Compulsive Study Behaviors Are Associated with Eating Disorders and Have Independent Negative Effects on Well-Being: A Structural Equation Model Study among Young Musicians

Abstract

Compulsive overworking and eating disorders (EDs) show considerable similarities in terms of risk factors (e.g., rigid perfectionism), clinical manifestation (e.g., excessive controlling behaviors), and consequences (e.g., physical exhaustion and depression). This study aimed to examine the hypotheses that compulsive study behaviors (conceptualized as study addiction) are related to EDs and that they have independent negative effects on well-being among young musicians, who constitute a highly vulnerable population for these types of problematic behaviors. The relatively high prevalence of study addiction and its pronounced negative relationship with psychosocial functioning make it a pending challenge for sustainable education. A total of 255 students from various music academies in Poland took part in the study. The Bergen Study Addiction Scale, assessing compulsive studying (conceptualized as addictive behavior), the Eating Attitude Test-26 (EAT-26), the Perceived Stress Scale, the Hospital Anxiety and Depression Scale, and the quality-of-life measure were used. A structural equation model was investigated. Study addiction was positively related to the general factor of EDs and the social pressure component. Both problematic behaviors showed negative and independent effects on the well-being of young musicians. EDs may be 8 to 16 times more prevalent among the students of music academies who are addicted to studying than among the general population. About 80% of those students showing all seven symptoms of study addiction exhibited at least mild depression, while more than half had clinically significant levels of depression. Almost 90% had clinically significant levels of anxiety. Without addressing co-occurring study addiction and eating disorders, including their commonalities and idiosyncrasies, their prevention and treatment cannot be effective and it will substantially affect the sustainability of education and work.

1. Introduction

School, academic, and work stress is recognized as a substantial factor affecting the mental and physical health of students and the working population, negatively impacting their productivity [1,2,3,4]. Consequently, it poses a considerable challenge to sustainable education and work. Professional burnout, defined as a syndrome resulting from ill-managed work stress, is a growing rather than diminishing problem, and an analogous phenomenon is already apparent among school and undergraduate students [5,6,7,8]. A tendency to overburden oneself with work-related activities that becomes compulsive constitutes a specific problem in this context. It is conceptualized as an addictive disorder, known as work addiction or workaholism, and is strictly related to chronic stress, burnout, and mental and physical health problems [9,10].

While compulsive overworking is typically associated with behaviors related to professional work, eating disorders (EDs), especially anorexia nervosa (AN), are an important cause of unemployment and result in severe work impairment [11,12]. At the same time, these two classes of problematic behaviors are closely related, due to the shared risk factors (e.g., rigid perfectionism), similar clinical manifestations (e.g., excessive controlling behaviors), and consequences (e.g., physical exhaustion and depression [13]). This study aimed to examine the hypotheses that compulsive study behaviors are related to EDs, and that they have independent negative effects on well-being. Without addressing both substantially co-occurring psychological problems, including their commonalities and idiosyncrasies, their prevention and treatment cannot be effective, and this will substantially affect the sustainability of education and work.

Compulsive study behaviors are conceptualized as an addictive disorder and an early form of work addiction [14,15] with which they share profound phenomenological similarities, risk factors, consequences, prevalence rates, and longitudinal associations (for an overview, see [9]) [16]. These behaviors are tentatively and descriptively called “study addiction”; however, they are not formally recognized as an addictive disorder among the official classifications of diseases and disorders. Importantly, compulsive behaviors related to the uncontrollable need for productivity and a complete focus on work-related activities, to the exclusion of other spheres of life, are currently diagnosed as symptoms of obsessive-compulsive personality disorder in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders [17], corresponding to the anankastia domain in terms of personality disorder in the International Classification of Diseases, 11th revision [18].

However, a comparatively long history of research [19,20,21,22] and recent advances in the work addiction literature (see [9,23]) have provided a considerable body of evidence suggesting that compulsive overworking represents an addictive process and OCPD/anankastia may be its major risk factor [10]. Figure 1 represents a schematic summary of the relationships among OCPD/anankastia, study addiction, and work addiction. It provides a theoretical framework by which to understand the current conceptualization of study addiction within the existing classifications of diseases and disorders, and the work addiction framework within the behavioral addiction field of research. Study addiction is relatively highly prevalent (from 6 to 17% depending on country and course of study; [14,24]) in comparison to other addictive and compulsive behaviors, and it closely parallels the work addiction prevalence (for comparisons and analyses of prevalence rates, see [25]). Work and study addictions are related to considerable harm to the individual and society. For a comprehensive overview, see [9].

On the individual level, work and study addictions are related to chronic distress, anxiety, depression, suicide risk (substantiated in longitudinal research on work addiction), and likely all stress-related diseases and disorders, due to the high workload and prolonged distress [10]. On the social level, compulsive overworking is related to impaired social functioning, while work addiction is associated with harm to coworkers, work recipients (e.g., the potentially increased risk of medical errors, due to burnout and exhaustion), the organization, partners, and family members [9]. Study addiction is also related to loneliness [14]. On a macro-level, a theoretical model based on empirical findings identifies work and study addictions as potentially major contributing factors to social disintegration [9]. High prevalence and substantial harm make study and work addictions some of the most significant challenges in the addictive behavior field today [9]. If compulsive studying is an addictive disorder, then its reclassification from OCPD will be required in the official classifications of diseases and disorders. One step toward this is examining its associations with other specific disorders, such as EDs, and its independent effects on harm, such as the deterioration of well-being.

Similar to work addiction, study addiction shows comorbidities with a host of other mental health problems, including substantial co-occurrences with other addictive behaviors [25,26], depression, anxiety, and social anxiety [14,24,27]. Following previous findings on the association between compulsive working and food-starving and food-binging behaviors [28,29,30], a recent study showed considerable comorbidity between work addiction and EDs [13]. This is particularly true in the case of anorexia nervosa and bulimia nervosa, where as many as five out of six women previously diagnosed with EDs fulfilled the cut-off score for work addiction. Such substantial co-occurrence may be due to the significant similarities between these problematic behaviors’ risk factors and clinical manifestations. The main common features of work addiction, anorexia nervosa, and bulimia nervosa include a high need for control and the use of particular behaviors to cope with stress and underlying problems [13,31]. These problematic behaviors are also highly associated with mood disorders, perfectionism, OCPD, and personality traits such as neuroticism and competitiveness [10,12,32,33,34,35]. Furthermore, sociocultural factors (e.g., media and peer influences) and family factors (e.g., parental pressure, high expectations, family criticism, family discord, and parental psychopathology) have been suggested to play a significant role in the development and maintenance of work addiction and EDs [36,37,38,39,40,41].

Based on previous studies, a general model explaining the relationship between compulsive learning and EDs, particularly anorexia nervosa and bulimia nervosa, has been suggested. Etiologically, both study addiction and these EDs, in many cases, may have their roots in specific family dynamics that are regulated by strong parental control, parental criticism, and family discord [14,31,42]. In response, compensatory mechanisms related to perfectionism and a high need for control may develop. These may be particularly strong in relation to self-image and the feeling of competence in specific domains, and individualsmay be especially vulnerable to social expectations [14,15,24,37,38,39,40,41]. In general, a need for an impeccable self-image as a physically attractive and academically/professionally competent person may drive exceptionally destructive weight-control behaviors and study-related behaviors. These are, to some degree, regulated by personality traits that are specific and shared in both classes of problematic behaviors. They include narcissistic concerns, high competitiveness, rigid perfectionism, and emotional instability [9,38,39,40,41,43,44]. As a result, individuals use eating and study behaviors to cope with stress and to manage other underlying problems. However, since these are ineffective coping mechanisms, the unsolved problems, together with the growing stress and exhaustion, further aggravate the symptoms of EDs and study addiction. It can be expected that considerable interaction and mutual causal effects may manifest when they co-occur.

On the one hand, compulsive learning behaviors are associated with chronic stress due to several mechanisms. These include: i) study overload and the related health problems due to long study hours, sitting position, and sleep deprivation; ii) examination stress and evaluation anxiety, related to dysfunctional perfectionism and social anxiety; iii) neglecting social relationships, and, in consequence, growing loneliness and lack of social support [9,14,24]. This increased stress may trigger or aggravate EDs. On the other hand, individuals suffering from EDs may use study behaviors to distract themselves mentally from hunger. Stress is also known to affect appetite and food intake, so students addicted to studying may purposefully increase their stress, via excessive learning behaviors and long hours of studying, to down-regulate their food cravings. A detailed analysis of the possible mutual relationships between compulsive overworking and EDs, together with a comprehensive literature overview, can be found in [13].

This study aimed to investigate the relationship between compulsive study behaviors and EDs among the students of music academies, to analyze further the similarities between study and work addiction. Particularly, analogous specific comorbid disorders, showing considerable similarities in risk factors and clinical manifestation, may provide some additional indirect support to the hypothesis that study addiction is an early form of work addiction. Moreover, if these symptoms represent a separate clinical entity and an addictive disorder, then unique negative effects on well-being, above and beyond other similar mental health problems such as EDs, and their shared commonalities, can be expected.

Previous research has used the former clinical diagnoses of EDs [13]. However, the low prevalence of anorexia nervosa or bulimia nervosa makes it challenging to investigate the more nuanced aspects of their relationship with work-related compulsive behaviors and well-being. Addictive behaviors can be effectively studied as continuous phenomena and such an approach provides considerable advantages, especially in terms of statistical power and the more precise estimation of addiction risk or level [25,26]. Therefore, continuous scores on EDs were used in the current study to allow for more advanced analyses of the relationships between them and study addiction.

Moreover, previous studies suggest that particular groups are more likely to be affected by behavioral addictions, such as work/study addiction and/or EDs, for instance, musicians, actors, or dancers [27,39,40,45]. The reasons are the lifestyle, a demanding performance schedule, a shortage of privacy, peer pressure, stress, depression, and performance anxiety [38,40,46,47]. The study’s objective was to examine the relationship between study addiction and eating disorders, and their individual effects on well-being. Investigating the comorbidity of compulsive studying and EDs among young musicians may better capture their associations, due to this group’s relatively high prevalence of both problematic behaviors. The research question is whether and how eating disorders and study addiction are related and what are their relative impacts on well-being among the students of music academies.

Hypotheses

Based on previous research and existing theoretical frameworks, it was hypothesized that: (i) study addiction is positively related to EDs (H1), (ii) study addiction offers a unique contribution to well-being above and beyond EDs (H2). To the authors’ knowledge, the current study is the first to examine these relationships. Its primary aim is to establish the basic effects that can be investigated in more detail in future studies, including those on mutual causal relationships between study addiction and EDs, examined with proper longitudinal designs.

2. Materials and Methods

2.1. Sample

The sample consisted of 255 music academies’ students: 184 (72.2%) were female and 71 (27.8%) were male, with a mean age of M = 23.05 years (SD = 3.46). The mean number of years of formal musical education attained was M = 12.95 years (SD = 4.50). The participants attended different courses, faculties, and years of study at music academies in Poland: the Fryderyk Chopin University of Music in Warsaw, the Krzysztof Penderecki Academy of Music in Cracow, the Ignacy Jan Paderewski Academy of Music in Poznań, the Karol Lipiński Academy of Music in Wrocław, the Stanisław Moniuszko Academy of Music in Gdańsk, the Grażyna and Kiejstut Bacewicz Academy of Music in Łódź, the Karol Szymanowski Academy of Music in Katowice, the Feliks Nowowiejski Academy of Music in Bydgoszcz and the Academy of Art in Szczecin.

2.2. Instruments

Data on the validity and reliability of each measure, including the CFA results within the structural model, average variance extracted (AVE), omega, omega hierarchical, omega subscale, omega hierarchical subscale, composite reliability (CR), and explained common variance (ECV) can be found in

Table 1. SEM model results: standardized estimates, standard errors, statistics, p-values, average variance extracted (AVE), composite reliability (CR), omega for general factor and subscales, omega hierarchical, and explained common variance (ECV). Note: SA = study addiction; G = general factor of EDs; SP = social pressure; FA = food awareness; FP = food preoccupation; WB = well-being.

			Estimate	S.E.	Est./S.E.	Two-Tailed p-Value	AVE	CR	Omega/ Omega Subscale	Omega Hierarchical/Omega Hierarchical Subscale	ECV
SA							0.49	0.87	0.87	-	-
	BSTAS 1	Salience	0.48	0.05	9.68	<0.001
	BSTAS 2	Tolerance	0.55	0.05	11.45	<0.001
	BSTAS 3	Mood modification	0.65	0.04	15.09	<0.001
	BSTAS 4	Relapse	0.81	0.03	24.51	<0.001
	BSTAS 5	Withdrawal	0.78	0.03	25.83	<0.001
	BSTAS 6	Conflict	0.74	0.03	21.66	<0.001
	BSTAS 7	Problems	0.82	0.03	26.99	<0.001
G							0.21	0.67	0.90	0.46	0.35
	EATS 3	Food preoccupation	0.48	0.08	5.75	<0.001
	EATS 4	Binge-eating	0.41	0.09	4.75	<0.001
	EATS 6	Calorie awareness	0.55	0.07	8.33	<0.001
	EATS 7	Carbohydrates avoidance	0.51	0.08	6.47	<0.001
	EATS 8	Social pressure to eat more	−0.05	0.08	−0.98	0.325
	EATS 13	Others’ concerns	−0.25	0.08	−3.37	0.001
	EATS 16	Sugar avoidance	0.32	0.08	4.03	<0.001
	EATS 17	Dieting food	0.41	0.08	5.24	<0.001
	EATS 18	Lack of control	0.72	0.07	9.67	<0.001
	EATS 20	External pressure to eat	0.24	0.08	3.05	0.002
	EATS 21	Food preoccupation	0.81	0.08	11.27	<0.001
SP							0.67	0.86	0.88	0.87	-
	EATS 8	Social pressure to eat more	0.92	0.04	22.26	<0.001
	EATS 13	Others’ concerns	0.75	0.05	15.46	<0.001
	EATS 20	External pressure to eat	0.78	0.05	15.16	<0.001
FA							0.36	0.77	0.83	0.53	-
	EATS 6	Calorie awareness	0.44	0.07	6.50	<0.001
	EATS 7	Carbohydrates avoidance	0.62	0.07	8.76	<0.001
	EATS 16	Sugar avoidance	0.75	0.06	12.77	<0.001
	EATS 17	Dieting foods	0.53	0.07	8.06	<0.001
FP							0.31	0.51	0.90	0.41	-
	EATS 3	Food preoccupation	0.73	0.09	8.49	<0.001
	EATS 4	Binge-eating	0.50	0.09	5.62	<0.001
	EATS 18	Lack of control	0.48	0.10	4.77	<0.001
	EATS 21	Food preoccupation	0.49	0.11	4.63	<0.001
WB							0.55	0.58	0.58	-	-
	Anxiety	Anxiety	−0.93	0.04	−24.49	<0.001
	Depression	Depression	−0.67	0.04	−15.33	<0.001
	Quality of life	Quality of life	0.67	0.05	13.76	<0.001
	Stress	Stress	−0.66	0.05	−13.21	<0.001

. Data on the discriminant validity of the measures can be found in

Table 2. The square root of AVE (SR AVE) and latent variable correlations.

	General	Social Pressure	Food Awareness	Food Preoccupation	Study Addiction	Well-Being
SR AVE	0.46	0.82	0.60	0.56	0.70	0.74
Study Addiction	0.20	0.19	−0.10	−0.02	-	-
Well-being	−0.49	0.00	0.35	−0.11	−0.50	-

. It should be noted that indicators for the bifactor model of the eating attitude test 26 (EAT-26), such as AVE, ECV, and hierarchical reliability coefficients, reflect the fact that the measure is clearly multidimensional, with a general factor of eating disorders and specific factors. Somewhat lower AVE and hierarchical reliability coefficients for the general factor suggest that specific factors explain a considerable amount of variance in the results. At the same time, shared variance with a general factor may explain, to some extent, the mediocre indices for specific factors. Such results are expected for a multidimensional scale with a bifactor structure. These results also support the need to analyze EAT-26 with a bifactor model. Considerable correlation of the general factor with well-being, in comparison to the square root of AVE of the general factor (Table 2), supports the notion that the general factor represents the pathological aspects of eating disorders. Positive or no correlations of specific factors with well-being further support the argument that these factors represent the non-pathological aspects of eating behaviors. The bifactor model allows us to parcel out the variance in items representing the pathological and non-pathological aspects of eating behaviors.

Eating disorders. The eating attitude test 26 (EAT-26; Garner, Olmstedt, Bohr, and Garfinkel, 1982) was used here to identify EDs risk. The original EAT-26 is a 26-item questionnaire investigating three dimensions of EDs: dieting (e.g., “Am terrified about being overweight”), bulimia and food preoccupation (e.g., “Find myself preoccupied with food”), and oral control (e.g., “Avoid eating when I am hungry”). Participants provided answers on a six-point Likert scale, with response categories from 1 (never) to 6 (always). A recoding procedure for the diagnostic response categories is applied. A score of 20 or higher is considered to be an indicator of possible ED behaviors and eating problems [41]. A subsequent study found that a modified factorial structure of EAT-26 better fits the data, with a bifactor model of general EDs factors and four specific factors: social pressure, food awareness, purging behaviors, and food preoccupation [44]. Purging behaviors and food preoccupation factors were assumed to be correlated. This bifactor model was used in the present study, requiring proper model specification.

Study addiction. The Bergen study addiction scale (BStAS; [14,15]) was used to measure compulsive studying, conceptualized as addictive behavior. The instrument consists of seven items that refer to experiences during the past 12 months (e.g., “You studied so much that it has negatively impacted your health”), with a Likert-type response scale, from 1 (never) to 5 (always). This showed good psychometric properties in previous research, including its use in samples of music academies’ students [24], cross-cultural samples [14,48], and high school students [27]. The Cronbach’s alpha reliability coefficient in the current sample was 0.83.

Anxiety and depression. The hospital anxiety and depression scale (HADS; [49]) was used to measure anxiety and depression. The scale consists of seven items that measure the level of anxiety (e.g., “I feel tense or ‘wound up’”) and seven that measure the level of depression (e.g., “I can laugh and see the funny side of things”). Respondents provided answers on a four-point scale, with response categories from 0 to 3, the labels of which vary depending on the items. The scale showed good validity and reliability in previous research, including in screening studies with student samples [50,51]. The Cronbach’s alpha reliability coefficients in the current sample were 0.74 for anxiety and 0.67 for depression. HADS was used for three main reasons: (i) it has been used worldwide for decades and its results can be compared to those in other countries and populations, (ii) it has a well-established and thoroughly empirically validated cut-off score that allows evaluating the comorbidity of study addiction and EDs with anxiety and depression, (iii) clinical levels of anxiety and depression can be approximated (similarly to study addiction and EAT-26) in the general population of students. This is of special importance since all these disorders, but especially study addiction and EDs, are underdiagnosed.

Perceived stress. The perceived stress scale 4 (PSS-4; [52]) was used to measure the perceived stress. The PSS-4 consists of 4 items (e.g., “In the last month, how often have you felt that things were going your way?”), with a Likert-type response ranging from 0 (never) to 4 (very often). The Polish version of the scale showed good validity and reliability in previous research [53]. In the present sample, the Cronbach’s alpha reliability coefficient was 0.72.

General quality of life. This was measured on a short scale based on eight single-item measures of the different aspects of quality of life, developed from the WHOQOL-BREF (short version of the World Health Organization Quality of Life Assessment) [54]. They include the physical domain of quality of life (general health and quality of sleep), the social domain (satisfaction with personal relationships, satisfaction with the support received from friends) and self-esteem, the psychological domain (satisfaction with life and the meaning of life), and general quality of life. All the items were measured using a 9-point Likert-type scale. The first five items ranged from 1 (very dissatisfied) to 9 (very satisfied). Two consecutive items varied from 1 (not at all) to 9 (an extreme amount). The last item was described from 1 (very poor) to 9 (very good). These measures showed good validity and reliability in previous research [55,56,57,58]. In the present sample, the general score was used as a sum of the items, following proper confirmatory factorial validity testing. Cronbach’s alpha coefficient for the general score was 0.83.

2.3. Procedure

Data collection took place from April to May 2021. Students were invited to participate anonymously in the study by lecturers and student councils, via various educational and social portals (convenience sampling). The inclusion criteria were being an adult student of a music academy in Poland. The exclusion criteria included being underaged and not studying at a music academy. In addition, participants giving incomplete answers were excluded from the analyses. Data have been collected using an online survey (Profitest.pl). The estimated response rate was above 62%, based on the number of individuals who completed the survey in relation to those who accessed the link with the survey. No monetary or other material rewards were offered.

2.4. Statistical Analyses

Means, standard deviations, percentages, and Spearman’s rank correlation coefficients were calculated using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, NY, USA). Correlations with all items of EAT-26 were calculated to provide more information on the associations between the particular symptoms of EAT-26 and study addiction. Data on the validity and reliability of each measure were calculated, including the CFA results within the structural model, average variance extracted (AVE), composite reliability, and omega, and for the bifactor model omega hierarchical, omega subscale, omega hierarchical subscale, and explained common variance (ECV). Data on the discriminant validity of the measures were calculated and included the square root of AVE, compared to the latent factor correlations. Correlations between the general and specific factors of EAT-26 were not calculated, as the bifactor model assumes no such correlations. The structural equation model (SEM), with the bifactor specification of EDs, was investigated with Mplus, version 7.2, Muthén & Muthén (Los Angeles, CA, USA). The model assumes that EDs and study addiction are correlated constructs with independent negative effects on well-being, measured with such indicators as depression, anxiety, quality of life, and perceived stress. The model was investigated using a WLSMV estimator, due to the ordinal character of the EAT-26 and BStAS items. All categories of EAT items were also used to maximize the information, which somewhat differs from how items are scored for calculating diagnostic scores on the scale [41,59]. As is congruent with the results of the recent study [44], we tested a bi-factor model that comprised: 13 items, measuring three distinct eating pathology domains (social pressure, food awareness, food preoccupation, and bulimia) and the general factor of EDs loading on all items. Figure 2 shows details of the full specification of the model. The significance level for all analyses was set at α = 0.05.

2.5. Ethics Statement

The study was carried out in accordance with the Declaration of Helsinki. All gathered data were anonymous. Participants were notified of all relevant information concerning the study and their role in it, including the fact that they could withdraw at any point. Consent was required to participate in the study. No medical information was collected. The authors had no contact with the participants before, during, and after data collection.

3. Results

3.1. Descriptive Statistics and Correlation Analysis

Table 3. Mean scores, standard deviations, and percentages for study variables between genders.

Variables	Gender	N	M (SD)/%
Age	Women	184	23.09 (3.37)
	Men	71	22.94 (3.71)
Study addiction	Women	184	19.64 (5.67)
	Men	71	18.07 (5.85)
Anxiety	Women	184	11.55 (3.77)
	Men	71	9.00 (3.67)
Depression	Women	184	5.96 (3.36)
	Men	71	6.14 (3.20)
Quality of life	Women	184	36.69 (9.92)
	Men	71	38.42 (10.34)
Stress	Women	184	12.40 (2.83)
	Men	71	11.11 (2.97)

presents the mean scores, standard deviations, and percentages for the study variables.

Table 4. Spearman correlation coefficients between study addiction and all EAT-26 items.

	Garner et al. (1982) [41]	Rogoza et al. (2016) [44]	Study Addiction
Am terrified about being overweight.	Dieting	-	0.13 *
Avoid eating when I am hungry.	Oral Control	-	0.25 **
Find myself preoccupied with food.	Bulimia and Food Preoccupation	Food preoccupation	0.03
Have gone on eating binges where I feel that I may not be able to stop.	Bulimia and Food Preoccupation	Food preoccupation	0.11
Cut my food into small pieces.	Oral Control	-	0.08
Aware of the calorie content of foods that I eat.	Dieting	Food awareness	0.08
Particularly avoid food with a high carbohydrate content (i.e., bread, rice, potatoes, etc.).	Dieting	Food awareness	0.06
Feel that others would prefer if I ate more.	Oral Control	Social pressure	0.19 **
Vomit after I have eaten.	Bulimia and Food Preoccupation	Purging behaviors	0.10
Feel extremely guilty after eating.	Dieting	-	0.19 **
Am preoccupied with a desire to be thinner.	Dieting	-	0.16 *
Think about burning up calories when I exercise.	Dieting	-	0.17 **
Other people think that I am too thin.	Oral Control	Social pressure	0.19 **
Am preoccupied with the thought of having fat on my body.	Dieting	-	0.14 *
Take longer than others to eat my meals.	Oral Control	-	0.08
Avoid foods with sugar in them.	Dieting	Food awareness	0.01
Eat diet foods.	Dieting	Food awareness	0.12
Feel that food controls my life.	Bulimia and Food Preoccupation	Food preoccupation	0.07
Display self-control around food.	Oral Control	-	0.13 *
Feel that others pressure me to eat.	Oral Control	Social pressure	0.21 **
Give too much time and thought to food.	Bulimia and Food Preoccupation	Food preoccupation	0.17 **
Feel uncomfortable after eating sweets.	Dieting	-	0.14 *
Engage in dieting behavior.	Dieting	-	0.09
Like my stomach to be empty.	Dieting	-	0.18 **
Have the impulse to vomit after meals.	Bulimia and Food Preoccupation	Purging behaviors	0.24 **
Enjoy trying new rich foods.	Dieting	-	0.04

Note. * p < 0.05; ** p < 0.01. In bold are the significant correlations for items that were not included in the SEM model. All these items pertaining to the dieting or oral control factors are from the Garner et al. (1982) [41] structure of the EAT-26.

presents the correlation coefficients between particular EAT-26 items measuring the symptoms of EDs and study addiction. Not all the items are included in the current scale version used for SEM analysis. Notably, some of the items that were part of the original structure of the scale [59] showed significant associations with study addiction, particularly items representing symptoms, such as dieting (and reducing body weight by exercise) and oral control, that are typical for anorexia nervosa.

3.2. Prevalence

Table 5. Prevalence (number and percent) and the co-occurrence of study addiction, EDs, depression, and anxiety, based on the established cut-off scores. The prevalences are on the diagonal. Co-occurrence rates show the number of cases of the disorder in the column that are present among the cases of the disorder in the raw.

	Study Addiction (At Least 4 Symptoms)	Study Addiction (All 7 Symptoms)	Eating Disorders (Possible Disorder)	Depression (Possible, At Least Mild Disorder)	Depression Clinical (Clinically Significant Disorder)	Anxiety (Possible, At Least Mild Disorder)	Anxiety Clinical (Clinically Significant Disorder)
Study addictionpolythetic	63 (24.7%)	9 (14.3%)	10 (15.9%)	29 (46.0%)	13 (20.6%)	61 (96.8%)	55 (87.3%)
Study addiction monothetic	9 (100%)	9 (3.5%)	3 (33.3%)	7 (77.8%)	5 (55.6%)	9 (100%)	8 (88.9%)
Eating disorders	10 (37.0%)	3 (11.1%)	27 (10.6%)	15 (55.6%)	6 (22.2%)	27 (100%)	25 (92.6%)
Depression (>= 8)	29 (36.7%)	7 (8.9%)	15 (19.0%)	79 (31.0%)	26 (33.0%)	75 (95.0%)	60 (76.0%)
Depression (>= 11)	13 (50.0%)	5 (19.2%)	6 (23.1%)	26 (100.0%)	26 (10.2%)	24 (92.3%)	21 (80.8%)
Anxiety (>= 8)	61 (29.8%)	9 (4.4%)	27 (13.2%)	75 (36.6%)	24 (11.7%)	205 (80.4%)	130 (63.4%)
Anxiety (>= 11)	55 (42.3%)	8 (6.2%)	25 (19.2%)	60 (23.1%)	21 (16.2%)	130 (100%)	130 (51.0%)

Note. The bold type at the diagonal highlights the prevalence of the disorders among participants.

shows the prevalence and co-occurrence of study addiction, EDs, depression, and anxiety. Based on the polythetic cut-off score (high scores on at least 4 symptoms), 24.7% of students fulfilled the criteria for study addiction in the present sample, with 3.5% fulfilling the monothetic (high scores on all symptoms) cut-off score. About 10.6% of the sample showed results that indicate a high level of concern about EDs, based on the cut-off used for the full EAT-26 questionnaire. According to the instructions for the scale, these scores may indicate a need to consult a health professional to determine whether the problem requires clinical attention. Such an estimate may allow comparisons with previous studies because there is no cut-off for the shortened version [44]. Based on the 8+ cut-off for depression and anxiety, which showed optimal and stable sensitivity and specificity across the samples in the previous research, 31.0% and 80.4% of participants could be considered ‘possible cases’ of depression and anxiety, respectively. In addition, 10.2% of participants showed clinically significant levels of depression, while 51.0% showed clinically significant levels of anxiety.

Almost 16.0% of those fulfilling the polythetic cut-off and 33.3% of those fulfilling the monothetic cut-off for study addiction exhibited EDs. On the other hand, among those fulfilling the cut-off for EDs, 37.0% fulfilled the polythetic cut-off, and 11.1% fulfilled the monothetic cut-off for study addiction. Study addiction was four times more prevalent among those classified as possibly having EDs, in comparison to those classified as being in the low-risk group. EDs were almost two times more prevalent among those with at least four symptoms of study addiction than those without them, and more than three times more prevalent among those with all seven study addiction symptoms present.

Almost half of the students fulfilling the polythetic cut-off and about four out of five students fulfilling the monothetic cut-off for study addiction were depressed. On the other hand, among those who fulfilled the cut-off for depression, more than 15% fulfilled the polythetic cut-off, and 33.0% fulfilled the monothetic cut-off for study addiction. Among those who showed clinically significant levels of depression, 50% fulfilled the polythetic cut-off, and almost 20.0% fulfilled the monothetic cut-off for study addiction. Practically all the students who fulfilled the study addiction criteria showed at least mild anxiety levels, and nearly 90% showed clinically significant levels of anxiety. Generally, depression and anxiety showed similar rates among those fulfilling the cut-off for study addiction and those fulfilling it for EDs.

3.3. The SEM Model

Figure 3 presents a simplified representation of the SEM model with a bi-factor specification of EAT-26. The model including the factor of purging behaviors, with loadings only on two items, showed problems with identification. Therefore, a model without this factor was tested. The model showed the following fit indices: χ2/df = 333.815/189, CFI = 0.955, TLI = 0.945, RMSEA = 0.055. All items except for item eight were significantly loaded on the general score of EDs.

Table 6. Correlation coefficients between study addiction and EDs and the standardized direct effects of the general factors of EDs, social pressure, food awareness, food preoccupation, and study addiction on well-being. Note: SA = study addiction; G = general factor of EDs; SP = social pressure; FA = food awareness; FP = food preoccupation; WB = well-being.

			Estimate	S.E.	Est./S.E.	Two-Tailed p-Value
SA	WITH	G	0.12	0.10	2.07	0.038
SA	WITH	SP FA FP	0.34 −0.03 −0.05	0.07 0.10 0.11	4.76 −0.26 −0.45	<0.001 0.798 0.650
WB	ON	G SP FA FP SA	−0.46 −0.23 0.19 −0.05 −0.33	0.09 0.08 0.08 0.12 0.08	−4.83 −3.02 2.50 −0.40 −4.06	<0.001 0.003 0.012 0.692 <0.001

presents all correlation coefficients and all direct effects. The general factors of EDs and social pressure correlated significantly with study addiction. The general factors of EDs and study addiction showed the relatively strongest negative independent effects on well-being. Table 6 presents the square root of convergent validity and latent variable correlation.

4. Discussion

The objective of the study was to investigate the relationship between study addiction and eating disorders and their individual effects on well-being among music academies’ students. As expected, the results showed that, generally, study addiction was positively related to EDs (H1 is mostly substantiated). The relationship of study addiction to particular components of EDs needs to be interpreted within the framework of the bi-factor model investigated in the study. First of all, study addiction was positively related to the general factor of EDs, which also showed the relatively strongest negative association with well-being. It confirms that the general factor represents the most pathological aspects of EDs. Study addiction was positively related to the social pressure component, which may be explained by the common factors underlying both psychological problems, such as social anxiety and the fear of evaluation [13]. Notably, food awareness showed a positive relationship to well-being, which suggests that after controlling for the general pathological factor, this component may represent a healthy approach to food and diet. If so, then this would explain the finding of no association between food awareness and study addiction. Food preoccupation also showed no relationship with study addiction. It may show that when the general pathological factor of EDs and social pressures are controlled, study addiction is not related to a greater concern about food and eating in and of itself. This could indicate that factors other than the food-related psychological determinants common to EDs and study addiction link these two pathological classes of behavior. These may include rigid perfectionism and obsessive-compulsiveness, social anxiety, or emotional instability [13,60,61,62].

The co-occurrence analysis, based on validated cut-off scores, shows that study addiction and EDs are highly comorbid. Study addiction was four times more prevalent among those classified as possibly having EDs than among those classified in the low-risk group. EDs were almost two times more prevalent among those with at least four symptoms of study addiction than among those without them and were more than three times more prevalent among those with all seven study addiction symptoms present. These are somewhat lower estimates than those offered previously for the comorbidity of work addiction and EDs among professionally active women [13]. This can be explained by the fact that the clinical diagnosis of EDs was used in the previous study, which may have limited the number of cases due to the general underdiagnosis of these disorders in the general population. As a result, it may have somewhat overestimated the number of cases of compulsive overworking among those with EDs. However, as is congruent with this explanation, the number of cases of EDs among those with study addiction (almost 16%) in the present study was much higher (about 8 to 16 times) when compared to the prevalence rates of EDs in the general population (about 1–2% of anorexia and bulimia [12]). It is even more pronounced among those participants who showed all seven study addiction symptoms; in which case, 33% fulfilled the cut-off for eating disorders. It suggests up to 16 to 33 times higher rates of eating disorders among those most addicted to studying.

The correlation analysis with particular items showed that study addiction had the relatively strongest association with purging behaviors (item 25) that are indicative of bulimia, restrictive food intake, indicative of anorexia nervosa (item 2), and other items (1, 10, 11, 12) that are suggestive of the preoccupation with being thin that is typical for both disorders. This finding is consistent with research on work addiction that showed the association of bulimia and anorexia with workaholic behaviors [13]. It is also congruent with the theoretical distinction, supported by empirical data, that apart from those highly controlling individuals addicted to work, there are also more impulsive types [10,63]. Furthermore, this supports the theoretical assumption that study addiction and EDs may share common etiological pathways and clinical manifestations. These mainly include those factors contributing to maintaining a self-image of physical attractiveness and academic competence through highly controlling and perfectionistic behaviors. The relationship between study addiction and the social pressure component of EDs further supports the role of external validation in these behaviors. Family functioning, including critical and controlling parents, and family discord may contribute to both classes of psychopathology. Moreover, personality traits conducive to these behavioral patterns are shared in EDs and study addiction, including rigid perfectionism, narcissism, competitiveness, and general emotional instability [13,14,18,25,26,64].

As expected, study addiction had a unique contribution to a deterioration in well-being that was above and beyond EDs (H2 is substantiated). Study addiction, together with the general factor of EDs, had the relatively strongest negative effect on well-being. These results are congruent with previous studies showing that study addiction is related to high anxiety, depression, stress, and a low general quality of life [14]; (for an overview, see [9]). These findings also further support the notion that study addiction is a separate entity and is a psychopathological problem that contributes to impaired psychosocial functioning above and beyond other mental health problems and the shared underlying pathological processes.

The co-occurrence analysis based on the validated cut-off scores shows that study addiction and EDs are highly comorbid with depression and anxiety. Almost half of the students fulfilling the polythetic cut-off and nearly 80% of those fulfilling the monothetic cut-off for study addiction were depressed. Practically all students who fulfilled the study addiction criteria showed at least mild anxiety levels, and nearly 90% showed clinically significant anxiety levels. This finding is congruent with previous studies showing high levels of depression and anxiety among students from various universities, faculties, courses, and years of study, who show study addiction symptoms [14]. Generally, depression and anxiety showed similar rates among those fulfilling the cut-off for study addiction and those fulfilling it for EDs. Depression and anxiety were also highly prevalent in this sample of young musicians, which is congruent with a vast body of research documenting mental health problems among musicians [24].

The results of this study indicate that compulsive behaviors related to studying need to be recognized as a genuine psychopathology that is comorbid with EDs, as well as with a host of other mental health problems (e.g., anxiety, depression, social anxiety, and OCPD), and that may aggravate these disorders by independently contributing to harm and functioning impairments. Study addiction and EDs may share many risk factors, such as rigid perfectionism and obsessive-compulsiveness, anxiety and social anxiety, emotional instability, and difficult family dynamics (enmeshment, high parental control, expectations, and criticism), as well as early life traumas [12,13,27,33,65]. Therefore, some treatment approaches addressing these underlying causes may prove useful in managing both problems. Such interventions as those focused on family or cognitive-behavioral treatments may help to alleviate or treat symptoms of EDs and study addiction.

However, certain differences also need to be acknowledged. The independent effects of study addiction on well-being suggest that it has some idiosyncratic pathological mechanisms. These may include a preoccupation with academic performance and the related anxiety and test anxiety, obsession about success in school and work, neglecting social relationships, extreme study effort, and related mental and physical fatigue and health problems, such as musculoskeletal disorders and sleep problems. All these problems were associated with study addiction in previous studies [14,15,24,27]. Ignoring health problems was identified as an important symptom of study addiction [14].

In addition, without recognizing compulsive learning as a psychopathological problem, the treatment of EDs may have limited efficacy in those cases of their comorbidity because the underlying core and general tendencies (such as the need for external validation of perfectionistic standards [17]) are not addressed. In such cases, if only their manifestation in relation to eating behaviors is addressed, then compulsive studying may develop as a compensatory behavior in other domains of life. Moreover, high stress aggravates EDs, and study addiction was consistently related to high academic and general stress and anxiety [9]. Therefore, study addiction may exacerbate EDs symptoms or may affect their recurrence and worsen the prognosis [13].

Finally, when taking into account (i) the very high co-occurrence of work addiction and EDs among professionally active women [13], (ii) the relatively high prevalence of work addiction in working populations [9], and (iii) the broad consequences of work addiction in the workplace [66] and beyond [9], it is necessary to introduce the early prevention of both disorders. The current study corroborated the co-occurrence of study addiction and EDs in music academies’ students. Study addiction is related to work addiction after students graduate and enter the labor market [16]. Study addiction was already identified among high-school students [27]. Early prevention strategies should arguably focus on two areas: first, increasing the awareness of study addiction and EDs and their relationship; second, addressing their common risk factors, especially the undue need for control and perfection and an excessive preoccupation with self-image, by the development of healthy coping skills and a supportive social environment in both schools and universities.

4.1. Strengths and Limitations

In terms of the study’s strengths, to the authors’ knowledge, this is the first study to investigate the relationship between study addiction and EDs, and their unique contributions to well-being. Commonly used valid and reliable psychological questionnaires were applied. A specific sample of music academies’ students who present a high risk of EDs and study addiction (confirmed with estimates from this sample) increased the probability of capturing the associations between the investigated variables. Sophisticated statistical analyses using SEM allowed for disentangling of the associations and independent effects of variables, substantiating robust conclusions concerning the studied phenomena. The results are highly consistent with the theoretical frameworks of study addiction and EDs and with previous research [10,13]. They support the comorbidity of study addiction and EDs, and, more generally, the consistent associations between EDs and compulsive work-related behaviors in different sociodemographic groups. These results may guide prevention and treatment programs by accounting for the independent contribution of each of the behaviors to impaired well-being and psychosocial functioning.

As far as the limitations of the present study are concerned, the sample was relatively small, which somewhat reduces the power of the analyses. Taking into account the specificity of the group and the fact that the sample was predominantly female, the findings cannot be generalized to other samples without some reservations. Moreover, all the data collected were self-reported, which is related to certain typical weaknesses of such data, e.g., common method bias. Nevertheless, the observed effect sizes were large enough to be detected, and the discussed limitations did not seem to prevent identifying the most important and expected effects. The study was conducted during the COVID-19 pandemic, which could affect the subjects’ well-being due to the related stressors or diminished social support [67].

4.2. Conclusions and Future Research Directions

In general, it can be concluded that compulsive studying, conceptualized as study addiction, is positively related to EDs and both psychological problems have negative and independent effects on the well-being of young musicians. EDs may be 8 to 16 times more prevalent among the students of music academies addicted to studying (and up to 16 to 33 times more prevalent among the most addicted) than among the general population. Study addiction is four times more prevalent among those with EDs than those without them. About 80% of students showing all seven symptoms of study addiction showed the symptoms of at least mild depression, and more than half had clinically significant levels of depression. Almost 90% had clinically significant levels of anxiety. This further supports the close and consistent association and comorbidity of compulsive work-related behaviors and EDs across the socio-demographic spectrum. Experienced social pressure is an important common aspect of both classes of problematic behaviors, which may share numerous risk factors and clinical manifestations. The results further support the notion that study addiction is a separate entity and a psychopathological problem of an addictive nature that needs to be recognized, prevented, and treated. The study also identifies young musicians as being particularly vulnerable to study addiction. This shows an urgent need for prevention and awareness about the problem within this group. Without addressing compulsive studying and EDs, including their commonalities and idiosyncrasies, the prevention and treatment of these problems cannot be effective, and this will substantially affect the sustainability of education and work.

Future research should further investigate the theoretical and empirical associations between study addiction and EDs because both problems greatly burden society. Moreover, compulsive work-related behaviors and subclinical EDs are highly prevalent, including in the youngest age groups, and their comorbidity is still poorly understood. Future studies may consider the influence of personality factors, such as neuroticism and conscientiousness, in the relationship between study addiction, EDs, and well-being [68]. More representative samples from the general population and longitudinal studies, including detailed analyses of the common risk factors, clinical manifestations, and developmental trajectories, are warranted. Prevention and treatment programs designed to account for the comorbidity of these disorders are highly needed. Due to the interdisciplinary character of compulsive overworking, a greater interest among researchers from various disciplines, including the fields of psychology, psychiatry, economy, education, and sociology, and integrated efforts to incorporate the concept of detrimental and unprofitable hard work into the frameworks of the free-market economy is highly warranted.