Advertisement

Evaluation of the reliability and validity of the Italian version of the schema mode inventory for eating disorders: short form for adult with dysfunctional eating behavior

  • Giada PietrabissaEmail author
  • Alessandro Rossi
  • Susan Simpson
  • Andrea Tagliagambe
  • Venessa Bertuzzi
  • Clarissa Volpi
  • Giulia Fava
  • Gian Mauro Manzoni
  • Giovanni Gravina
  • Gianluca Castelnuovo
Open Access
Original Article
  • 117 Downloads

Abstract

Purpose

To examine the psychometric properties and the factorial structure of the Italian version of the schema mode inventory for eating disorders—short form (SMI-ED-SF) for adults with dysfunctional eating patterns.

Methods

649 participants (72.1% females) completed the 64-item Italian version of the SMI-ED-SF and the eating disorder examination questionnaire (EDE-Q) for measuring eating disorder symptoms. Psychometric testing included confirmatory factor analysis (CFA) and internal consistency. Multivariate analysis of covariance (MANCOVA) was also run to test statistical differences between the EDE-Q subscales on the SMI-ED-SF modes, while controlling for possible confounding variables.

Results

Factorial analysis confirmed the 16-factors structure for the SMI-ED-SF [S–Bχ2 (1832) = 3324.799; p < .001; RMSEA = 0.045; 90% CI 0.043–0.048; CFI = 0.880; SRMR = 0.066; χ2/df = 1.81; < 3]. Internal consistency was acceptable in all scales, with Cronbach’s Alpha coefficients ranging from 0.635 to 0.873.

Conclusions

The SMI-ED-SF represents a reliable and valid alternative to the long-form SMI-ED for assessment and conceptualization of schema modes in Italian adults with disordered eating habits. Its use is recommended for clinical and research purposes.

Level of evidence

Level V, descriptive study.

Keywords

Factorial structure Psychometric properties Schema therapy Modes Eating disorders 

Introduction

Eating disorders (EDs) are serious and difficult-to-treat mental illnesses, often showing ego-syntonic features and resistance to treatments. Epidemiological studies usually underestimate the occurrence of EDs in the general population, since individuals are rarely aware of their illness and only occasionally refer to mental health care [1]. Many factors conspire to impede the treatment of EDs, including entrenched thinking, ambivalence about change, avoidant and perfectionistic personality traits, and comorbidity of trauma symptoms [2, 3].

Cognitive behavioural therapy (CBT) is widely recognized as the treatment of choice for adults with EDs [4]. Despite the widespread support for its efficacy [5, 6], therapy is often hampered by the well-known phenomenon of dropout [5].

Schema therapy (ST) is an integrative and multi-modal approach developed to address deeper levels of cognition and entrenched behaviours that do not respond to first-line treatments [7].

The goal of the ST treatment for EDs is to enable core psychological (and physiological) needs to be met [8], and to bring about change in eating habits by breaking enduring and self-defeating patterns of thinking, feeling, and behaving that typically begin early in life as a result of the interaction between temperament and unmet core emotional needs—referred to as early maladaptive schemas (EMS)—whilst developing healthy coping mechanisms [9, 10]. Indeed, research, suggests that those who suffer from EDs experience significantly higher levels of maladaptive modes than community samples [11, 12]. The ST treatment for EDs includes recognizing and challenging Internalized Critic Modes, re-parenting to heal the vulnerable child mode, and bypassing the resulting coping modes that are linked to the over-evaluation of shape, weight, and self-starvation. Limits are also set on Angry and Impulsive Child Modes that drive a self-destructive “acting out” of needs (i.e., bingeing). Cognitive and behavioural techniques are considered core aspects of ST, but the model gives equal weight to emotion-focused work and experiential techniques, in addition to the basic healing components of the therapeutic relationship. As with CBT, ST is structured, systematic and specific, following a sequence of assessment and treatment procedures. However, the pace and emphasis on aspects of treatment may vary depending on the individual needs.

To facilitate more precise measurement of mode states within the ED population, the schema mode inventory for eating disorders (SMI-ED) was recently developed, showing adequate validity and reliability [13]. Given the large number of items in the SMI-ED (n = 190)—which make it cumbersome for everyday clinical practice—the purpose of the present study was to develop a shortened Italian version of the SMI-ED, to assess its psychometric proprieties, and to determine the internal reliability of its subscales. The relationship between ED symptoms (restraint, binge eating and purging) and schema modes was also explored.

Materials and methods

Participants

The sample comprised 649 participants [181 males (27.9%) and 468 females (72.1%)] aged from 18 to 91 years (mean = 40.66, SD = 18.27). The study was open to individuals (1) aged over 18 years old, (2) who were Italian-speaking and that (3) signed digital informed consent to participate in the study. Exclusion criteria included the inability to complete the questionnaire due to visual or cognitive impairments. Participation was voluntary, and respondents did not receive remuneration.

Sample size calculation

Sample size calculation was based on two recommendations: first, that 500 or more observations can be considered “very good” for conducting a confirmatory factor analyses [14]; second, using the rule of ten subjects per item [15].

Measures

Demographics Information including age, gender, education, relationships, and employment status were collected.

Biomedical data Data on height and weight were registered and BMI was calculated as weight (kg) divided by height squared (m2). Participants were also asked to report on the presence of existing diagnosis of eating disorders through a multiple choice question (“Have you ever been diagnosed with one of the following eating disorder?”) [16].

The Italian version of the Eating Disorder Examination Questionnaire (EDE-Q) [17] The EDE-Q 6.0 is a 28-item self-report measure of ED attitudes psychopathology and behaviours in both community and clinical populations. The questions concern the frequency of key behavioural features of EDs in which the person engages over the preceding 28 days. The questionnaire is scored on a 7-point Likert scale (0–6), rated using four subscales (restraint—R; eating concern—EC; shape concern—SC; and weight concern—WC) and a global score.

The EDE-Q has generally received support as an adequately reliable and valid measure of eating-related pathology [13]. Similarly, in the present sample, the dimensions of the EDE-Q have demonstrated acceptable internal consistency (R-α = 0.804; EC-α = 0.822; SC-α = 0.900; WC-α = 0.800; General/Total-α = 0.944).

The Italian version of schema mode inventory for eating disorders—short form (SMI-ED-SF) The item-pool (n = 64) for the new SMI-ED-SF was first created independently by two clinicians/researchers specialized both in ST and in the treatment of ED (authors GP and SS), who listed the items under each of the 16 modes in order of relevance in observance of the ST conceptualization for EDs.

Simultaneously, and blinded from the other authors, a third researcher (not specialized in ST; author AR) identified those items showing higher factor loading for each dimension of the original SMI-ED [13]. Conclusions from the authors were matched and discussed until agreement on the final set of items for the SMI-ED-SF was reached. Four items (three general, and one EDs-specific statement—where applicable) per mode were retained—thus to overcome the limitation of the previous version of the tool—where the number of items was highly heterogeneous between modes.

The SMI-ED is a 190-item self-report questionnaire with sixteen different modes clustered thematically: (A) five innate child modes (1. vulnerable child—VC, 2. angry child—AC, 3. enraged child—EC, 4. impulsive child—IC and 5. undisciplined child—UC); (B) two maladaptive (internalized/introject) modes (6. punitive mode—PM and 7. demanding mode—DM); (C) seven maladaptive coping modes (8. compliant surrenderer—CS, 9. helpless surrenderer—DS, 10. detached protector—Det.P, 11. detached self-soother—Det.SS, 12. self-aggrandizer—SA, 13. bully and attack—BA 14. eating disorder overcontroller—EDO); and (D) two healthy factors (15. happy child—HC and 16. healthy adult—HA). Notably, two modes (IC and EC) only included items retrieved from the original version of the SMI [18], while the HS and the EDO modes exclusively consisted of new ED-specific statements.

The SMI-ED revealed acceptable internal consistency, with Cronbach’s alpha coefficients ranging from 0.807 (Det.SS) to 0.976 (PM) across subscales (meanα-factors = 0.914; SDα-factors = 0.048).

Contrary to its full-length version—in which the number of items between scales varies from 5 (DS) to 20 (VC)—a fixed list of four statements was ensured for each of the SMI-ED-SF subscales (n = 16). Specifically, except for those modes only including either items retrieved from the original SMI or consisting of EDs-specific statements, the remaining subscales comprised three general statements and one item representative of the ED population.

Consistent with the previous versions of the tool [13, 18], items were scored on a six-point Likert scale ranging from 0 (“never or hardly ever”) to 5 (“all of the time”) and the score for each mode was computed dividing the sum scores by the number of items in each subscale. The higher the score, the more frequent were the manifestations of the modes.

Translation and cross-cultural adaptation

The SMI-ED-SF was independently translated from the original English version into Italian by two bilingual experts in the field, with one of them also having good knowledge of the measure. Any inconsistencies were revised and adjusted by a third investigator independent from the study using culturally and clinically fitting expressions. Also, to ensure conceptual equivalence between translations, a blind back translation of the Italian version of the SMI-ED-SF into English was conducted by an independent bilingual translator. Prior to the main study, the approved Italian version of the questionnaire was trialed with a random sample of 15 patients with EDs and 23 non-clinical participants, to assess item comprehensibility for the target population. No further adjustment was required.

Procedure

This study was completed entirely online, hosted by the questionnaire tool Qualtrics. Recruitment advertisements included a link placed on the main social networks (i.e., Facebook, Twitter) and websites of various local clinical centers specialized in the treatment and rehabilitation of EDs in Italy. In addition, flyers were placed around University campuses and in clinical waiting rooms of local ED services. The initial page contained a detailed description of the study, inclusion, and exclusion criteria along with any potential risks that may occur as a result of participation. Subjects were then asked to acknowledge they had read the terms and conditions and were aware of any potential risks by signing an informed consent form. Following informed consent, participants were asked to report demographic information and to answer the study questionnaires. After completing the survey, they were given access to a debriefing page of the study aims, and methodology, and received contact details for support services.

Statistical analyses

To test the factorial structural model of the SMI-ED-SF a Confirmatory Factor Analysis (CFA) was performed using ‘lavaan’ package [19, 20] for R software (R-core project [21, 22]). All the other statistical analysis were carried out with SPSS software (version 20.0, SPSS Inc., Bologna, Italy) [23].

As reported in Table 2, items’ descriptive statistics showed a non-normal distribution of some indicators. Therefore, in line with the previous study [13], the robust maximum likelihood method (MLM) [24, 25, 26, 27] was chosen as estimator for the CFA. The MLM is a robust variant of the Maximum likelihood [27] that provides robust standard errors and is also referred to as the Satorra–Bentler Chi square (S–Bχ2) [19, 28, 29] to assess the model fit. Other fit indexes used to assess the model fit [30] were: the root mean square error of approximation (RMSEA) [31, 32], the comparative fit index (CFI) [33], and the standard root mean square residual (SRMR) [27], and the ratio of S–Bχ2 to the degrees of freedom (df) [34]. A S–Bχ2 test non-significant is desirable [35]. The RMSEA expresses fit per degrees of freedom of the model, with values lower than 0.08 suggesting an acceptable model fit [36] and values below 0.05 indicating a good fit [37]. The CFI designates the amount of variance and covariance accounted by the model compared with a baseline model, with values between 0.90 and 0.95 considered an acceptable fit [38, 39], and values > 0.95 indicating a good fit [36].

However, Kenny and McCoach mathematically demonstrate that a higher number of indicators analyzed negatively affects this fit index [40, 41, 42]. The SRMR derives from the residual correlation matrix and represents the average discrepancy between the correlations observed in the input matrix and those predicted by the model [27, 38]. A cutoff value higher than 0.08 is considered good [26, 36]. Also, the χ2/df ratio is considered as an easily computable measure of fit [26, 43], and a χ2/df ratio value of 3 or less indicates good fit [44, 45, 46, 47].

The Cronbach’s alpha coefficient was used as measure of internal consistency for each SMI-ED-SF subscale—and values higher than 0.7 are deemed acceptable [48]. However, considering the differences in the magnitude of SMI-ED-SF’s factor loadings, Cronbach’s alpha was supported by Raykov’s maximal reliability (MR) [49] and the Bentler’s “Model-Based Internal Consistency Coefficient” (MBICC) [50]. These two indices were, respectively, chosen as measures of internal consistency of each single factor and multidimensional (overall) reliability: values higher than 0.6 suggest good reliability [51].

In addition, a MANCOVA was conducted to assess for possible statistical differences between the disordered eating subgroups simultaneously, on the SMI-ED-SF subscales, while adjusting for differences in age and gender.

Results

Sample characteristics

Participants’ self-reported BMI ranged from 13.71 to 65.31 (mean = 28.26; SD = 10.54), with 15.7% of the sample having a BMI below 18.5 and 38.4% of the respondents having a BMI above 30.1.

Of 649 participants, 46 self-reported a diagnosis of anorexia nervosa (AN), 31 were diagnosed with bulimia nervosa (BN), 64 suffered from binge eating disorder (BED), and 58 declared eating disorders not otherwise specified (EDNOS)—while the remaining 450 participants did not self-report a diagnosis of EDs. Descriptive statistics are presented in Table 1.

Table 1

Descriptive statistics for the EDE-Q 6.0 subscales

 

Overall sample (n = 649)

AN (n = 46)

BN (n = 31)

BED (n = 64)

EDNOS (n = 58)

No diagnosis (n = 450)

Statisticsa

p value

Weight—in kg (mean; SD)

82.21

30.94

50.69

13.55

75.66

24.87

107.32

23.56

103.67

29.82

79.53

29.56

H = 143.227

< 0.001

Height—in m (mean; SD)

1.67

0.08

1.64

0.08

1.66

0.07

1.66

0.08

1.64

0.11

1.67

0.08

H = 10.276

0.036

BMI (mean; SD)

28.26

10.54

18.23

3.14

27.13

9.37

38.17

7.40

37.27

10.41

26.81

9.77

H = 131.005

< 0.001

Age (mean; SD)

40.66

18.27

26.46

9.11

40.97

10.59

51.27

17.86

51.59

16.01

39.17

17.98

H = 76.277

< 0.001

Gender (n; %)

            

V = 0.187

< 0.001

 Male

181

27.9%

3

6.5%

1

3.2%

19

29.7%

17

29.3

141

31.3%

  

 Female

468

72.1%

43

93.5%

30

96.8%

45

70.3%

41

70.7

309

68.7%

  

Relationships status (n; %)

            

V = 0.149

< 0.001

 Single/never married

357

55.0%

39

84.8%

18

58.1%

24

37.5%

15

25.9%

261

58.0%

  

 In a de-facto relationship

42

6.5%

2

4.3%

3

9.7%

6

9.4%

2

3.4%

29

6.4%

  

 Married

175

27.0%

4

8.7%

5

16.1%

25

39.1%

26

44.8%

115

25.6%

  

 Separated/divorced

52

8.0%

0

0.0%

4

12.9%

6

9.4%

10

17.2%

32

7.1%

  

 Widowed

23

3.5%

1

2.2%

1

3.2%

3

4.7%

5

8.6%

13

2.9%

  

Education status (n; %)

            

V = 0.114

0.006

 Elementary school

21

3.2%

0

0.0%

0

0.0%

3

4.7%

6

10.3%

12

2.7%

  

 Middle school

114

17.6%

11

23.9%

6

19.4%

18

28.1%

14

24.1%

65

14.4%

  

 High school

381

58.7%

30

65.2%

16

51.6%

37

57.8%

29

50.0%

269

59.8%

  

 Bachelor’s degree

127

19.6%

5

10.9%

9

29.0%

6

9.4%

8

13.8%

99

22.0%

  

 Postgraduate degree/PhD

6

0.9%

0

0.0%

0

0.0%

0

0.0%

1

1.7%

5

1.1%

  

Employment status (n; %)

            

V = 0.189

< 0.001

 Student

253

39.0%

30

65.2%

8

25.8%

7

10.9%

9

15.5%

199

44.2%

  

 Employees

149

23.0%

3

6.5%

10

32.3%

17

26.6%

14

24.1%

105

23.3%

  

 Freelancers

55

8.5%

2

4.3%

3

9.7%

6

9.4%

4

6.9%

40

8.9%

  

 Homemaker

36

5.5%

0

0.0%

3

9.7%

10

15.6%

6

10.3%

17

3.8%

  

 Unemployed

63

9.7%

10

21.7%

3

9.7%

8

12.5%

10

17.2%

32

7.1%

  

 Retried

93

14.3%

1

2.2%

4

12.9%

16

25.0%

15

25.9%

57

12.7%

  

AN Anorexia nervosa, BN bulimia nervosa, BED binge-eating disorder, EDNOS eating disorder not otherwise specified

aDue to variable sample sizes from each group and the non-normal distributions of some variables, Cramér’s V and Kruskal–Wallis’s tests were performed to assess associations across socio-demographics variables

Structural validity

Item analysis revealed a non-perfect normal distribution, with Kolmogorov–Smirnov and Shapiro–Wilk tests being significant (p < .001). Skewness ranged between − 1.18 and 2.76 (meansk = 0.79, SDsk = 0.81), and kurtosis ranged between − 1.03 and 8.09 (meank = 0.64, SDk = 2.01) (Table 2).

Table 2

Factor loading of the SMI-ED-SF items

Factor

Item

Item descriptive statistics

CFA

Mean

Median

SD

Sk

K

%Min (%)

%Max (%)

λ

R 2

VC—vulnerable child

 

Item1

1.568

1

1.294

0.600

− 0.284

24.5

2.3

0.707

0.500

 

Item2

1.156

1

1.391

1.134

0.432

45.6

3.7

0.643

0.414

 

Item3

1.017

1

1.192

1.158

0.804

45.1

1.3

0.869

0.754

 

Item4

1.173

1

1.215

0.959

0.359

37.7

1.3

0.830

0.689

AC—angry child

 

Item5

1.188

1

1.368

1.026

0.206

44.0

3.0

0.684

0.468

 

Item6

1.406

1

1.257

0.752

0.106

29.1

2.5

0.693

0.481

 

Item7

1.585

1

1.392

0.648

− 0.362

27.5

3.9

0.848

0.719

 

Item8

0.875

0

1.162

1.488

1.958

51.4

1.9

0.698

0.487

EC—enraged child

 

Item9

0.447

0

0.951

2.593

6.908

75.0

0.8

0.682

0.465

 

Item10

0.928

1

1.147

1.316

1.362

47.8

1.1

0.873

0.762

 

Item11

0.671

0

1.034

1.822

3.409

60.4

1.0

0.901

0.811

 

Item12

1.259

1

1.141

0.989

0.943

28.4

1.7

0.731

0.535

IC—impulsive child

 

Item13

1.531

1

1.264

0.801

0.052

21.1

2.2

0.753

0.568

 

Item14

0.917

1

1.182

1.503

1.891

48.0

1.5

0.779

0.607

 

Item15

1.217

1

1.199

1.062

0.744

32.1

1.6

0.820

0.673

 

Item16

1.831

2

1.361

0.510

− 0.373

18.2

4.9

0.591

0.349

UC—undisciplined child

 

Item17

1.261

1

1.286

1.014

0.438

34.5

2.5

0.782

0.611

 

Item18

1.396

1

1.273

0.697

− 0.271

30.2

1.4

0.851

0.724

 

Item19

1.089

1

1.196

1.141

0.888

40.1

1.4

0.557

0.310

 

Item20

1.535

1

1.353

0.766

− 0.121

25.8

3.6

0.609

0.371

HC—happy child

 

Item21

3.136

3

1.393

− 0.425

− 0.623

4.5

18.8

0.675

0.456

 

Item22

2.911

3

1.367

− 0.196

− 0.708

4.7

14.2

0.339

0.115

 

Item23

2.791

3

1.345

− 0.221

− 0.616

6.0

10.2

0.864

0.746

 

Item24

2.894

3

1.257

− 0.337

− 0.479

4.0

8.2

0.780

0.608

PM—punitive mode

 

Item25

0.740

0

1.137

1.713

2.412

59.5

0.9

0.650

0.422

 

Item26

0.577

0

1.007

2.199

5.198

65.8

1.2

0.767

0.588

 

Item27

0.435

0

0.958

2.762

8.028

75.8

1.2

0.839

0.704

 

Item28

0.445

0

0.961

2.737

8.087

75.0

1.5

0.861

0.742

DM—demanding mode

 

Item29

1.426

1

1.463

0.943

0.009

34.3

5.5

0.703

0.494

 

Item30

1.145

1

1.339

1.259

0.974

41.8

3.9

0.712

0.507

 

Item31

2.580

3

1.501

− 0.034

− 0.964

9.6

12.4

0.396

0.157

 

Item32

2.699

3

1.484

− 0.044

− 1.010

7.0

13.8

0.461

0.212

HA—healthy adult

 

Item33

3.808

4

1.186

− 1.179

1.293

2.5

32.1

0.669

0.447

 

Item34

2.938

3

1.296

− 0.330

− 0.400

4.8

11.5

0.605

0.366

 

Item35

3.270

4

1.398

− 0.622

− 0.411

4.6

20.8

0.794

0.630

 

Item36

3.651

4

1.193

− 0.868

0.495

2.0

27.3

0.732

0.536

CS—compliant surrender

 

Item37

2.305

2

1.385

0.117

− 0.776

10.5

6.2

0.560

0.314

 

Item38

1.553

1

1.319

0.612

− 0.354

25.9

2.5

0.775

0.601

 

Item39

2.137

2

1.488

0.195

− 0.949

16.8

6.3

0.715

0.511

 

Item40

1.223

1

1.446

1.062

0.144

44.7

4.0

0.724

0.524

Det.P—detached protector

 

Item41

1.526

1

1.440

0.759

− 0.336

30.4

4.2

0.749

0.561

 

Item42

1.503

1

1.458

0.778

− 0.346

32.1

4.5

0.650

0.423

 

Item43

0.836

0

1.171

1.633

2.475

54.0

2.2

0.730

0.533

 

Item44

0.679

0

1.129

1.815

2.809

64.9

1.1

0.674

0.454

Det.SS—detached self-soother

 

Item45

1.714

1

1.611

0.615

− 0.815

30.8

7.3

0.659

0.434

 

Item46

2.124

2

1.499

0.236

− 0.952

18.8

6.6

0.744

0.553

 

Item47

1.368

1

1.447

0.909

− 0.103

37.6

4.3

0.631

0.398

 

Item48

2.371

2

1.445

0.143

− 0.844

10.4

9.0

0.676

0.457

SA—self-aggrandizer

 

Item49

2.224

2

1.285

0.152

− 0.615

9.0

4.0

0.511

0.262

 

Item50

1.755

2

1.476

0.556

− 0.628

24.4

5.6

0.365

0.133

 

Item51

1.103

0

1.495

1.195

0.271

54.3

4.4

0.601

0.362

 

Item52

2.133

2

1.446

0.128

− 0.935

16.3

5.1

0.543

0.294

BA—bully and attack

 

Item53

0.715

0

1.112

1.772

2.705

59.8

1.0

0.710

0.504

 

Item54

0.675

0

0.945

1.460

1.848

57.2

0.3

0.610

0.372

 

Item55

0.962

0

1.266

1.360

1.185

50.9

2.0

0.729

0.532

 

Item56

1.071

1

1.307

1.284

1.033

45.5

3.0

0.628

0.395

HS—helpless surrenderer

 

Item57

2.079

2

1.528

0.328

− 0.902

17.7

8.2

0.599

0.359

 

Item58

2.764

3

1.482

− 0.093

− 0.915

7.3

15.7

0.717

0.514

 

Item59

2.083

2

1.470

0.337

− 0.732

16.4

7.7

0.614

0.377

 

Item60

1.904

2

1.515

0.443

− 0.791

21.7

7.0

0.800

0.641

EDO—eating disorder overcontroller

 

Item61

1.355

1

1.517

0.887

− 0.364

42.0

4.2

0.753

0.567

 

Item62

1.870

2

1.678

0.463

− 1.027

29.7

9.8

0.788

0.621

 

Item63

1.284

1

1.548

0.993

− 0.210

46.5

5.1

0.876

0.768

 

Item64

0.974

0

1.361

1.370

0.955

55.0

2.8

0.776

0.603

The scale revealed acceptable internal consistency, with Cronbach’s Alpha coefficients ranging from 0.635 (SA) to 0.873 (EC and EDO); meanα-factors = 0.787; SDα-factors = 0.06. Furthermore, the Raykov’s MR ranged from 0.664 (SA) to 0.905 (EC); meanMR-factors = 0.816 SDMR-factors = 0.06—suggesting each scale to be adequately reliable (Table 3). Also, the Bentler’s MBICC was equal to 0.951—indicating a good overall reliability of the scale

In line with the SMI-ED validation study [13], results from the CFA suggested an acceptable 16-correlated-factors solution for the SMI-ED-SF, despite not all the model’s fit indexes reaching the desired value [36]. Indeed, the Satorra–Bentler Chi square model for fit was statistically significant [S–Bχ2 (1832) = 3324.799; p < .001] and the CFI value did not achieve the threshold (CFI > 0.90 [38, 39]: CFI = 0.880). However, the RMSEA showed a good approximation fit of the model to the data [RMSEA = 0.045 (90% CI from 0.043 to 0.048), p(RMSEA < 0.05) = 1], and the SRMR also accounted for the goodness of the model (SRMR = 0.066 [36]). By dividing the χ2 for the degrees of freedom (df) of the model [34, 36], the model further resulted acceptable (χ2/df = 1.81; < 3) [26].

As reported in Table 2, each item loaded significantly on its associated factor (p < .001), meanloadings = 0.698; SDloadings = 0.122; ranging from 0.339 (item#22) to 0.901 (item#11). Correlations between the 16 factors ranged from |0.065| to |0.654|; meanr-factors = 0.238; SDr-factors = 0.297 (Table 3).

Table 3

Mean values, standard deviations, correlations between subscales of the SMI-ED-SF and reliability of each subscale (Cronbach’s Alpha and MR)

  

M

SD

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Alpha

MR

1

VC

1.24

1.05

               

0.829

0.875

2

AC

1.26

1.04

0.617

              

0.813

0.842

3

EC

0.82

0.91

0.363

0.654

             

0.873

0.905

4

IC

1.37

1.01

0.419

0.523

0.607

            

0.823

0.845

5

UC

1.32

0.99

0.461

0.462

0.339

0.502

           

0.786

0.839

6

HC

2.93

1.01

− 0.479

− 0.351

− 0.167

− 0.142

− 0.251

          

0.748

0.845

7

PM

0.55

0.85

0.541

0.440

0.226

0.277

0.347

− 0.375

         

0.856

0.881

8

DM

1.96

1.05

0.384

0.317

0.211

0.189

0.118

− 0.152

0.416

        

0.690

0.711

9

HA

3.42

0.99

− 0.426

− 0.323

− 0.196

− 0.181

− 0.288

0.528

− 0.429

− 0.065*

       

0.793

0.809

10

CS

1.81

1.10

0.423

0.295

0.128

0.192

0.390

− 0.281

0.450

0.351

− 0.249

      

0.784

0.804

11

Det.P

1.14

1.03

0.508

0.478

0.305

0.316

0.423

− 0.380

0.526

0.312

− 0.388

0.544

     

0.785

0.799

12

Det.SS

1.89

1.15

0.501

0.533

0.305

0.343

0.333

− 0.262

0.390

0.366

− 0.222

0.357

0.425

    

0.769

0.778

13

SA

1.80

0.95

0.381

0.446

0.376

0.368

0.292

− 0.252

0.261

0.373

− 0.148

0.231

0.367

0.440

   

0.635

0.664

14

BA

0.87

0.90

0.333

0.487

0.442

0.345

0.296

− 0.266

0.304

0.276

− 0.229

0.226

0.390

0.361

0.542

  

0.753

0.772

15

HS

2.20

1.16

0.539

0.568

0.403

0.425

0.467

− 0.312

0.369

0.342

− 0.273

0.442

0.430

0.546

0.503

0.449

 

0.778

0.800

16

EDO

1.37

1.30

0.446

0.387

0.225

0.311

0.271

− 0.173

0.362

0.385

− 0.180

0.275

0.336

0.520

0.393

0.291

0.374

0.873

0.886

All correlations are significant at p < .001, except for *(p > .05; ns.)

VC Vulnerable child, AC angry child, EC enraged child, IC impulsive child, UC undisciplined child, HC happy child, PM punitive mode, DM demanding mode, HA healthy adult, CS compliant surrender, Det.P detached protector, Det.SS detached self-soother, SA self-aggrandizer, BA bully and attack, HS helpless surrenderer, EDO eating disorder overcontroller, MR maximum residual

Concurrent validity: correlation between SMI-ED-SF factors and eating disorder variables

Most SMI-ED-SF factors were significantly associated (ranging from |0.088| to |0.855|) with the EDE-Q subscales and ED symptoms (Table 4). In line with the original SMI-ED the adaptive modes (happy child and healthy adult) were negatively correlated with all the ED variables.

Table 4

Correlations between SMI-ED-SF subscales

 

Times OE

Times bingeing

Days bingeing

Vomit

Laxatives

Exercise

Restraint

Eating concerns

Shape concerns

Weight concerns

Global

VC

0.291

0.477

0.377

0.228

0.333

0.377

0.368

0.530

0.449

0.506

0.536

AC

0.404

0.569

0.456

0.281

0.357

0.414

0.272

0.424

0.392

0.413

0.425

EC

0.504

0.572

0.538

0.366

0.320

0.290

0.144

0.219

0.178

0.198

0.205

IC

0.835

0.788

0.855

0.761

0.450

0.449

0.203

0.269

0.243

0.261

0.282

UC

0.367

0.452

0.451

0.360

0.814

0.853

0.157

0.385

0.322

0.316

0.330

HC

− 0.074§

− 0.234

− 0.145

− 0.022§

− 0.195

− 0.196

− 0.238

− 0.313

− 0.273

− 0.326

− 0.352

PM

0.168

0.362

0.260

0.120

0.230

0.297

0.316

0.437

0.401

0.419

0.441

DM

0.088*

0.230

0.142

0.152

0.070§

0.068§

0.321

0.356

0.351

0.356

0.407

HA

− 0.091*

− 0.303

− 0.209

− 0.011

− 0.192

− 0.206

− 0.202

− 0.294

− 0.229

− 0.283

− 0.290

CS

0.078*

0.264

0.186

0.100*

0.256

0.302

0.249

0.354

0.358

0.332

0.375

Det.P

0.200

0.394

0.318

0.126

0.316

0.341

0.286

0.407

0.356

0.391

0.412

Det.SS

0.258

0.351

0.300

0.211

0.249

0.296

0.358

0.471

0.445

0.466

0.500

SA

0.324

0.338

0.287

0.249

0.231

0.295

0.278

0.346

0.329

0.331

0.371

BA

0.302

0.327

0.313

0.190

0.261

0.289

0.230

0.317

0.215

0.240

0.260

HS

0.332

0.432

0.363

0.263

0.349

0.391

0.301

0.453

0.404

0.407

0.451

EDO

0.210

0.333

0.266

0.214

0.218

0.249

0.490

0.527

0.557

0.563

0.613

All correlations are significant at p < .001, except for * (p < .025) and §(p > .05; ns)

Times OE Over the past 28 days, how many TIMES have you eaten what other people would regard as an unusually large amount of food (given the circumstances)?

Times bingeing On how many of these TIMES did you have a sense of having lost control over your eating (at the time that you were eating)?

Days bingeing Over the past 28 days, on how many DAYS have such episodes of overeating occurred (i.e., you have eaten an unusually large amount of food and have had a sense of loss of control at the time)?

Vomit Over the past 28 days, how many TIMES have you made yourself sick (vomit) as a means of controlling your shape or weight?

Laxatives Over the past 28 days, how many TIMES have you taken laxatives as a means of controlling your shape or weight?

Exercise Over the past 28 days, how many TIMES have you exercised in a “driven” or “compulsive” way as a means of controlling your weight, shape or amount of fat, or to burn off calories?

VC Vulnerable child, AC angry child, EC enraged child, IC impulsive child, UC undisciplined child, HC happy child, PM punitive mode, DM demanding mode, HA healthy adult, CS compliant surrender, Det.P detached protector, Det.SS detached self-soother, SA self-aggrandizer, BA bully and attack, HS helpless surrenderer, EDO eating disorder overcontroller

EDs symptoms

EDE-Q subscales

Correlation between SMI-ED-SF factors, gender, age, and BMI

Most of the SMI-ED-SF factors were not significantly associated with gender, age and BMI (Table 5). Regarding gender, significant associations ranged from |0.084| (angry child) to |0.235| (vulnerable child). Considering age, statistically significant correlations ranged from |0.079| (happy child) to |0.197| (helpless surrenderer). Also, significant correlations between the SMI-ED-SF factors and BMI ranged from |0.099| (self-aggrandizer) and |0.168| (eating disorder overcontroller).

Table 5

Correlations between SMI-ED-SF subscales, gender, age, and BMI across EDs

 

Gender

Age

BMI

VC

0.235***

− 0.154***

− 0.155**

AC

0.084*

− 0.074

− 0.024

EC

0.022

− 0.128**

− 0.110*

IC

0.064

− 0.063

− 0.115*

UC

0.034

0.058

0.050

HC

− 0.063

0.079*

0.074

PM

0.038

0.011

− 0.019

DM

0.009

− 0.050

− 0.049

HA

− 0.108**

0.109**

0.109*

CS

0.071

0.064

− 0.007

Det.P

0.005

− 0.011

0.009

Det.SS

0.121**

− 0.111**

− 0.002

SA

− 0.030

− 0.167***

− 0.099*

BA

− 0.120**

− 0.119**

− 0.070

HS

0.229***

− 0.197***

− 0.164***

EDO

0.124**

0.122**

0.168**

Associations between SMI-ED-SF subscales and gender were computed with point-biserial (polychoric) correlations; whereas, associations regarding SMI-ED-SF subscales, age and BMI were calculated on Pearson’s product–moment correlation

*p < .050; **p < .010; ***p < .001

Mode scores across disordered eating subscales

While controlling for age and gender as possible confounding variables, the MANCOVA revealed a significant difference between the presence of a self-reported diagnosis of ED and most of the SMI-ED-SF subscales: Wilks’s Λ = 0.638, F = 4.587, p < .001, partial η2 = 106. No differences emerged between ED diagnoses and the enraged child mode measured by the SMI-ED-SF. Also, to test differences between groups within the SMI-ED-SF subscales, ANCOVAs with focused contrasts were conducted for each dependent variable (Table 6).

Table 6

Mean (SD) for the ED diagnosis resulting from the MANCOVA

 

AN (n = 46)

BN (n = 31)

BED (n = 64)

EDNOS (n = 58)

No diagnosis (n = 450)

F (4637)

η p 2

Focused contrasta

Mean (SD)

Mean (SD)

Mean (SD)

Mean (SD)

Mean (SD)

VC

2.51 (1.23)

2.30 (1.27)

1.50 (1.07)

1.37 (1.11)

0.97 (0.82)

36.909

0.190

1 > 3**; 1 > 4***; 1 > 5; 2 > 3**; 2 > 4**; 2 > 5; 3 > 5; 4 > 5

AC

2.04 (1.01)

2.13 (1.13)

1.60 (1.14)

1.38 (1.21)

1.04 (0.90)

20.236

0.114

1 > 5; 2 > 4**; 2 > 5; 3 > 5; 4 > 5**

EC

1.05 (1.02)

1.05 (0.93)

0.94 (1.03)

0.75 (0.90)

0.76 (0.84)

2.317§

0.014

3 > 5*

IC

1.78 (1.13)

2.07 (1.14)

1.77 (1.20)

1.35 (1.04)

1.22 (0.90)

11.038

0.065

1 > 5; 2 > 4; 2 > 5; 3 > 4*; 3 > 5

UC

1.82 (1.25)

1.79 (0.87)

1.70 (1.10)

1.48 (0.86)

1.14 (0.91)

11.100

0.066

1 > 5; 2 > 5; 3 > 5; 4 > 5**

HC

2.16 (1.02)

2.16 (0.64)

2.66 (1.01)

2.80 (1.23)

3.13 (0.93)

17.933

0.102

1 < 4*; 1 < 5; 2 < 3***; 2 < 4***; 2 < 5; 3 < 5; 4 < 5**

PM

1.40 (1.64)

1.09 (1.13)

0.77 (0.94)

0.47 (0.73)

0.40 (0.60)

21.501

0.120

1 > 4*; 1 > 5; 2 > 4**; 1 > 5; 3 > 5

DM

2.74 (1.37)

2.48 (1.26)

2.02 (1.08)

1.87 (0.97)

1.85 (0.94)

10.310

0.061

1 > 5; 2 > 5

HA

2.70 (1.01)

2.83 (0.94)

3.39 (0.99)

3.43 (1.04)

3.55 (0.94)

9.459

0.057

1 < 3*; 1 < 4*; 1 < 5; 2 < 3*;2 < 4*; 2 < 5

CS

2.40 (1.30)

2.11 (1.25)

2.12 (1.19)

1.84 (1.05)

1.67 (1.03)

6.706

0.041

1 > 4***; 1 > 5; 3 > 5**

Det.P

1.94 (1.25)

1.69 (1.09)

1.32 (1.01)

1.23 (1.01)

0.97 (0.94)

13.947

0.081

1 > 4**; 1 > 5; 2 > 5; 3 > 5**

Det.SS

2.73 (1.11)

2.94 (1.04)

2.12 (1.18)

1.92 (1.13)

1.69 (1.08)

16.599

0.095

1 > 5; 2 > 3**; 2 > 4***; 2 > 5; 3 > 5***

SA

2.28 (1.02)

2.37 (1.04)

2.08 (0.92)

1.68 (0.87)

1.68 (0.90)

11.244

0.067

1 > 5; 2 > 4***; 2 > 5; 3 > 4**; 3 > 5

BA

1.15 (0.91)

1.17 (0.88)

0.94 (0.97)

0.79 (0.93)

0.81 (0.87)

3.902*

0.024

1 > 5**; 2 > 4*; 2 > 5**; 3 > 5*

HS

2.89 (1.13)

2.72 (1.19)

2.52 (1.20)

2.31 (1.17)

2.03 (1.10)

8.656

0.052

1 > 5; 2 > 5***; 3 > 5; 4 > 5**

EDO

2.70 (1.52)

2.64 (1.26)

1.85 (1.23)

1.54 (1.33)

1.05 (1.11)

33.089

0.174

1 > 4**; 1 > 5; 2 > 3**; 2 > 4***; 2 > 5; 3 > 5

All contrasts are significant at p < .001, except for ***(p < .005), **(p < .020), *(p < .050) and § (p > .050; ns)

VC Vulnerable child, AC angry child, EC enraged child, IC impulsive child, UC undisciplined child, HC happy child, PM punitive mode, DM demanding mode, HA healthy adult, CS compliant surrender, Det.P detached protector, Det.SS detached self-soother, SA self-aggrandizer, BA bully and attack, HS helpless surrenderer, EDO eating disorder overcontroller

aFocused contrast with covariates (ANCOVAs) was performed to test potential differences between EDs (1. AN anorexia nervosa, 2. BN bulimia nervosa, 3. BED binge eating disorder, 4. EDNOS eating disorder not otherwise specified, 5. no diagnosis) and SMI-ED-SF dimensions. Age and gender were used as covariates

Participants with no self-reported diagnosis of EDs showed lower means for each maladaptive mode as well as higher means for the adaptive modes, thus suggesting the goodness of the SMI-ED-SF in discriminating between the clinical and the general population.

Discussion

This study tested the psychometric properties of the shorter version of the Schema Mode Inventory for disordered eating both for the general population and a clinical sample, in Italy.

Findings confirmed an adequate fit for the 16-factor model, with moderate intercorrelations between subscales. However, the Satorra-Bentler Chi square was statistically significant and the CFI values did not achieve the desired cutoff score (CFI > 0.90 [38, 39]: CFI = 0.880). They may have been affected by the sample size (i.e., Chi square [34, 35, 52, 53, 54]) and the number of considered indicators, (i.e., CFI [36, 40, 41, 42, 46, 54, 55, 56]) respectively, but, since both the SRMR and RMSEA accounted for the goodness of the model, this is not reason for concern [40]. Also, internal consistency within subscales was high, and the scale showed good overall reliability.

As expected, disordered eating behaviours were positively correlated with most of the negative coping modes, and negatively related to the healthy modes (healthy adult and happy child). Specifically, the overcontroller mode and the helpless surrenderer dimensions (explicitly designating the presence of disordered eating patterns) showed moderate-to-high correlations with the eating/weight/shape concerns subscales of the EDE-Q, as well as with the EDE-Q global score. Consistently, higher mean scores for the Healthy Modes were noticed in respondents with no self-reported diagnosis of EDs.

Findings from this study reflect those observed by testing the psychometric properties of the Schema Mode Inventory for eating Disorders (SMI-ED) [13]—the adapted version of the Schema Mode Inventory (SMI) for the measurement of mode states within a population with self-reported disordered eating behaviours [18]—but overcome some of its methodological and practical limitations. In fact, unlike for the SMI-ED validation study, participants were recruited from both clinical and non-clinical populations, thus supporting the discriminatory power of the tool and its ability to identify individuals at risk/with disordered eating behaviours. By assessing the psychometric proprieties of the questionnaire in Italian—and demonstrating their goodness of fit—further evidence was also reached for both its construct and external validity. Moreover, a meaningful item reduction resulting in the development of a new shorter instrument in Italian increases the scale usability for both clinical and research purposes.

Nonetheless, these results should be considered a first step in the validation process of the SMI-ED-SF, and as a promising starting point for future research on the topic. In fact, as the sample was purely recruited via online survey, it has its limitations. First, it was not possible to ensure gender homogeneity among respondents—although a smaller proportion of males is representative of the gender ratio usually found in clinical settings [57]. Also, a relatively low proportion of participants revealed binge eating behaviours compared with other dysfunctional eating patterns, and the percentage of respondents who had never been diagnosed with an ED doubled its counterpart. In addition, asking people to self-report an existing diagnosis of EDs may have led to under-represent both those with reduced capacity to acknowledge their ED patterns, and individuals with severe EDs but avoidant of support services.

Future studies should ideally include a larger percentage of males in the sample, and all ED subgroups should be adequately represented within the sample to more precisely determine whether specific profiles of schema modes exist within a given diagnostic group, and the degree to which this is statistically feasible. The measurement invariance between clinical and non-clinical populations should also be tested to ascertain whether the questionnaire is valid to measure schema modes in each group separately.

Conclusion

This scale is of significant value for clinicians and researchers in identifying and exploring mechanisms through which schema modes are expressed within the ED population—both quantitatively and qualitatively. In fact,—as the SMI-ED—the SMI-ED-SF not only provides information regarding modes that would not be otherwise accessible in the original SMI [18], but—because of its reduced number of items—it facilitates the capacity to make important links between ED symptoms and schema modes, and in developing individually tailored case conceptualizations and treatments.

In fact, although CBT is widely recognized as the gold standard intervention for adults with EDs, it is still restricted to the ineffective coping mechanisms maintaining the problem [58], without adequately addressing early life experiences often at the root of the painful or unhelpful ways of thinking, feeling and behaving typical of clients with EDs. Evidence supports the effectiveness of ST in facilitating behavioural change both through diminishing the emotional intensity of memories linked to EMS [and associated ED symptoms], alongside direct behavioural pattern-breaking. The development of a measure specifically aimed at facilitating a more precise measurement of mode states within the ED population will enable clinicians to provide more sophisticated conceptualizations and therapeutic opportunities for those with EDs, and to enhance long-term maintenance of the achieved results [10].

Notes

Acknowledgements

The author(s) received no financial support for the research, authorship and/or publication of this article.

Compliance with ethical standards

Conflict of interest

The authors declare that there are no conflicts of interest.

Research involving human participants and/or animals

All procedures performed in studies were run in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Informed consent

Written informed consent was obtained from all participants. The Medical Ethics Committee of Istituto Auxologico Italiano approved the study protocol and the informed consent process.

References

  1. 1.
    Fassino S, Abbate-Daga G (2013) Resistance to treatment in eating disorders: a critical challenge. BMC Psychiatry 13:282.  https://doi.org/10.1186/1471-244X-13-282 CrossRefGoogle Scholar
  2. 2.
    Pietrabissa G (2018) Group motivation-focused interventions for patients with obesity and binge eating disorder. Front Psychol 9:1104.  https://doi.org/10.3389/fpsyg.2018.01104 CrossRefGoogle Scholar
  3. 3.
    Sorgente A et al (2017) Web-based interventions for weight loss or weight loss maintenance in overweight and obese people: a systematic review of systematic reviews. J Med Internet Res 19(6):e229.  https://doi.org/10.2196/jmir.6972 CrossRefGoogle Scholar
  4. 4.
    Grilo CM (2017) Psychological and behavioral treatments for binge-eating disorder. J Clin Psychiatry 78(Suppl 1):20–24.  https://doi.org/10.4088/JCP.sh16003su1c.04 CrossRefGoogle Scholar
  5. 5.
    Linardon J, Brennan L (2017) The effects of cognitive-behavioral therapy for eating disorders on quality of life: a meta-analysis. Int J Eat Disord 50(7):715–730.  https://doi.org/10.1002/eat.22719 CrossRefGoogle Scholar
  6. 6.
    Agras WS, Fitzsimmons-Craft EE, Wilfley DE (2017) Evolution of cognitive-behavioral therapy for eating disorders. Behav Res Ther 88:26–36.  https://doi.org/10.1016/j.brat.2016.09.004 CrossRefGoogle Scholar
  7. 7.
    Castelnuovo G et al (2017) Cognitive behavioral therapy to aid weight loss in obese patients: current perspectives. Psychol Res Behav Manag 10:165–173.  https://doi.org/10.2147/PRBM.S113278 CrossRefGoogle Scholar
  8. 8.
    Young JE, Klosko JS, Weishaar ME (2003) Schema therapy: a practitioner’s guide. The Guilford Press, New YorkGoogle Scholar
  9. 9.
    Bamelis LL et al (2014) Results of a multicenter randomized controlled trial of the clinical effectiveness of schema therapy for personality disorders. Am J Psychiatry 171(3):305–322.  https://doi.org/10.1176/appi.ajp.2013.12040518 CrossRefGoogle Scholar
  10. 10.
    Castelnuovo G et al (2016) Not only clinical efficacy in psychological treatments: clinical psychology must promote cost-benefit, cost-effectiveness, and cost-utility analysis. Front Psychol 7:563.  https://doi.org/10.3389/fpsyg.2016.00563 Google Scholar
  11. 11.
    Talbot D et al (2015) Schema modes in eating disorders compared to a community sample. J Eat Disord 3:41.  https://doi.org/10.1186/s40337-015-0082-y CrossRefGoogle Scholar
  12. 12.
    Voderholzer U et al (2014) A comparison of schemas, schema modes and childhood traumas in obsessive–compulsive disorder, chronic pain disorder and eating disorders. Psychopathology 47(1):24–31.  https://doi.org/10.1159/000348484 CrossRefGoogle Scholar
  13. 13.
    Simpson SG et al (2018) Factorial structure and preliminary validation of the schema mode inventory for eating disorders (SMI-ED). Front Psychol 9:600.  https://doi.org/10.3389/fpsyg.2018.00600 CrossRefGoogle Scholar
  14. 14.
    MacCallum RC et al (1999) Sample size in factor analysis. Psychol Methods 4:84–99.  https://doi.org/10.1037/1082-989X.4.1.84 CrossRefGoogle Scholar
  15. 15.
    Terwee CB et al (2007) Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol 60:34–42.  https://doi.org/10.1016/j.jclinepi.2006.03.012 CrossRefGoogle Scholar
  16. 16.
    American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders, 5th edn. American Psychiatric Publishing, ArlingtonCrossRefGoogle Scholar
  17. 17.
    Calugi S et al (2017) The eating disorder examination questionnaire: reliability and validity of the Italian version. Eat Weight Disord 22(3):509–514.  https://doi.org/10.1007/s40519-016-0276-6 CrossRefGoogle Scholar
  18. 18.
    Young JE et al (2007) The schema mode inventory. Schema Therapy Institute, New YorkGoogle Scholar
  19. 19.
    Rosseel Y (2012) lavaan: an R package for structural equation modeling. J Stat Softw 48(2):1–36.  https://doi.org/10.18637/jss.v048.i02 CrossRefGoogle Scholar
  20. 20.
    Rosseel Y et al (2015) Package ‘lavaan’. http://cran.r-project.org/web/packages/lavaan/lavaan.pdf. pp 1–89
  21. 21.
    Team RC (2015) R: a language and environment for statistical computing. R Foundation for Statistical ComputingGoogle Scholar
  22. 22.
    Team RC (2014) The R project for statistical computing.Google Scholar
  23. 23.
    Nie NH, Bent DH, Hull CH (1970) SPSS: statistical package for the social sciences. McGraw-Hill, New YorkGoogle Scholar
  24. 24.
    Muthén LK, Muthén BO (1998–2012) Mplus user’s guide, 7th edn. Muthén & Muthén, Los AngelesGoogle Scholar
  25. 25.
    Muthén B, du SHC, Toit, Spisic D (1997) Robust inference using weighted least squares and quadratic estimating equations in latent variable modeling with categorical and continuous outcomes. Psychometrika 75(1):40–45Google Scholar
  26. 26.
    Hoyle RH (2012) Handbook of strucural equation modeling. The Guilford Press, New YorkGoogle Scholar
  27. 27.
    Bentler PM (1995) EQS structural equation program manual, in multivariate software. CA, EncinoGoogle Scholar
  28. 28.
    Satorra A, Bentler PM (1988) Scaling corrections for chi-square statistics in covariance structure analysis. In: Business and economic section of the American Statistical Association. American Statistical Association, AlexandriaGoogle Scholar
  29. 29.
    Satorra A, Bentler PM (1994) Corrections to test statistics and standard errors in covariance structure analysis. In: von Eye A, CC Clogg (eds) Latent variables analysis: applications for developmental research. Sage, Thousand OaksGoogle Scholar
  30. 30.
    Barrett P (2007) Structural equation modelling: adjudging model fit. Person Indiv Diff 42(5):815–824.  https://doi.org/10.1016/j.paid.2006.09.018 CrossRefGoogle Scholar
  31. 31.
    Steiger JH, Lind JC (1980) Statistically-based test for the number of common factors, in annual meeting of the Psychometric Society, Iowa City, IAGoogle Scholar
  32. 32.
    Steiger JH (1990) Structural model evaluation and modification: an interval estimation approach. Multivariate Behav Res 25(2):173–180.  https://doi.org/10.1207/s15327906mbr2502_4 CrossRefGoogle Scholar
  33. 33.
    Bentler PM (1990) Comparative fit indexes in structural models. Psychol Bull 107(2):238–246.  https://doi.org/10.1037/0033-2909.107.2.238 CrossRefGoogle Scholar
  34. 34.
    Jöreskog KG (1969) A general approach to confirmatory maximum likelihood factor analysis. Psychometrika 34:183–202CrossRefGoogle Scholar
  35. 35.
    Bentler PM, Bonett DG (1980) Significance tests and goodness of fit in the analysis of covariance structures. Psychol Bull 88:588–606.  https://doi.org/10.1037/0033-2909.88.3.588 CrossRefGoogle Scholar
  36. 36.
    Hu LT, Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling 6:1–55.  https://doi.org/10.1080/10705519909540118 CrossRefGoogle Scholar
  37. 37.
    Browne MW, Cudeck R (1990) Single sample cross-validation indices for covariance structures. Multivariate Behav Res 24:445–455.  https://doi.org/10.1207/s15327906mbr2404_4 CrossRefGoogle Scholar
  38. 38.
    Brown TA (2015) Confirmatory factor analysis for applied research, 2nd edn. In: TD Little (ed) The Guilford Press, New YorkGoogle Scholar
  39. 39.
    Browne MW, Cudeck R (1993) Alternative ways of assessing model fit. Sociol Methods Res 21(2):230–258.  https://doi.org/10.1177/0049124192021002005 CrossRefGoogle Scholar
  40. 40.
    Kenny DA, McCoach DB (2003) Effect of the number of variables on measures of fit in structural equation modeling. Struct Equ Model 10(3):333–351.  https://doi.org/10.1207/S15328007SEM1003_1 CrossRefGoogle Scholar
  41. 41.
    Russell DW (2002) In search of underlying dimensions: the use (and abuse) of factor analysis in Personality and Social Psychology Bulletin. Person Soc Psychol Bull 28(12):1629–1646.  https://doi.org/10.1177/014616702237645 CrossRefGoogle Scholar
  42. 42.
    Iacobucci D (2010) Structural equations modeling: Fit indices, sample size, and advanced topics. J Consumer Psychol 20(1):90–98.  https://doi.org/10.1016/j.jcps.2009.09.003 CrossRefGoogle Scholar
  43. 43.
    Marsh HW, Hocevar D (1985) Application of confirmatory factor analysis to the study of self-concept: first-and higher order factor models and their invariance across groups. Psychol Bull.  https://doi.org/10.1037/0033-2909.97.3.562 Google Scholar
  44. 44.
    Wheaton B (1977) Assessing reliability and stability in panel models. In: Heise DR (ed) Sociological methodology. Jossey-Bass, Inc., San Francisco, pp 84–136Google Scholar
  45. 45.
    Manzoni GM et al (2018) Feasibility, validity, and reliability of the italian pediatric quality of life inventory multidimensional fatigue scale for adults in inpatients with severe obesity. Obes Facts 11(1):25–36.  https://doi.org/10.1159/000484565 CrossRefGoogle Scholar
  46. 46.
    Manzoni GM et al (2018) Validation of the Italian Yale Food Addiction Scale in postgraduate university students. Eat Weight Disord 23(2):167–176.  https://doi.org/10.1007/s40519-018-0495-0 CrossRefGoogle Scholar
  47. 47.
    Pietrabissa G et al (2017) Stages of change in obesity and weight management: factorial structure of the Italian version of the University of Rhode Island Change Assessment Scale. Eat Weight Disord 22(2):361–367.  https://doi.org/10.1007/s40519-016-0289-1 CrossRefGoogle Scholar
  48. 48.
    Cronbach LJ (1951) Coefficient alpha and the internal structure of tests. Psychometrika 16(3):297–334.  https://doi.org/10.1007/BF02310555 CrossRefGoogle Scholar
  49. 49.
    Raykov T (2012) Scale construction and development using structural equation modeling. In: Hoyle RH (ed) Handbook of structural equation modeling. The Guilford Press, New York, pp 472–492Google Scholar
  50. 50.
    Bentler PM (2009) Alpha, dimension-free, and model-based internal consistency reiability. Psychometrika 74(1):137–143.  https://doi.org/10.1007/s11336-008-9100-1 CrossRefGoogle Scholar
  51. 51.
    Barbaranelli C et al (2014) Dimensionality and reliability of the self-care of heart failure index scales: further evidence from confirmatory factor analysis. Res Nurs Health 37(6):524–537.  https://doi.org/10.1002/nur.21623 CrossRefGoogle Scholar
  52. 52.
    Tuker LR, Lewis C (1973) A reliability coefficient for maximum likelihood factor analysis. Psychometrika 38:1–10.  https://doi.org/10.1007/BF02291170 CrossRefGoogle Scholar
  53. 53.
    James LR, Mulaik SA (1982) Causal analysis: assumptions, models and data. Sage, Beverly HillsGoogle Scholar
  54. 54.
    Kline RB (2015) Principles and practice of structural equation modeling. The Guilford Press, New YorkGoogle Scholar
  55. 55.
    Fan X, Sivo SA (2005) Sensitivity of fit indexes to misspecified structural or measurement model components. Struct Equ Model 12(3):343–367CrossRefGoogle Scholar
  56. 56.
    Hu L, Bentler PM (1998) Fit indexes in covariance structure modeling: sensitivity to under parameterized model misspecification. Sychol Methods 3(4):424–453CrossRefGoogle Scholar
  57. 57.
    Striegel-Moore RH et al (2009) Gender difference in the prevalence of eating disorder symptoms. Int J Eat Disord 42(5):471–474.  https://doi.org/10.1002/eat.20625 CrossRefGoogle Scholar
  58. 58.
    Waller G, Kennerley H (2003) Cognitive-behavioral treatments. In: Treasure J, Schmidt U, Furth E (eds) Handbook of eating disorders. Wiley, Chichester, pp 233–252Google Scholar

Copyright information

© The Author(s) 2019

Authors and Affiliations

  1. 1.Psychology Research LaboratoryIstituto Auxologico Italiano IRCCSMilanItaly
  2. 2.Department of PsychologyCatholic University of MilanMilanItaly
  3. 3.Department of Philosophy, Sociology, Education, and Applied Psychology, Section of Applied PsychologyUniversity of PaduaPaduaItaly
  4. 4.School of Psychology, Social Work and Social PolicyUniversity of South AustraliaAdelaideAustralia
  5. 5.Regional Eating Disorders Unit, NHS LothianSt. John’s HospitalLivingstonUK
  6. 6.Casa di Cura San RossorePisaItaly
  7. 7.Faculty of PsychologyeCampus UniversityComoItaly

Personalised recommendations