INTRODUCTION

Childhood experiences are known to be important for lifelong development, just as negative experiences like neglect and abuse may cause negative results including adjustment disorder; thus, positive childhood experiences may perserve wellbeing in adults (1-3).

In the literature, the majority of research is focusing on the effects of negative childhood experiences. This research found experiences like neglect, abuse, long-term separation, and early loss to be related with several factors, such as psychiatric disorders (4-6), personality disorders (7-10), a risk of developing chronic disease (11,12), alcohol and substance abuse (13), and delinquency (14) in adulthood. In this area, Felitti et al. (15) developed a 10-item Adverse Childhood Experiences (ACE) questionnaire to assess negative childhood experiences such as neglect and abuse suffered during the first 18 years of a person’s life. Validity and reliability of the Turkish version have been confirmed by Gunduz et al. (16).

Though not to the same extent as studies on the effect of ACE on adulthood, recently work on the effect of beneficial childhood experiences like healthy attachment, effective parenting, and availability of social resources has begun to appear in the literature, particularly regarding the acquisition of resilience in adversity and lifelong development (1-3). In parallel with the ACE questionnaire (15), Narayan et al. (3) developed the Benevolent Childhood Experiences (BCEs) Scale to evaluate positive experiences during the first 18 years of life, such as internal and external perceived safety and support or positive and predictable qualities, thought to help prevent or override long-term effects of adverse experiences in early life.

Positive childhood experiences have been seen to protect not only against certain psychopathologies like adult depression (17) and personality disorder (18), but generally in a range of different areas of physical, cognitive, and social health (19). Therefore, evaluating BCEs is not only relevant in its own field but may offer directions to healthcare professionals, families, and decision-makers in order to strengthen healthy child development and reduce the effect of ACE. Thus, in addition to reducing those negative childhood experiences, increasing positive childhood experiences can be initiated.

Our study aimed to adapt the recently developed original BCEs to Turkish and test its adequacy for use in Turkey, examining its psychometric properties in order to introduce a scale for measuring positive childhood experiences to a country where such a research instrument is not yet available.

METHOD

Participants

A total of 175 individuals participated in this study, 76 students enrolled at Hacettepe University and 99 persons completing an online form. The number of participants was calculated in view of a 1:10 ratio: A number of studies recommends this ratio between sample size and number of items (20,21). A 1:10 ratio means that for every item on the scale, data of at least 10 participants should be collected for an ideal item:participant number ratio (22). 

As an exclusion criterion, age below 18 and above 65 years was used (mean age=25.07 years, standard deviation [SD]=6.43) (Table 1). The sample included 137 females (78%) and 38 males (22%) (Table 1). There were 74 university graduates (42.3%), 67 high school graduates (38.3%), and 32 participants had a postgraduate degree (18.3%). As their marital status, 140 (80%) recorded single and 32 (18.3%) married. The mean score for the data retrieved from the first section of the Post-Traumatic Stress Symptom Scale (PTSSS) was 1.21 for the total sample, 1.25 for the online participants, and 1.16 for the university students (Table 1).

Procedure

The required permission was received from the developer of the original BCEs Scale (3), Dr. Angela Narayan, to adapt the instrument to Turkish and examine its psychometric properties. For the other scales to be used in the study, permission was obtained from the developers of the respective Turkish versions. After the ethics committee of Hacettepe University granted the required approval. After the research team completed the translation-retranslation of the original items, the resulting Turkish items were forwarded to 3 area specialists (1 associate professor and 2 assistant professors), and in the light of their feedback, the definitive Turkish form was designed.

After the Turkish translation was completed, participants were contacted in two different ways: Primarily, to constitute the group consisting of university students, the researchers explained students from the Psychology department of Hacettepe university the research project before the start of the lecture and collected data from students volunteering to participate in the research. A similar announcement of the project was posted on social media accounts via the internet, including a link to a form created with an online survey tool used for scientific studies, Qualtrics, reaching the remaining participants through the online survey. In the informed consent form, the aim of the study, adapting the scale to Turkish, was explained to reduce the potential bias towards positive experiences. The aim of administering the form online was to generate a sample with a more generalizable age and sex distribution, given that the student group was of similar age and mainly female. Of the participants completing the preliminary measurements during the lecture, 52 persons were administered the test again 3 weeks later during lecture time. None of the participants using pen and paper or those filling in the online form rejected the research or left their forms incomplete.

Measures

All participants were given the following forms in turn:

Sociodemographic Data Form: This form recorded the participants’ sociodemographic information including age, sex, marital state, and duration of education. In addition, it inquired about the frequency of having experienced traumatic events, using the questions included in the first section of the PTSSS (23). The sociodemographic form was used to investigate potentially confounding variables. As the frequencies of experiencing traumatic events such as abuse and neglect are considered potential confounding variables, this survey was used in order to determine any possible difference in the frequency of traumatic experiences between university students completing the pen-and-paper forms and participants in the online survey.

Benevolent Childhood Experiences (BCEs) Scale: This self-report scale consisting of 10 items in the form of yes/no questions has been developed to collect data about positive experiences during childhood. The validity and reliability of the original English form has been studied by Narayan et al. (3).

Symptom Check List 90 (SCL-90-R): This instrument consists of 90 items scored on a 5-point Likert-type scale. The SCL-90-R is a self-report scale screening for psychiatric symptoms. The latest version of the original form was developed by Derogatis (24), and the validity and reliability of the Turkish version was confirmed by Dag (25). As benevolent childhood experiences were found to be negatively correlated with psychopathologic symptoms (3,17,19), the SCL-90-R was used to investigate criterion-related validity.

Satisfaction with Life Scale (SWLS): This instrument, developed by Diener et al. (26), consists of 5 items on a 5-point Likert-type scale. A study to confirm the validity and reliability of the Turkish version was made by Dagli and Baysal (27). As positive childhood experiences were found positively related with wellbeing in adulthood (17,19), the SWLS was used to investigate criterion-related validity.

Statistical Analysis

The adaptation of the form was carried out following the qualitative and statistical methods laid out by Savasir (28), especially the translation-retranslation procedure from English to Turkish and back. SPSS version 24 was used to for the statistical analysis of the study. In the first step, descriptive statistics for all values in the study were calculated, including minimum and maximum values, arithmetic means, and standard deviation. Independent groups t-test was used to determine if there was a significant difference in age and PTSSS scores between the online survey participants and the university students participating in pen-and-paper format; chi-square test was performed to examine if there was a significant sex difference between the participant groups.

The construct validity of the scale was tested by exploratory factor analysis (EFA); additionally, the factor structure was investigated through parallel analysis. The factors established by EFA were tested by confirmatory factor analysis (CFA) using the AMOS package. For criterion-related validity, the correlation of data from the Turkish form of the BCEs with results from the SWLS and SCL-90-R was investigated. In order to establish the reliability of the instrument, test-retest reliability was used, administering the form to 52 participants from the sample group 3 weeks later under identical conditions for a second time. For the scores reached in the two iterations, Pearson correlation coefficient was calculated to obtain the reliability coefficient for the scale. For internal consistency, Cronbach’s alpha coefficients were calculated for all items of the scale and for the subdimensions of the scale.

RESULTS

This section lists the results for the psychometric characteristics of the BCEs adapted to Turkish.

Independent Groups T-Test and Chi-Square Test

Independent groups t-test was used to establish possible differences in mean age between the university student group and the online survey participant group. This test, comparing mean values between two groups (female-male, control-trial group, etc.) is used in order to decide if differences between groups are random or statistically significant (29). Independent groups t-test showed a statistically significantly higher age in the online participant group (mean=27.76 years, SD=6.94, n=99) compared to the university students completing the form with pen and paper (mean=21.57 years, SD=3.30, n=76), t (173)=-7.17, p<0.001 (Table 1).

The same analysis was performed to test for a possible difference between university students filling in the paper forms and online survey participants regarding the incidence of having experienced traumatic events. No statistically significant difference was found in the incidence of traumatic events experienced between the online survey participants (mean=1.25, SD=1.22, n=99) and the university students using paper forms (mean=1.16, SD=1.15, n=76), t (173)=-0.52, p>0.05 (Table 1).

Chi-square test of independence was used to determine if there was a sex difference between the pen-and-paper group and the online group. This test is used to decide if there is a correlation between 2 categorical variables or if the two categorical variables are independent from each another. If the results are statistically significant (p<0.05), a correlation between the variables has been demonstrated, while if they are not statistically significant (p>0.05), the two variables are independent (30). Our analysis found a significant correlation between group membership (pen-and-paper or online format) and participants’ sex, χ2 (1)=12.36, p<0.001 (Table 1).

Validity Analysis

Construct Validity: EFA was performed to test the construct validity of the BCEs. Factor analysis aims to reveal unobservable and non-measurable hidden dimensions as multiple causes of a measurable and observable characteristic. EFA is used to reveal and discover the factor structure underlying the expressions representing the items of a scale that has been translated from one language to another (31). Before performing EFA, the adequacy of the data for factor analysis was assessed through Bartlett test and Kaiser-Meyer-Olkin (KMO) measure. For data to be adequate for factor analysis, the Bartlett test needs to be statistically significant (p<0.01) (32) and the KMO measure is expected to be above 0.50 (33). Bartlett test and KMO measure showed our study data to be adequate for analysis (KMO=0.66 and Bartlett’s test χ2=167.81, SD=45, p<0.001). To examine the factor structure of the scale and determine subdimensions, factor analysis was performed using varimax rotation. In the first step, a 4-factor structure with an eigenvalue above 1 explaining 57.43% of the total variance was found. However, when evaluating the eigenvalue graph (scree plot), the elbow of the curve was found after the 2nd factor; therefore, the number of factors to retain was cut off at 2, because the contribution of factors beyond that point is relatively small and almost identical (34). 

To determine the number of factors of the scale, Horn’s (35) parallel analysis based on random data production was used. In parallel analysis, a random data set parallel to the actual data is generated with the Monte Carlo simulation method and the expected eigenvalue of the parallel data is calculated. Subsequently, the eigenvalue of the parallel data and the expected eigenvalues are compared. As a result of this comparison, the point where the eigenvalue of the parallel data is greater than the eigenvalue in the actual data set is considered significant to determine the number of factors (35). Having decided for a 2-factor structure of the scale, parallel analysis was performed using the syntax developed by O’Connor (36). Comparison of the eigenvalues obtained from parallel analysis and the eigenvalues resulting from EFA supported a 2-factor structure of the instrument (Table 2). Items 8 and 10 were found to load on both factors. However, given that item 8 loaded on factor 1 with a higher value and considering its content, it was decided to keep it in factor 1. Though item 10 also loaded with a higher value on factor 1, it was kept in factor 2 as it seemed to fit its content better.

When the analysis was repeated on this basis, eigenvalues for the factors of the instruments were found to be 1.96 for the first factor and 1.68 for the second. The first factor explained 19.62% of the total variance, the second factor 16.83%, and both factors together explained 36.45% of the total variance (Table 3). In the end, in the light of examining the items constituting the factors, the first factor consisting of items 1, 2, 6, 7, and 8 was named “perceived safety and support” and the second factor, including items 3, 4, 5, 9, and 10, was called “internal and environmental motivation.” Results of the EFA of BCEs are presented in Table 3.

As the factor structure had not been investigated in the process of developing the original instrument, we performed CFA with the AMOS package for the factor structure emerging as a result of the EFA in our study. The analysis showed that the first model tested had an adequate fit index (Model 1; χ2 [34]=51.421, p<0.05; χ2/df=1.512; GFI=0.95, CFI=0.86, NFI=0.70, RMSEA=0.05). A ratio of chi-square to degrees of freedom (χ2/df) smaller than 5, as in our model, indicates that the fit is good (37); a CFI value close to 1 points to a good fit (38), an RMSEA value ≤0.05 means an excellent fit (39,40), and a GFI index of 0.90 or above shows a good fit (41.42). All of these indices confirm the factor structure of the BCEs to be at an acceptable level. The factor loadings of items in the resulting model and their error variances are shown in Figure 1.

Criterion-related Validity: This type of validity tests if an instrument measures the intended construct, looking at the relation between scores from the instrument to be evaluated and results measured with an appropriate instrument (43). In this sense, the relevant literature indicates that positive childhood experiences are related with a person’s psychological health and life satisfaction, and it is known that benevolent childhood experiences can be protective against psychopathologies in adult life. In this context, criterion-related validity was tested using SCL-90-R, expected to correlate negatively with BCEs, and SWLS, which should show a positive correlation. The correlation of participants’ SCL-90-R scores for severity of psychologic symptoms and their BCEs scores was examined using Pearson product-moment correlation coefficient, and a statistically significant weak negative correlation was found (r=-0.25, p<0.001). Between SWLS scores for life satisfaction and benevolent childhood experiences, a statistically significant moderate positive correlation was detected (r=0.42, p<0.001). Results of the correlation analysis performed to investigate the relation of BCEs with SCL-90-R and SWLS is presented in Table 4.

Reliability Analysis

Test-retest Reliability: To assess test-retest reliability, an instrument is administered for a second time to the same study group under similar conditions after a period that is long enough to ensure that participants do not remember the scale, but short enough to avoid significant changes in the construct to be measured. The correlation coefficient resulting from these two measurements is the reliability coefficient of the instrument (43). Thus, to examine the test-retest reliability of the BCEs, the scale was readministered to 52 participants after an interval of 3 weeks, resulting in a test-retest reliability coefficient of 0.91 for the entire scale. Test-retest reliability coefficients for the subdimensions perceived safety and support and internal and environmental motivation were 0.85 and 0.84, respectively.

Internal Consistency Reliability: This is a reliability measure used to establish the mutual consistency of all items included in an instrument (43). In this context, internal consistency reliability is evaluated by calculating Cronbach’s alpha coefficient. This coefficient is used to test the reliability of the scale by calculating the ratio of the variances of each individual item on the scale to the total variance of the instrument (43). The analysis of our data found a reliability coefficient for the entire BCEs of 0.61; for the subdimension perceived safety and support, the value was 0.55 and for internal and environmental motivation, 0.45. An alpha coefficient of 0.6 or above shows that the items are unidirectional and can be gathered in one scale (44).

DISCUSSION

While numerous studies in the literature are investigating ACE, the number of publications examining positive childhood experiences is quite limited. Until now, no valid and reliable instrument to assess benevolent childhood experiences has been available in Turkey. Aim of our study was to adapt the BCEs, developed by Narayan et al. (3), to Turkish and thereby introduce an instrument to the Turkish literature that allows an evaluation of positive childhood experiences. The findings of this adaptation study indicate that the BCEs is a valid and reliable tool to be used in subsequent research.

While Narayan et al. (3) asserted that the items on their scale were related with perceived internal and external safety (beliefs that give comfort, at least one caregiver with whom one feels safe, etc.), positive and predictable quality of life (liking school, predictable home routine, etc.), and relational support (caring teacher, supportive adult other than caregiver, etc.), no factor analysis for the instrument had been carried out. Having contacted the authors with regard to the issue of factor analysis, it was learned that they are in the process of collecting data from a larger sample with the intention to perform factor analysis on the basis of this broader data. Later studies using the original scale also did not attempt factor analysis (19,45), only test-retest reliability, predictive validity, and validity of the scale in a different sample were included (45). Findings from EFA and CFA in the Turkish adaptation study revealed a 2-factor structure of the instrument.

In view of the item content, these 2 factors were named the “perceived safety and support” and the “internal and environmental motivation” subdimensions. According to our findings, all items except for item 8 (“Did you have opportunities to have a good time?”) and item 10 (“Did you have a predictable home routine, like regular meals and a regular bedtime?”) loaded on one subdimension only (either “perceived safety and support” or “internal and environmental motivation”), while items 8 and 10 loaded on both subdimensions. As the loading ratio of item 8 on the subdimension perceived safety and support was greater and its content was relevant to this subdimension, it was included with items 1, 2, 6, and 7 in the former. The difference in factor loadings of item 10 was small and its content was more suitable for the subdimension “internal and environmental motivation”; hence, it was included in this subdimension together with items 3, 4, 5, and 9. Parallel analysis and CFA showed that the data were consistent with this 2-factor model. 

In our study, the subdimensions perceived safety and support and internal and environmental motivation were found to explain 19.62 and 16.83% of the variance, respectively. As Narayan et al. (3) did not perform factor analysis, it is not possible to compare the resulting variance values. The inadequate percentage of variance explained by the factors might be understood as an effect of the small number of participants. While the sample size in this study was sufficient for the number of items (46), the value for variance explained by the factors was low, which might be due to the sample size. In future studies with a larger sample, variance explained by the factors may be reassessed.

The original publication of the instrument focused on test-retest reliability and predictive validity (3), while another study using the original scale dealt with validity in a different sample and predictive validity (45). From this perspective, the only joint approach between previous research and our study is the examination of test-retest reliability. Similar to the original scale with a test-retest reliability of 0.80 and a high level of reliability (3), we found a test-retest reliability coefficient of 0.91, which is consistent with the original instrument. 

For internal consistency reliability, a reliability coefficient of 0.61 was found. As is known, Cronbach’s alpha value is related with the number of items, a smaller item number leading to a lower alpha value (33). Thus, the alpha value for a 10-item scale is expected to be low. However, a value of or above 0.6 shows that the items are unidirectional and can be included in a single scale (44). Therefore, the alpha value of the BCEs can be considered as acceptable.

Our study found a number of important results for the psychometric properties of the BCEs, but certain limitations remain. The first issue is sample size. Although several sources consider a number of 10 participants per item to be sufficient (20,21), others maintain that a sample size of at least 200 individuals is required (47). Especially for the BCEs subdimension items 5, 6, and 10, factor loadings close to the cutoff point (0.30) and high item-related error variances emerged. Furthermore, the results showed that in the Turkish adaptation of the BCEs, total variance explained was 36.45%, and criterion-related validity found a weak correlation with SCL-90-R (r=-0.25, p<0.001) and a moderate correlation with SWLS (r=0.42, p<0.001). As all these problems may be related to the small sample size, they should be reassessed in studies with larger samples. Another limitation might be the failure to use any instrument that includes ACE to investigate criterion-related validity. A recent study found a moderate correlation (r=-0.32, p<0.001) between BCEs and ACE (18).

Another limitation of this study is the uneven distribution of participants’ sex, age, and level of education, not being representative of Turkish society as a whole: as we prioritized accessibility, the sample can be described as lopsided in this sense. Considering the significant age and sex differences between pen-and-paper and online survey participants, we can say that including the online data into the research created a more generalizable sample than the pen-and-paper group on its own. But even so, 78% of the sample consisted of female participants. This limitation could be overcome by repeating the study with a more balanced sample regarding the distribution of sex, age, and level of education.

Another limitation lies in the imponderability involved with online administration of the tests, even though this approach facilitated the recruitment of participants with more varied demographic characteristics. Considering the disadvantages of online test application compared to face-to-face administration, the test might not have been completed in a single session but been restarted after an interruption; it was not possible to control for confounding factors (like the presence of another person, noise, etc.) that may have arisen during the completion of the test, and online participants had no opportunity to ask the researcher directly when they did not understand certain sections. On the other hand, especially through the use of IP address control, confounding situations like an individual participant’s repeated access to the research could be averted. 

Finally, the scale that we have adapted to Turkish is a self-report instrument requiring retrospective memory. As Narayan et al. (3) and Merrick et al. (45) already pointed out, reporting previous experiences may give rise to bias – which is the case for all self-report instruments based on retrospective memory. Future studies might help increase validity by administering the scale to individuals still in childhood age. 

Despite these limitations, the results of our validity and reliability analyses reached an acceptable level; therefore, we have been able to demonstrate that the BCEs can be used in Turkey as a valid and reliable tool. 

Ethics Committee Approval: Hacettepe University Ethics Board has approved this study.

Informed Consent: Informed consent of pariticipants were obtained.

Peer-review: Externally peer-reviewed.

Conflict of Interest: The authors declare that there are no conflicts of interest.

Financial Disclosure: All financial expenses have been met by the authors.

REFERENCES

1. Chung EK, Mathew L, Elo IT, Coyne JC, Culhane JF. Depressive symptoms in disadvantaged women receiving prenatal care: the influence of adverse and positive childhood experiences. Ambul Pediatr 2008; 8:109-116.

2. Hillis SD, Anda RF, Dube SR, Felitti VJ, Marchbanks PA, Macaluso M, et al. The protective effect of family strengths in childhood against adolescent pregnancy and it's long-term psychosocial consequences. Perm J 2010; 14:18-27. 

3. Narayan AJ, Rivera LM, Bernstein RE, Harris WW, Lieberman AF. Positive childhood experiences predict less psychopathology and stress in pregnant women with childhood adversity: A pilot study of the Benevolent Childhood Experiences (BCEs) Scale. Child Abuse Negl 2018; 78:19-30. 

4. Briere J, Elliott DM. Prevalence and psychological sequelae of self-reported childhood physical and sexual abuse in a general population sample of men and women. Child Abuse Negl 2003; 27:1205-1222. 

5. Chapman DP, Whitfield CL, Felitti VJ, Dube SR, Edwards VJ, Anda RF. Adverse childhood experiences and the risk of depressive disorders in adulthood. J Affect Disord 2004; 82:217-225.

6. Anda RF, Brown DW, Felitti VJ, Bremner JD, Dube SR, Giles WH. Adverse childhood experiences and prescribed psychotropic medications in adults. Am J Prev Med 2007; 32:389-394.

7. Johnson JG, Cohen P, Brown J, Smailes EM, Bernstein DP. Childhood maltreatment increases risk for personality disorders during early adulthood. Arch Gen Psychiatry 1999; 56:600-606.

8. Lobbestael J, Arntz A, Bernstein DP. Disentangling the relationship between different types of childhood maltreatment and personality disorders. J Pers Disord 2010; 24:285-295.

9. Zhang T, Chow A, Wang L, Dai Y, Xiao Z. Role of childhood traumatic experience in personality disorders in China. Compr Psychiatry. 2012; 53:829-836. 

10. Cohen LJ, Tanis T, Bhattacharjee R, Nesci C, Halmi W, Galynker I. Are there differential relationships between different types of childhood maltreatment and different types of adult personality pathology? Psychiatry Res 2014; 215:192-201. 

11. Anda RF, Brown DW, Dube SR, Bremner JD, Felitti VJ, Giles WH. Adverse childhood experiences and chronic obstructive pulmonary disease in adults. Am J Prev Med 2008; 34:396-403.

12. Dube SR, Fairweather D, Pearson WS, Felitti VJ, Anda RF, Croft JB. Cumulative childhood stress and autoimmune diseases in adults. Psychosom Med 2009; 71:243-250.

13. Anda RF, Croft JB, Felitti VJ, Nordenberg D, Giles WH, Williamson DF, et al. Adverse childhood experiences and smoking during adolescence and adulthood. JAMA 1999; 282:1652-1658.

14. Dube SR, Anda RF, Felitti VJ, Chapman DP, Williamson DF, Giles WH. Childhood abuse, household dysfunction, and the risk of attempted suicide throughout the life span: findings from the Adverse Childhood Experiences Study. JAMA 2001; 286:3089-3096.

15. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. Am J Prev Med 1998; 14:245-258.

16. Gunduz A, Yasar AB, Gundogmus I, Savran C, Konuk E. Adverse Childhood Events Turkish Form: validity and reliability study. Anadolu Psikiyatri Derg 2018; 19:68-75.

17. Bethell C, Jones J, Gombojav N, Linkenbach J, Sege R. Positive Childhood Experiences and Adult Mental and Relational Health in a Statewide Sample: Associations Across Adverse Childhood Experiences Levels. JAMA Pediatr 2019; 173:e193007.

18. Gunay-Oge R, Pehlivan FZ, Isikli S. The effect of positive childhood experiences on adult personality psychopathology. Pers Individ Dif 2020; 158:109862. 

19. Crandall A, Miller JR, Cheung A, Novilla LK, Glade R, Novilla MLB, et al. ACEs and counter-ACEs: How positive and negative childhood experiences influence adult health. Child Abuse Negl 2019; 96:104089.

20. Turker D. Measuring corporate social responsibility: a scale development study. J Bus Ethics 2009; 85:411-427.

21. Zheng J, You LM, Lou TQ, Chen NC, Lai DY, Liang YY, et al. Development and psychometric evaluation of the Dialysis patient-perceived Exercise Benefits and Barriers Scale. Int J Nurs Stud 2010; 47:166-180. 

22. Boateng GO, Neilands TB, Frongillo EA, Melgar-Quiñonez HR, Young SL. Best practices for developing and validating scales for health, social, and behavioral research: a primer. Front Public Health 2018; 6:149.

23. Isikli S. Correlations Between Event-Related Attention Bias, Level of Dissociation, and Working Memory Space in Individuals with Post-Traumatic Stress Symptoms. Doctoral Thesis, Hacettepe University, Institute for Social Sciences, Ankara, 2006. (Turkish)

24. Derogatis LR. SCL-90-R: Administration, Scoring And Procedures Manual. Baltimore: Clinical Psychometric Research 1977.

25. Dag I. Reliability and validity of the Symptom Check List (SCL-90-R) for university students. Turk Psikiyatri Derg 1991; 1:5-12. (Turkish).

26. Diener E, Emmons RA, Larsen RJ, Griffin S. The Satisfaction With Life Scale. J Pers Assess 1985; 49:71-75.

27. Dagli A, Baysal N. Adaptation of the Satisfaction With Life Scale into Turkish: the study of validity and reliability. Electronic Journal of Social Sciences 2016; 15:1250-1262. (Turkish)

28. Savasir I. Problems in the adaptation of scales and some solutions. Turk Psikoloji Dergisi 1994; 9:27-32. (Turkish)

29. Johnson RA, Wichern DW. Applied Multivariate Statistical Analysis. Sixth ed., Prentice-Hall: Pearson, 2007.

30. Miller R, Siegmund D. Maximally selected chi square statistics. Biometrics 1982; 38:1011-1016. 

31. Suhr D. Exploratory or Confirmatory Factor Analysis? Cary, NC: SAS Institute, 2006.

32. Bartlett MS. Tests of significance in factor analysis. Br J Stat Psychol 1950; 3:77-85. 

33. Field A. Discovering Statistics Using SPSS for Windows. London: Sage Publications Ltd., 2000.

34. Cokluk O, Sekercioglu G, Buyukozturk S. Multivariate Statistics for Social Sciences: SPSS and LISREL Applications. Second ed., Ankara: Pegem Akademi Yayincilik, 2012.

35. Horn JL. A rationale and test for the number of factors in factor analysis. Psychometrika 1965; 30:179-185.

36. O’Connor BP. SPSS and SAS programs for determining the number of components using parallel analysis and velicer’s MAP test. Behav Res Methods Instrum Comput 2000; 32:396-402.

37. Kelloway EK. Using Lisrel for Structural Equation Modeling: A Researcher’s Guide. Third ed, Boston: Sage Publications, 1998.

38. Bentler PM. Comparative fit indexes in structural models. Psychol Bull 1990; 107:238-246.

39. Anderson J, Gerbing D. The effect of sampling error on convergene, improper solutions and goodness of fit indices for maximum likelihood confirmartory factor analysis. Psychometrika 1984; 49:155-173.

40. Browne MW, Cudeck R. Alternative ways of assessing model fit. Sociol Method Res 1992; 21:230-258.

41. Schumacker RE, Lomax RG. A Beginner’s Guide to Structural Equation Modeling. Second ed., New Jersey: Lawrance Erlbaum Associates, 2004.

42. Kline RB. Beyond Significance Testing: Reforming Data Analysis Methods in Behavioral Research. Washington, DC: American Psychological Association, 2004.

43. Ercan I, Kan I. Reliability and validity in the scales. Uludag Universitesi Tip Fakültesi Dergisi 2004; 30:211-216. (Turkish)

44. Huck SW. Statistical Misconceptions. Firsth ed., London: Routledge, 2008.

45. Merrick JS, Narayan AJ, DePasquale CE, Masten AS. Benevolent Childhood Experiences (BCEs) in homeless parents: A validation and replication study. J Fam Psychol 2019; 33:493-498.

46. Buyukozturk S. Handbook of Data Analysis for Social Sciences. Fifteenth ed. Ankara: Pegem Akademi Yayinlari, 2011.

47. Pituch KA, Stevens J, Stevens J. Applied Multivariate Statistics for The Social Sciences. Sixth ed., London: Routledge, 2016.