., 2007). An effect size of .15, alpha value of .05, a confidence interval of .95, and variable count of 10 were used as the inputs used for calculation. The effect size of .15 was decided on in light of Dodell-Feder and Tamir’s (2018) meta-analysis regarding readership and social cognition, which determined the relationship between the two variables to have an effect size between .15 and .16. The target sample size was calculated as 172. The total number of participants who completed the study was 284. Participants were excluded from analysis based on age and highest education level. The criteria for inclusion were:

1. must be between 18 and 30 (inclusive at both ends)

2. must either:

   1. hold at least a Bachelor’s degree, or

   2. hold at least a high school diploma and currently be enrolled in a Bachelor’s degree programme

After the inclusion criteria were applied, the final sample consisted of 258 participants. As all participants completed the survey in full, there are no missing data. Accordingly, all tables, figures, and analyses in this thesis are based on a sample size of 258.

*Table 1* Distribution of demographic variables

Variable	n	%
Sex
Male	50	19.4
Sex
Male	50	19.4
Female	208	80.6
Age
19	13	5.0
20	50	19.4
21	60	23.3
22	59	22.9
23	24	9.3
24	18	7.0
25	6	2.3
26	8	3.1
27	7	2.7
28	9	3.5
29	1	0.4
30	3	1.2
Marital status
Single	147	57.0
Partnered	104	40.3
Married	7	2.7
Monthly spending (TL)
0–15,000	102	39.5
15,001–30,000	88	34.1
30,001–45,000	35	13.6
45,001–60,000	15	5.8
60,001–75,000	4	1.6
75,001–90,000	9	3.5
Over 90,000	5	1.9
Highest education
High school degree	177	68.6
Associate degree	4	1.6
Bachelor’s degree	67	26.0
Master’s degree	9	3.5
PhD	1	0.4

Note. TL = Turkish lira.

3.2. Instruments

The instruments used in the study include an Informed Consent Form (see Appendix A), a Demographic Information Form (see Appendix B), a Canon Readership Checklist (see Appendix C), a Reading Habits Self-Report Form (see Appendix D), the Turkish adaptation of the State–Trait Anxiety Inventory (STAI; see Appendices E and F), the Turkish adaptation of the Existential Concerns Questionnaire (ECQ; see Appendix G), the Turkish adaptation of the Mentalization Scale (MentS; see Appendix H), the Turkish adaptation of the Questionnaire of Cognitive and Affective Empathy (QCAE; see Appendix I), and a Debriefing Form presented after participation (see Appendix J). The following subsections describe these instruments, excluding the Informed Consent Form and the Debriefing Form.

3.2.1. Demographic Information Form

Through the demographic information form, which was prepared by the author of this study, participants’ background information was collected regarding sex, age, marital status, average monthly expenditure, and formal education level. See Appendix B for a copy of the form.

3.2.2. Canon Readership Checklist

This checklist was not designed as a psychometric measure but was adapted from a national literary list to serve as an index of literary exposure. Curated by the Turkish Ministry of Education (Türkiye Cumhuriyeti Milli Eğitim Bakanlığı [MEB]), a literary canon for high school students was officially released to the public on 19 August 2004 through a circular signed by then-Minister Hüseyin Çelik, under the name of 100 Temel Eser (100 Essential Works). The circular notes the reason behind the creation of this canon as instilling reading habits and aesthetic appreciation in students, improving their language abilities, and enriching cultural heritage (T.C. Milli Eğitim Bakanlığı, 2004). As for the selection process of the canon, the circular asserts that it was done in cooperation with various non-governmental organisations, scientists, journalists, and public intellectuals. For Turkish literature, only works by deceased Turkish authors were included to avoid controversy. The canon was modified once on 29 April 2008, with the removal of Yusuf Atılgan’s Anayurt Oteli and inclusion of Tarık Buğra’s Osmancık instead (T.C. Milli Eğitim Bakanlığı, 2008).

In the present study, the canon was presented to the participants as a checklist, who were asked to check the boxes for the works they had read. The total number of books presented to the participants was 101, including both the removed and added books from 2008. The total number of books read by the participants were used as their scores, with possible scores ranging from 0 (no books read) to 101 (all books read). Though not a validated scale like the Author Recognition Test (ART) used in similar studies outside of Turkey, the canon was selected as a proxy due to its inclusion of a broad range of literary works and the significant exposure students had to it through formal education between 2004 and 2018. Even after its official rescission in 2018 (T.C. Milli Eğitim Bakanlığı, 2018), the canon maintains a strong cultural influence. A copy of the form, including the instructions and all items is available in Appendix C.

3.2.3. Reading Habits Self-Reports

Participants’ general reading habits were assessed using a set of questions inspired by the reading time estimation section in the study conducted by Acheson and colleagues (2008). In their study, the authors created a self-report format to measure the participants’ weekly reading time across different materials, not as a validated scale but a practical method for capturing their print exposure.

In the present study, participants were asked to indicate how many hours they spent reading on average in a week with respect to six different materials: university textbooks, academic articles (such as peer-reviewed articles), magazines, newspapers, e-mails, and written internet content (such as blogs and textual social media like Twitter or Reddit). The response options were ordinal, ranging from “0 hours” to “7 or more hours”. These items were used solely for descriptive purposes and were not included in the main statistical analyses. See Appendix D for a copy.

3.2.4. State-Trait Anxiety Inventory (STAI)

The State-Trait Anxiety Inventory developed by Spielberger and colleagues (1970) is a widely used self-report measure distinguishing between two types of anxiety: state anxiety, defined as “a temporary emotional condition in response to situational stress”, and trait anxiety, which is defined as “a general and stable tendency to perceive situations as threatening”.

The inventory consists of two scales, State Anxiety (STAI–S) and Trait Anxiety (STAI–T), which include 20 items each for a total of 40. The items are rated on a 4-point Likert-type scale, ranging between 1 (“Not at all”) and 4 (“Very much so”), indicating the intensity or frequency of symptoms related to anxiety. Several items of the scale are reverse-coded to control for acquiescence bias. For the State Anxiety scale, respondents are instructed to answer based on how they feel “right now, at this moment”, and “how they generally feel” for the Trait Anxiety scale (Spielberger et al., 1970). Each scale is scored separately, with higher scores indicating greater levels of anxiety. The scores range between 20 and 80. The STAI does not provide diagnostic categories determined by cut-off points, and is rather used to assess the relative differences in anxiety scores across individuals.

Although an updated version of the inventory (Form Y) was later published (Spielberger et al., 1983), a Turkish adaptation exists only for the original Form X (Spielberger et al., 1970), which was translated and validated by Öner and LeCompte (1985). This version, used in the present study, demonstrates excellent psychometric properties (Cronbach’s α = .94) and has been used extensively with both clinical and non-clinical populations, including university students, making it suitable for the present study’s aims and sample demographic.

In the present study, Cronbach’s α coefficients for the State and Trait forms were .94 and .83, indicating excellent and good internal consistency, respectively. Consult Appendix E for the State Form, and Appendix F for the Trait Form.

3.2.5. Existential Concerns Questionnaire (ECQ)

The Existential Concerns Questionnaire (ECQ) is a questionnaire developed by van Bruggen and colleagues (2017) in order to measure sensitivity to existential themes such as death, meaninglessness, guilt, social isolation, and identity disturbance. The scale was developed in response to the lack of a validated tool for assessing such concerns in existential theory research and existential therapy settings (van Bruggen et al., 2017).

The questionnaire has 3 subscales, namely General Existential Anxiety (General EA), Death Anxiety, and Avoidance. There are a total of 22 items, each rated on a 5-point Likert scale from 1 (“Never”) to 5 (“Always”). The respondent’s score is calculated by taking the sum of their ratings for each of the subscales. Furthermore, as the scale is unidimensional, an overall score is calculated by summing the ratings for all items. Hence, for each respondent, an Overall Existential Concerns, General EA, Death Anxiety, and Avoidance scores can be calculated. For the subscales, example items are I worry about not living the life that I could live, I worry about having to let go of everything at the moment of my death, and I try to avoid the question of who I really am respectively.

The Turkish adaptation of the questionnaire was done by Ümmet et al. (2018), keeping the original’s structure, item count, subscales, and scoring intact, with a Cronbach’s α coefficient of .81, indicating good internal consistency.

In the present study, Cronbach’s α coefficient for the overall scale was .93, indicating excellent internal consistency. The subscales had acceptable to excellent internal consistency, as the Cronbach’s α coefficients were .90 for General Existential Anxiety, .84 for Death, and .73 for Avoidance. Only the overall ECQ score was used in the main analyses; subscale scores were included for descriptive purposes. A copy of the questionnaire used in the study, including all items, instructions, subscale classifications, and reverse-coded items, can be consulted in Appendix G.

3.2.6. Mentalization Scale (MentS)

The Mentalization Scale (MentS) is a self-report scale developed by Dimitrijević and colleagues (2017) to measure the participants’ mentalisation, defined as the capacity to interpret one’s own and others’ behaviour in terms of mental states such as desires, feelings, and beliefs as a distinct construct from empathy, emotional intelligence, or general ToM capabilities, avoiding the overlap between these different constructs. The scale distinguishes between two aspects of mentalisation: self-related mentalisation (MentS–S), other-related mentalisation (MentS–O), as well as a third factor assessing the respondent’s level of motivation for analysing mental states, in other words, their motivation to mentalise (MentS–M).

The scale consists of 28 items which are rated on a 5-point Likert-type scale. The ratings range from 1 (“Completely incorrect”) to 5 (“Completely correct”). Four different scores can be calculated for each respondent: Overall Mentalization (by adding together the ratings for all items), Self-Related Mentalization, Other-Related Mentalization, and Motivation to Mentalize. Example statements for MentS–S, MentS–O, and MentS–M include Often I cannot explain, even to myself, why I did something, I can recognize other people’s feelings, and I do not like to waste time trying to understand in detail other people’s behavior, respectively.

The Turkish adaptation of the scale was carried out by Törenli Kaya and colleagues (2023) who reduced the item count to 25 because 3 items from the original study had low psychometric properties in the Turkish sample. The adaptation retains other properties of the original scale. After the removal of said items, a Cronbach’s α coefficient of .84 was reported (Törenli Kaya et al., 2023).

In the present study, Cronbach’s α coefficient for the overall MentS scale was .86. Among the subscales, Cronbach’s α was .83 for Self, .82 for Other, and .75 for Motivation. These values indicate good internal consistency for the overall scale and the Self and Other subscales, with acceptable reliability for the Motivation subscale. The Overall, Self, and Other scores were used in the main analyses; the Motivation subscale was included in descriptive analyses. For the complete questionnaire and its structure, refer to Appendix H

3.2.7. Questionnaire of Cognitive and Affective Empathy (QCAE)

The Questionnaire of Cognitive and Affective Empathy (QCAE) is a multidimensional self-report scale developed by Reniers and colleagues (2011), used to measure two distinct components of empathy, namely cognitive empathy, the ability to understand others’ thoughts and emotions, and affective empathy, the capacity to respond to and resonate with others’ emotions. The motivation behind Reniers and colleagues (2011) creating the questionnaire was the need for a scale that was in line with the developments in empathy research, clearly distinguishing cognitive aspects from the affective. Furthermore, the Cognitive Empathy and Affective Empathy subscales are made up of several components. The former encompasses Perspective Taking, Online Simulation, and the latter encompasses Emotion Contagion, Proximal Responsivity, and Peripheral Responsivity.

The scale consists of 31 items rated on a 4-point Likert-scale, ranging from 1 (“Strongly disagree”) to 5 (“Strongly agree”). Eight different scores can be calculated from the responses: Overall Empathy (the sum of all ratings), Cognitive Empathy, Affective Empathy, Perspective Taking, Online Simulation, Emotion Contagion, Proximal Responsivity, and Peripheral Responsivity.

Beginning with the Cognitive Empathy components, Perspective Taking is defined as “intuitively putting oneself in another’s point of view” (Reniers et al., 2011), assessed by items such as I can easily work out what another person might want to talk about. Online Simulation, on the other hand, is “an effortful attempt to put oneself in another person’s position by imagining what the person is feeling” (Reniers et al., 2011), which is assessed by an item such as Before criticizing somebody, I try to imagine how I would feel if I was in their place. As for Affective Empathy components, Emotion Contagion is the “automatic mirroring of the feelings of others” (Reniers et al., 2011). An example item for this component is I am happy when I am with a cheerful group and sad when the others are glum. The component Proximal Responsivity is defined as the “affective response when witnessing the mood of others in a close social context” (Reniers et al., 2011), an example item for which is It pains me to see young people in wheelchairs. Finally, Peripheral Responsivity is similar to Proximal Responsivity, the only difference being that the target is not in a close social context but rather a distant context, assessed by an item such as I usually stay emotionally detached when watching a film.

The questionnaire was adapted to Turkish by Gıca and colleagues. (2021), all the items and the structure being kept as is. The internal validity of the adaptation is good with a Cronbach’s alpha coefficient of α = .82 (Gıca et al. 2021).

In the present study, Cronbach’s α coefficient for the overall QCAE scale was .88. Subscale alphas were .87 for Cognitive Empathy (with .88 for Perspective Taking and .78 for Online Simulation) and .76 for Affective Empathy (with .72 for Emotion Contagion, .63 for Proximal Responsivity, and .49 for Peripheral Responsivity). These values indicate good internal consistency for the overall scale and most subscales. Proximal Responsivity showed acceptable reliability (α = .63), while Peripheral Responsivity displayed low internal consistency (α = .49), suggesting caution in interpreting results involving this component. However, this limitation is not critical for the present study, as only the overall, cognitive, and affective empathy scores were used in the main analyses, while individual component scores were examined only descriptively.

3.3. Procedures

Participants were selected through convenience sampling, and consisted entirely of voluntary individuals, who were recruited online through social media platforms, face-to-face through acquaintances and social circles of the author. They took part in the study through a Google Form survey using their personal smartphones, tablets or computers. The data was collected between 26 September 2024 and 30 April 2025. Participants digitally confirmed an informed consent form before they could participate. No information which could expose the participants’ anonymity was collected. The author’s and the thesis supervisor’s contact information were shared for possible queries. In the digital survey, the structure was as follows: (1) Informed Consent, (2) Demographics Form, (3) Book Checklist, (4) Reading Habits Self-Report, (5) STAI, (6) ECQ, (7) MentS, (8) QCAE, and finally (9) Debriefing. The procedure of the study was approved by the Science-Ethics Committee of Yeditepe University (see Appendix K).

3.4. Data Analysis

All analyses were conducted using Python (version 3.13.3; Python Software Foundation, 2025) on macOS Sonoma (version 14.4.1). The following open-source packages were used: pandas (version 2.1.2) for data structures and data manipulation (McKinney, 2010), NumPy (version 1.26.1) for numerical operations (Harris et al., 2020), Pingouin (version 0.5.5) for correlation analysis (Vallat, 2018), scikit-learn (version 1.3.2) for standardisation of continuous variables (Pedregosa et al., 2011), and Statsmodels (version 0.14.0) for linear and hierarchical regression modelling (Seabold & Perktold, 2010). Visualisations were created using Matplotlib (version 3.8.1; Hunter, 2007) and Seaborn (0.13.0; Waskom, 2021).

All participants (N = 258) completed every measure in full. No missing data imputation or listwise deletion was necessary. Logarithmic values for participants’ age and number of books read were used in correlation and regression analyses. Correlational analyses were conducted using Pearson’s r, and hierarchical regression models were built to examine the effects of demographic and psychological variables on empathy and mentalisation. Hierarchical multiple regression analyses were conducted using ordinary least squares (OLS) estimation with classical (non-robust) standard errors. Variables were entered in three blocks: demographic control variables, primary predictors, and interaction terms. Categorical variables in the control block were dummy-coded using reference categories. Predictors involved in interaction terms were mean-centred prior to computing product terms. Standardised beta coefficients (β) were calculated by multiplying each unstandardised coefficient by the ratio of the standard deviation of the predictor to that of the outcome (i.e., β = B × SDₓ / SDᵧ). Model assumptions, including normality of residuals, homoscedasticity, multicollinearity (VIF), and influence (Cook’s distance), were checked and met. Results were reported in accordance with APA 7 guidelines (American Psychological Association, 2020), with unstandardised coefficients (B), standard errors (SE), standardised beta (β), t-values, p-values, and model statistics (ΔR², R², F_change, p_change and total model F) for each step. An alpha level of .05 was used for significance testing in regression analyses, while both .05 and .001 thresholds were considered when interpreting correlation results.

All materials, including the dataset, analysis scripts, and survey preview, are available in an OSF repository (see Appendix L).