An Icelandic Version of McMasters Family Assessment Device (FAD)

Abstract

Purpose:

An analysis of the psychometric properties of an Icelandic version of McMasters Family Assessment Device (FAD) was conducted in this study.

Method:

Two groups, clinical and nonclinical, comprising of 529 parents answered the FAD. The study examined the internal reliability and discriminant validity of the instrument in addition to examining construct validity using exploratory factor analysis (EFA).

Results:

The theoretically based original FAD is reliable where the total scale had high α, .95. The factors found for the Icelandic FAD were not as reliable as the original FAD. The discriminant validity showed that the clinical group scored higher (more dysfunctional) than the nonclinical group in most dimensions of both original FAD and the Icelandic FAD. The eight factors extracted in an EFA were different from the original factors.

Discussion:

The Icelandic FAD possesses good psychometric properties, although its factor structure differs from the original FAD. The results are promising and merit continued application.

Keywords

instrument development psychometric study mental health field of practice child welfare evidence-based practice literature review methodological article family therapy qualitative outcome study youth population assessment internal consistency

The following study focuses on an assessment of psychometric properties of an Icelandic version of McMasters Family Assessment Device (FAD). The FAD is a self-report questionnaire that assesses family functioning on each of the six dimensions of the McMasters model of family functioning (MMFF). The six dimensions are Problem solving (PS), which reflects the family’s ability to resolve problems and steps taken to achieve this; Communication (CM), a dimension that refers to the verbal exchange of information within a family; Roles (RL), which assesses the repetitive patterns of behavior by which family members fulfill family functions; Affective responsiveness (AR), that assesses family member’s ability to respond with the appropriate emotions; Affective involvement (AI), which refers to the extent to which the family, as a group, values and shows interest in particular activities of individual family members and finally Behavioral control (BC), that describes the expected standards for behavior (Epstein, Baldwin, & Bishop, 1983; Ryan, Epstein, Keitner, Miller, & Bishop, 2005). In addition, the FAD includes a General Functioning (GF) scale that measures the overall level of family functioning (Epstein et al., 1983; Ryan et al., 2005).

The objective of this study was to validate the FAD in Icelandic, which could measure functioning of families in Iceland, both for epidemiological purposes and to be able to measure outcome of family interventions, such as family therapy or parental training programs.

Families are complex systems and are difficult to assess and quantify for many reasons. One reason is that parenting styles differ widely, the family’s life cycle stage is another factor and subsystems within the family influence each other. The pressure of larger systems like social systems can also be a factor and cultural factors can also further complicate valid measurements (Goldenberg & Goldenberg, 2008). Even if there are many factors that complicate measurements of families, documenting the family’s role as a stressor and as an essential support, when individuals are battling physical and mental illnesses, is an important focus for family social work. Family role is also vital in assessing recovery and performance of services, especially when user’s views are taken into account (Addington, Collins, McCleery, & Addington, 2005; Bhui, Mohamud, Warfa, Craig, & Stansfeld, 2003; Houtzager et al., 2004).

The concept of family functioning revolves around those factors that are seen as having the most impact on the emotional and physical health or problems of family members (Ryan et al., 2005). When families are being assessed, it is important to decide from what perspective the assessment is being made. Self-report instruments are useful in those kinds of assessments where the family gets its own voice and each individual can rapport his or her own experience (Goldenberg & Goldenberg, 2008). The development of family functioning scales has sought to cover this area of interest.

Although a self-report instrument is considered satisfactory in one country, research has shown that in cross-cultural comparative studies on family functioning, it is essential to establish equivalent measurement of relevant constructs across cultures (Kankaraš & Moors, 2010; Lauth, Magnússon, Ferrari, & Pétursson, 2008). The FAD has been translated and adapted in over 20 languages (Al-Krenawi, Graham, & Slonim-Nevo, 2002; Al-Krenawi, Lev-Wiesel, & Sehwail, 2007; Barroilhet, Cano-Prous, Cervera-Enguix, Forjaz, & Guillén-Grima, 2009; Chen et al., 2003; Kazarian, 2009; Keitner et al., 1991; Kiliç, Ozguven, & Sayil, 2003; Roncone et al., 1998; Ryan et al., 2005; Shek, 2002; Wenniger, Hageman, & Arrindell, 1993). Results from cross-cultural studies of the FAD suggest that cultural values can affect a family’s functioning and that differences in areas of family functioning can be captured using the FAD (Keitner et al., 1991). The FAD has also been used to assess family functioning in nonclinical families (Stevenson-Hinde & Akister, 1995), families with various medical (Chen & Clark, 2007; Hanks, Rapport, & Vangel, 2007; Malec, Testa, Rush, Brown, & Moessner, 2007; Miller, Bishop, Herman, & Stein, 2007; Schmitt et al., 2008) and psychiatric disorders (Celikel, Cumurcu, Koc, Etikan, & Yucel, 2008; Derisley, Libby, Clark, & Reynolds, 2005; Erol, Yazici, & Toprak, 2007; Kabacoff, Miller, Bishop, Epstein, & Keitner, 1990; Keitner et al., 1995; Lieb et al., 2000; Miller, McDermut, Gordon, Keitner, & Ryan, 2000; Sourander et al., 2006; Tamplin & Gooyer, 2001; Tiffin, Pearce, Kaplan, Fundudis, & Parker, 2007) as well as families with psychosocial problems (Keitner, Ryan, Miller, & Zlotnick, 1997; Koyama, Akyama, Miyake, & Kurita, 2004; Tiffin et al., 2007). The popularity of this instrument has been confirmed by the large number of studies it has been used in, but this is the first time the psychometric properties of it have been analyzed in Icelandic for the population of Iceland.

Regarding the clinical implications of the instrument, the FAD has a moderate correlation with other self-administered family functioning instruments, being capable of differentiating significantly between families that have been clinically assessed as functional or dysfunctional (Miller, Keitner, Bishop, & Epstein, 1985; Ryan et al., 2005). Regarding the concurrent validity for the FAD, the relationship between FAD and other family functioning measures such as the Family Adaptability and Cohesion Evaluation Scale (FACES-IV), Family Environment Scale, and Self-Report Family Inventory has been examined in the American culture and have been shown to be highly correlated with one another (Corcoran, 2009; Olson, 2011). To our knowledge, only one validation study has been published on an established family functioning self-report instrument in Iceland and that was for Icelandic FACES-IV. The result of that study indicated that it needed to be further examined before it could be used for Icelandic population (Gu㡲andsdóttir & Björgvinsdóttir, 2011). No other studies have been published regarding other established family functioning scales for the Icelandic population to our knowledge.

There have been discussions regarding studies done on the dimensionality of the FAD scales (Miller et al., 2000; Ridenour, Daley, & Reich, 1999; Shek, 2002). The authors of the MMFF and FAD have said that they want to distinguish between “rational-theoretical” and “internal-consistency” approaches to test construction of the FAD, since it was developed from the “rational-theoretical” approach to measure the constructs of the particular theory that lies behind the MMFF (Miller, Ryan, Keitner, Bishop, & Epstein, 2000a). In that regard, they say that since a factor analytic approach has no prior theory but develops constructs of a scale from the pattern of correlations among the items, they explicitly did not use an internal consistency or factor analytic method in developing the FAD (Miller et al., 2000a). They have also stated that different aspects of family functioning will never be totally independent of each other because problems in one area of family functioning will affect other areas and that is why there can never be the outcome of a study of dimensionalities, on this kind of instrument, that the scales will be totally independent of each other (Epstein et al., 1983).

Ridenour, Daley, and Reich (1999) concluded after doing a new factor analysis on the correlation matrix of the original analysis on the FAD (Epstein et al., 1983) that the seven dimensions of MMFF are not distinct components of the FAD and that subscales of the FAD should be two, GF and BC as well as suggesting that future studies should be done on two new subscales found in their study (Ridenour et al., 1999). In a rejoinder on that article, the authors of the MMFF and FAD disagreed and stated that multiple studies have shown significant differences in results between FAD subscales and ask that the original scoring of the FAD remain the current standard although it could be suggested that the original FAD scales operate as hypothesized and do manifest differences in specific populations and situations (Miller, Ryan, Keitner, Bishop, & Epstein, 2000b). In line with the results of the Ridenour et al. study, other studies of the dimensionality done on the FAD in non-English-speaking countries have not been conclusive: Shek (2002) found a three-factor solution with the Chinese FAD using principal components analysis, Barroilhet, Cano-Prous, Cervera-Enguix, Forjaz, and Guillen-Grima (2009) found a three-factor solution as well with the Spanish FAD using exploratory factor analysis (EFA), in a study on the Dutch version, Wenniger, Hageman, and Arrindell (1993) supported the structure of the FAD, Roncone et al. (1998) found seven factors in the Italian FAD different from the original seven factors using EFA and Chen et al. (2003) found eight factors in the Chinese FAD using confirmatory factor analysis (CFA) and EFA.

Evidence for the reliability of the original FAD scales has been reported in many studies. Epstein, Baldwin, and Bishop (1983) found in the original study of the instrument that the α values of the theoretically derived FAD scales ranged from .72 to .92. In other similar studies of comparative reliability of the FAD, the original scales have been evaluated as well as factor-derived scales (Shek, 2002). Alphas of the original scales have been lowest .20 in the subscale BC (Roncone et al., 1998) and highest in the GF scale .89 (Wenniger et al., 1993).

Aims and Research Questions

This study explores the validity and reliability of an Icelandic version of the FAD in a clinical and nonclinical sample. The study was designed to examine the internal reliability and discriminant validity of the instrument in addition to examining construct validity by using EFA.

Questions addressed in this study:
Research Question 1: Does the instrument measure the same conceptual domains as the original instrument?

Research Question 2: Is the internal consistency as good for the Icelandic version of the FAD as it is for the original FAD?

Research Question 3: Is there a significant difference in the discriminant validity for the nonclinical versus clinical families?

Method

Participants

The participants in the study were categorized into two groups, a clinical group and a nonclinical group. Since the FAD has been shown to differentiate between families living with a child suffering either from mental or medical problems or with a child without problems (Magill-Evans, Darrah, Pain, Adkins, & Kratochvil, 2001; Stein et al., 2000), it was decided to design the research to compare answers of these two groups. The clinical group contained parents of children who were patients at either the Department of child and adolescent psychiatry or the Department of pediatrics at Landspitali University Hospital (n = 289), which are the largest specialized children’s hospital departments in Iceland. These two children’s departments in the hospital service children from birth to 18 years of age and that is the age of the children in the clinical group. Only information regarding the age of the parents was collected for both groups and not the age of the child due to comments from the Data Protection Authority in Iceland.

The nonclinical group contained parents of children between the ages of 14–17 (n = 240) that were recruited through five primary and lower secondary schools by a sample of convenience. The sample was carefully chosen to represent different areas of demographic composition of inhabitants around the capital area based on statistical information from the Icelandic Statistical Office (Hagstofan; Gudjonsdottir, 2012).

Procedure

Participants were recruited during a 1-month period. The parents in the clinical group answered the self-administered instrument in the waiting rooms of the two departments at the Landspitali hospital and turned them in at the admittance desks. The nonclinical group was asked to answer the self-administered instrument at home and return it to the schools, but they also had the option of answering a web-based FAD and skip the paper version. This study was approved by the National Bioethics Committee (VSN11-072) in Iceland and the Data Protection Authority in Iceland (S5269). The families who participated in the study were offered family therapy if requested.

Instrument

The translation and adaptation of the original English language version of the FAD were done in accordance with standard protocols used for intercultural adaptation of assessment instruments (Prince, 2008). The translation and adaptation process consisted of forward translation from English to Icelandic by one of the authors of this article, an experienced English-speaking social worker and family therapist, Hrefna Ólafsdóttir. A pilot study was conducted at the department of child and adolescent psychiatry at Landspitali University Hospital in Iceland in September 2010. Five families participated in the pilot study. The FAD was back translated by a bilingual expert as a part of the translation and adaptation process.

The FAD is a 60-item self-administered instrument, made up of seven subscales. Six of the subscales correspond to the MMFF dimensions of family functioning and one subscale assesses the overall level of the family functioning, called GF (Epstein et al., 1983; Ryan et al., 2005).

The MMFF emphasizes a functional approach to understanding whether families accomplish basic tasks of daily life. The model focuses on the dimensions of family functioning that are seen as having the most impact on the emotional and physical health or problems within families. The functioning within each dimension ranges from most ineffective to most effective. By assessing a family from this perspective, the conclusion can be that “most ineffective” functioning in any of the dimensions can contribute to difficulties on a clinical level. Any family can be evaluated with this approach to determine the effectiveness of its functioning with respect to each dimension (Ryan et al., 2005).

The FAD evaluates family functioning on all dimension of the MMFF by assessing each member’s perception of their family’s functioning (Ryan et al., 2005). The third version of FAD was used in this study, which is designed to be completed by all family members over the age of 12 (Ryan et al., 2005). The items in the instrument consist of general statements about families and the respondent has to decide how well each statement represents his or her own family, using a 4-point Likert-type scale. The answers are scored from 1 to 4, with a higher score indicating poorer function (Ryan et al., 2005). The FAD’s cutoff points for distinguishing dysfunctional from functional families have adequate sensitivity, and in general the reliability of its subscales is acceptable (Miller et al., 1985). In addition to the 60 questions on the form, the participants were asked background questions about both parents in the family, their gender, age range, and years of education.

Data Analysis

For analytical purposes, Epstein and colleague’s (1983) original psychometric study on the FAD was used as a reference. In that study, the authors of the FAD used two groups as well, a nonclinical and a clinical group. The descriptive analysis consists of a frequency analysis for qualitative variables (sex, age, and education) along with means and standard deviation for those variables as well. A p value of ≤.05 was used as level of statistical significance.

As previously has been stated, although factor analysis was deliberately not used to develop the FAD items (Epstein et al., 1983; Miller et al., 2000), findings based on studies of the FAD have been analyzed with factor analysis (Barroilhet et al., 2009; Chen et al., 2003; Kabacoff et al., 1990; Roncone et al., 1998; Shek, 2002; Wenniger et al., 1993). In this study, the authors chose to use solely EFA using subjects from both groups to see how the items of the Icelandic FAD would load on the theoretically developed factors. The intention was also to explore new factors keeping in mind the debates regarding the previous studies of the instrument as well the wishes of the authors of the instrument regarding the dimensionality of the FAD, but they have asked that the original scoring of the FAD remain in the current standard even if factor analysis suggest changes in the instrument (Miller et al., 2000b).

The factor analysis was done according to the α method (Kaiser & Goffrey, 1965) with oblique rotation (promax), in order to find the simplest underlying factor structure (Fabrigar, Wegener, MacCallum, & Strahan, 1999). The general rules of factor analysis (Costello & Osbourne, 2005; Fabrigar et al., 1999; Matsunaga, 2010) were applied for this study. Sample size was acceptable according to the rule of minimum 5–10 cases per measure, which has commonly been recommended (Russel, 2002). In this study, the total sample size was 529 people for the 60-item list, which means 8.8 cases per measure. In line with other studies of the FAD, the 12 items of the GF subscale were excluded from the factor analysis since that subscale is designed to correlate with the other subscales and assumed to be multifactorial (Barroilhet et al., 2009; Roncone et al., 1998, Ryan et al., 2005). Communalities per item were in the range .3–.6 which according to MacCallum, Widaman, Zhang, and Sehee (1999) is in the lower range.

The internal consistency of the full scale and each of the original FAD subscales as well as the subscales of the Icelandic FAD was evaluated with the Cronbach’s α reliability coefficient (Cronbach, 1951). To explore the discriminant validity of the FAD, a comparison of the means and standard deviation was done between the two groups to see which group scored higher (more dysfunctional) in the subscales of the original as well as the Icelandic FAD. An interitem analysis was also done where the two groups, clinical and nonclinical, were compared by their extreme scores on the total scale (TS), the highest and lowest quartiles. Missing items were considered to be missing completely at random and were deleted listwise in the data analysis. The statistical analysis program Statistical Package for the Social Sciences (SPSS) 20.0 was used in this study.

Results

According to the descriptive characteristics of the sample, mothers (female) were the most common respondents in both groups, 75% in the clinical group and 77.9% in the nonclinical. The age range of the respondents was significantly higher in the nonclinical group, 40–45 years, whereas 35–40 years in the clinical group. The years of education were similar in the two groups, ranging from 15 to 17 years, compulsory education in Iceland being 10 years. A question was not addressed if the family was a one- or a two-parent family (Table 1).

Table 1.
Demographic and Social Characteristics of the Groups.

Nonclinical Group Clinical Group p

(n = 240) % (n = 289) %

Gender .819

Female 187 77.9 218 75.4

Male 48 20 53 18.3

Missing data 5 2.1 18 6.2

Age .000*

25–30 years 0 0 35 12.1

30–35 years 9 3.8 59 20.4

35–40 years 48 17.1 69 23.9

40–45 years 61 28.3 63 21.8

45–50 years 101 42.1 46 15.9

Average age range 40–45 years 35–40 years

Missing data 21 8.8 17 5.9

Education in years

Parent A .230

10 years or less 14 5.6 25 10.6

11–14 years 74 30 52 22.1

15–18 years 91 36.8 105 44.6

19 years or more 68 27.5 53 22.5

Average education in years (SD) 17 years (4.1) 16 years (3.5)

Missing data 12 55

Parent B .016*

10 years or less 14 7.4 30 15.5

11–14 years 53 28 49 25.3

15–18 years 80 42.3 74 38.3

19 years or more 42 22.2 40 20.7

Average education in years (SD) 17 years (3.9) 15 years (3.6)

Missing data 35 94

Note. SD = standard deviation.

p ≤ .05, two-tailed.

The factor analysis, shown in Table 2, yielded eight factors where all the items were grouped and which accounted for 49.3% of the total variance. The total of measures per factor which, according to Russel (2002), can be no less than 3, were sufficient. The Kaiser–Meyer–Olkin measure of sampling adequacy was good, 0.913, where the cutoff point is 0.5 and the Bartlett’s test of sphericity was significant (p = .000). The Scree plot, shown in Figure 1, the Keiser criterion, where Eigenvalue over 1 was considered, and subscales’ factor loadings were used to determine number of factors, which showed that eight factors gave the most stable factor loading after analyzing other factor loadings as well, which were more unstable.

Table 2.
Factor Loadings for FAD Obtained Using Exploratory Factor Analysis With α Method of Extraction and Promax Rotation.^a

Items Number Item Name FAD Scale Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Factor 7 Factor 8

Communicate Expressing Emotions Needs and Understanding Cooperation Behavior Family Roles Loyalty and Respect Danger in Family

59 When we don’t like what som, has done, we talk CM 0.868

60 We try to think of diff. ways to solve problems PS 0.838

18 People say things instead of hinting CM 0.700

43 We are frank w.each oth. CM 0.637 0.201

50 We confront problems involving feelings PS 0.592 0.242

29 We talk to people dir. rather than go betweens CM 0.580

38 We resolve most em. upsets that come up. PS 0.507

57 We cry openly AR 0.487 0.463 −0.250 −0.251

3 When someone is upset the others know why CM 0.457 0.211

52 We don’t talk to each other when we are angry CM 0.423 0.219 −0.264

20 We know what to do in an emergency BC 0.408 0.276

2 We resolve most everyday probl. in house PS 0.403

24 After fam. tries to solve probl., we disc. if worked PS 0.365 0.306 0.282 −0.346

12 We usually act on our decisions reg. problems PS 0.330 0.223

9 We are reluctant to show affection for each other AR 0.842 −0.302

28 We don’t show our love for each other AR 0.830

49 We express tenderness AR 0.296 0.681

33 We only get inv. with each oth. when interested AI 0.429 0.245

37 Interest when we can get s.th out of it personally AI −0.218 0.388 0.304

19 Some of us don’t respond emotionally AR 0.377 −0.234

22 It’s diffic. to talk to each oth ab. tender feelings CM 0.281 0.371 0.211 −0.291

39 Tenderness takes second place to other th. in fam. AR 0.364 −0.302 0.311 −0.333

8 We somet. run out of things we need RL 0.730

23 We have trouble meeting our bills RL 0.716 −0.263

58 We don’t have reasonab. transport RL 0.679

34 There’s little time for pers. interests RL 0.339 0.409

42 Fam. shows int. in e. other wh. get someth. out of it AI 0.302 0.367

35 We often don’t say what we mean CM 0.275 0.354

14 You can’t tell how a pers. is feeling from what they say CM 0.256 0.323 0.247 −0.266 0.249

5 If someone is in trouble, others get too involved AI 0.315 0.227 −0.217 0.262

45 If people are asked to do someth., they need reminding RL 0.796 0.282

4 When you ask someone to do someth., you have to check if done RL 0.691 −0.209

15 Family tasks don’t get spread around enough RL 0.493 0.367

13 You only get the interest of others when someth. is important to them AI 0.212 0.459

17 You can easily get away with breaking the rules BC 0.332 0.260 0.246 0.253

44 We don’t hold to any rules or standards BC −0.214 0.922 0.208

27 We have no clear expect. about toilet habits BC −0.201 0.753 0.382

48 Anything goes in our family BC −0.371 0.575 0.277

10 We make sure people meet their fam. responsib RL 0.216 0.624

30 Each of us have particul. duties and responsib. RL 0.621

40 We discuss who is to do househ. jobs RL 0.225 0.586

54 Even though we mean well, we intr too much into each other’s lives AI 0.779

25 We are too self-centered AI 0.581

53 We are generally dissatisfied with family duties assigned to us RL −0.347 0.564

47 If the rules are broken, we don’t know what to expect BC 0.276 0.215 0.334

32 We have rules about hitting people BC 0.343 0.749

55 There are rules about dangerous situations BC −0.216 0.373 0.233 0.387

7 We don’t know what to do when an emergency comes up BC 0.253 0.259 0.357

Note. n = 529. FAD = Family Assessment Device; PS = Problem solving; CM = Communication; RL = Roles; AR = Affective Responsiveness; AI = Affective Involvement; BC = Behavior Control; GF = General Functioning.

^aThe 12 items of GF scale are excluded in the factor analysis.

Figure 1.
Scree plot for exploratory factor analysis.

The eight factors derived from the EFA were different from the six factors in the MMFF model for the original FAD. Factor 1 accounted for 24.48% of the variance and included items 59, 60, 18, 43, 50, 29, 38, 57, 3, 52, 20, 2, 24, 12. This factor (14 items) is called Communicate (CC), since it included mostly items that address CM and PS. The items in this factor are from the original subscales CM (6 items), PS (6 items), and one from BC and AR. Factor 2 accounted for 5.71% of the variance and included Items 9, 28, 49, 33, 37, 19, 22, and 39. This factor (8 items) is called Expressing emotions (EE), since it included mostly items included in the AR subscale (5 of the 6 items that AR contains) but also 3 items regarding interest for each other and communication. Those 3 items are from the original subscales AI (2 items) and one from CM. Factor 3 accounted for 4.3% of the variance and included items 8, 23, 58, 34, 42, 35, 14, and 5. This factor (8 items) is called Needs and understanding (NU), since it included mostly items from the original subscales RL (4 items), CM (2 items), and AI (2 items). Factor 4 accounted for 3.7% of the total variance and included items number 45, 4, 15, 13, and 17. This factor (5 items) is called Cooperation (CO) since it included mostly items regarding RL (3 items) and other items addressing communication from the original subscales AI and BC. Factor 5 accounted for 3.2% of the variance and has 3 items, number 44, 27, and 48. This factor is called Behavior (BH), since it includes only items from the BC subscale of the FAD. Factor 6 accounted for 2.8% of the variance and includes 3 items, number 10, 30, and 40. This factor is called Family roles (FR) since it includes only the positive items from the RL subscale. Factor 7 accounted for 2.6% of the variance and includes 4 items, number 54, 25, 53, and 47. This factor is called Loyalty and respect (LR) since it includes items that address AI (2 items) and respect toward others (1 item from RL and 1 item from BC). Factor 8 accounted for 2.5% of the variance; items number 32, 7, and 55. This factor is called Danger in family (DF) since it includes only items from the BC subscale that addresses dangerous situations in the family.

In Table 3, the comparative reliability of the factors derived from the Icelandic FAD are presented. As Table 3 demonstrates, the comparative reliability of the factors derived in the factor analysis in this study is rather good. For the total sample, subscales CC, EE, NU, and CO scored average, α scores over .7 minimum. The same factors were over .7 in the clinical sample. In the nonclinical sample, the αs were also over .7 minimum in the CC, EE, and NU subscales but instead of the α being over .7 in the CO it was under .7 in CO but over .7 in the FR subscale.

Table 3.
Reliability of the Factors Derived From the Icelandic FAD.

Sample New FAD Subscales

CC EE NU CO BH FR LR DF

Nonclinical sample 0.85 0.79 0.74 0.68 0.54 0.79 0.62 0.39

Clinical sample 0.88 0.73 0.75 0.73 0.59 0.68 0.55 0.26

Total sample 0.87 0.76 0.75 0.71 0.57 0.64 0.59 0.31

Note. FAD = Family Assessment Device; CC = Communicate; EE = Expressing emotions; NU = Needs and understanding; CO = Cooperation; BH = Behavior; FR = Family roles; LR = Loyalty and respect; DF = Danger in family.

Table 4 compares the resulting Cronbach’s α with Epstein’s original psychometric study of the FAD (Epstein et al., 1983). For the total sample, internal consistency for the 60 items TS was high; Cronbach’s α = .95, which suggests that, in general, all items contributed to measuring this construct of family functioning. No items had to be eliminated for discriminant validity purposes.

Table 4.
Reliability of the Original FAD.

Author Sample n FAD Subscales Total Scale

PS CM RL AR AI BC GF

Epstein, Baldwin, and Bishop (1983) Community sample 503 0.74 0.75 0.72 0.83 0.78 0.72 0.92

Present study Nonclinical sample 240 0.76 0.79 0.69 0.73 0.73 0.70 0.86 0.94

Clinical sample 289 0.76 0.81 0.77 0.70 0.72 0.68 0.88 0.95

Total sample 529 0.76 0.81 0.75 0.71 0.73 0.68 0.87 0.95

Note. FAD = Family Assessment Device; PS = Problem solving; CM = Communication; RL = Roles; AR = Affective Responsiveness; AI = Affective Involvement; BC = Behavior Control; GF = General Functioning.

As Table 4 demonstrates, the comparative reliability for the original FAD in this study is rather good. For the total sample, subscales AR, AI, BC as well as the GF subscale, scored average and thus similar to the scores in the original study. For subscales PS, CM, and RL, the score was higher than in the original study.

In other studies of comparative reliability of the original FAD, the αs in PS have lowest been .63 (Barroilhet et al., 2009) and highest .80 (Kabacoff et al., 1990). In CM, the lowest αs have been .45 (Roncone et al., 1998) and highest .79 (Wenniger et al., 1993). In RL, the lowest αs have been .33 (Roncone et al., 1998) and highest .72 in the original study (Epstein et al., 1983). In the AR scale, the lowest αs have been .52 (Roncone et al., 1998) and the highest have been .83 also in the original study by Epstein et al. (1983). In AI, the lowest αs have been .39 (Roncone et al., 1998) and the highest .82 (Barroilhet et al., 2009) when 3 items had been excluded for discriminant validity purposes. In BC, the lowest αs have been .20 (Roncone et al., 1998) and the highest .78 (Barroilhet et al., 2009). The GF scale usually gets rather high αs, the lowest .69 (Roncone et al., 1998) and the highest .92 (Epstein et al., 1983). Finally, the TS has only been measured in few studies, but the lowest α score TS has got is .88 (Roncone et al., 1998) and the highest is .94 (Barroilhet et al., 2009).

Table 5 demonstrates the results of the discriminant validity of the original FAD as well as the subscales derived from the factor analysis done in this study. According to the authors of the FAD, the cutoff score for healthy/unhealthy families varies between the subscales and cultures (Ryan et al., 2005). Significant differences were observed between the parents in the clinical group and nonclinical group using t-test for independent groups in five of the seven subscales of the original FAD and five of the eight in the factor-based FAD.

Table 5.
Discriminant Validity of the Original FAD and the Icelandic FAD With t-Test for Independent Groups.

FAD Subscales Clinical Sample (n = 289) Nonclinical Sample (n = 240) p Cohen’s d

M SD M SD

Problem solving 1.87 0.43 1.76 0.39 .014 .27

Communication 1.95 0.44 1.81 0.62 .002* .22

Roles 2.17 0.44 2.03 0.35 .001* .35

Affective responsiveness 1.71 0.42 1.69 0.46 .486 .05

Affective involvement 1.97 0.43 1.80 0.41 .000* .41

Behavior control 1.74 0.40 1.75 0.43 .483 .02

General functioning 1.74 0.46 1.60 0.41 .007* .32

Communicate 1.85 0.42 1.75 0.35 .025* .26

Expressing emotions 1.80 0.44 1.71 0.46 .052 .20

Needs and understanding 1.89 0.48 1.71 0.43 .01 .40

Cooperation 2.44 0.54 2.25 0.48 .000* .40

Behavior 1.66 0.58 1.74 0.57 .065 .14

Family roles 2.1 0.53 2.02 0.45 .145 .16

Loyalty and respect 2.12 0.52 2.00 0.45 .001* .25

Danger in family 1.67 0.53 1.70 0.59 .389 .05

Note. FAD = Family Assessment Device; SD = standard deviation.

p ≤ .05, two-tailed. p ≤ .01, two-tailed. p ≤ .001, two-tailed.

Effect size was found for the differences using Cohen’s d index. The results indicate that the effect sizes found were small where significance was found, to trivial where there was no significance found. These findings are similar to other studies on the FAD in regard to the discriminant validity of the FAD (Shek, 2002).

Intercorrelations between the original and Icelandic FAD subscales are shown in Tables 6 and 7. Table 6 shows Pearson correlations between the original FAD subscales for both samples and Table 7 shows the same for the factor-derived FAD. Like the tables show, all but one of the predictors has significant correlations with each other and their intercorrelations are all well below .80, so multicollinearity is not a big problem.

Table 6.
Pearson Correlations for the Original FAD Subscales for the Nonclinical (n = 240) and Clinical Sample (n = 289).

PS CM RL AR AI BC

Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clin/nonclinical

CM 0.741/0.693

RL 0.601/0.503 0.656/0.536

AR 0.614/0.534 0.653/0.662 0.489/0.459

AI 0.473/0.390 0.565/0.590 0.589/0.540 0.535/0.633

BC 0.609/0.503 0.616/0.484 0.509/0.478 0.489/0.495 0.452/0.516

GF 0.752/0.667 0.807/0.774 0.698/0.633 0.749/0.761 0.638/0.646 0.593/0.546

Note. FAD = Family Assessment Device; PS = Problem solving; CM = Communication; RL = Roles; AR = Affective Responsiveness; AI = Affective Involvement; BC = Behavior Control; GF = General Functioning.

Correlation is significant at the .01 level (two-tailed).

Table 7.
Pearson Correlations for the Factors Derived From the Icelandic FAD for the Nonclinical (n = 240) and Clinical Sample (n = 289).

CO CC EE NU BH FR LR

Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical Clinical/nonclinical

CC 0.519/504

EE 0.438/0.462 0.644/0.646

NU 0.485/0.549 0.504/0.539 0.618/0.611

BH 0.217/0.237 0.324/0.399 0.415/0.370 0.368/0.294

FR 0.380/0.466 0.449/0.626 0.215/0.357** 0.156/0.376 0.290/0.322

LR 0.481/0.387 0.508/0.344 0.481/0.423 0.514/0.405 0.274/0.232 0.191/0.241

DF 0.201/0.125 0.353/0.452 0.243/0.325 0.262/0.250 0.362/0.327 0.232/0.337 0.128/0.210**

Note. FAD = Family Assessment Device; CC = Communicate; EE = Expressing emotions; NU = Needs and understanding; CO = Cooperation; BH = Behavior; FR = Family roles; LR = Loyalty and respect; DF = Danger in family.

*Correlation is significant at the .05 level (two-tailed). **Correlation is significant at the .01 level (two-tailed).

Discussion and Applications to Social Work Practice

In this study, an analysis of the psychometric properties of an Icelandic version of the FAD was conducted. The frequency analysis for qualitative variables showed that the age range of the parents in this study was significantly higher in the nonclinical group, 40–45, whereas it was 35–40 in the clinical group. This difference can be explained by the age range of the children in the nonclinical group which was fixed at 14–17, but the age of the children in the clinical group is from newborn to 18. There is a possibility that this difference can impact the results of this study and therefore it would be better to fix the ages of all the participants in future studies, so that the results will be more reliable.

The first research question for this study was if the Icelandic FAD measured the same conceptual domains as the original instrument does. In order to answer that question, it was important to clarify the Icelandic FAD’s factor structure. In an effort to do that, an EFA was conducted on the Icelandic FAD, considering all subjects, in order to evaluate if the translated FAD measures the same conceptual domains as the original FAD. The eight factors extracted in the EFA in this study were not the same as in the original measure and differed also in numbers, since in the original FAD, the factors normally used in the factor analysis are 6, but in this study, the factors were 8. The factors in this study are CC, EE, NU, CO, BH, FR, LR, and DF.

These results are in line with other similar cross-cultural studies that have used factor analysis on the FAD in non-English-speaking countries. In many of those studies, the results have not been conclusive, showing different factor solutions, from three factors (Barroilhet et al., 2009; Shek, 2002) to eight factors (Chen et al., 2003) using different methods of factor analysis.

The authors of the FAD have stated that there can be cultural differences in specific populations, which possibly are what we are seeing here. Although the factors derived from the EFA in this study do not show the same factors as the original FAD does, it is our opinion that the conceptual domains are there, based on the fact that many of the items that group together in the factor analysis done in this study are grouped together in the MMFF model and the original FAD (Epstein et al., 1983). Further analysis must be done to assess this question.

The second research question of this study is if the internal consistency is as good for the Icelandic version of the FAD as it is for the original FAD. The comparative reliability of the factors derived in the factor analysis in this study was analyzed. The results are that the α values of the Icelandic FAD scales ranged from .26 to .88, the lowest αs being in the subscale DF and the highest in CM. The αs in four of the subscales were acceptable but not in four of them, although they were found to be internally consistent. This means that although the α values in half of the factor-derived subscales for the Icelandic FAD in this study were generally acceptable, it is not enough to merit the use of these factors without further studies.

The comparative reliability of the original FAD is also assessed and compared to the original study of the FAD by Epstein et al. (1983). For the AR, AI, BC, and GF subscales, the score is lower or similar to the original study (Epstein et al., 1983), but higher than in many other similar studies (Barroilhet et al., 2009; Chen et al., 2003; Kazarian, 2009; Roncone et al., 1998; Shek, 2002; Wenniger et al., 1993). Similarly, the BC subscale is the lowest in the clinical and total samples in this study, which is similar to other studies of the original subscales of the FAD where the α for BC has also been the lowest (Roncone et al., 1998; Shek, 2002). For subscales PS, CM, and RL, the score was higher than in the original study (Epstein et al., 1983) and in fact the α for the CM subscale was higher than it has been in other similar studies of the FAD (Barroilhet et al., 2009; Chen et al., 2003; Kazarian, 2009; Roncone et al., 1998; Shek, 2002; Wenniger et al., 1993) and the α for RL in the clinical sample and the total sample was also higher than it has been in other similar studies (Barroilhet et al., 2009; Chen et al., 2003; Kazarian, 2009; Roncone et al., 1998; Shek, 2002; Wenniger et al., 1993). The subscales RL in the nonclinical group and BC in the clinical group and for the total sample the scores are under a minimum of .7 and need to be revised in further studies in order to try to get a higher internal consistency. The internal consistency for the TS was high; Cronbach’s α = .95 and that is also the highest score the TS scale has got in other similar studies of the FAD (Barroilhet et al., 2009; Chen et al., 2003; Kazarian, 2009; Roncone et al., 1998; Shek, 2002; Wenniger et al., 1993). This suggests that, in general, all items contributed to measuring the theoretically based construct of family functioning. These results indicate that the Icelandic version of the theoretically based original FAD is reliable and thus merits continued use, in line with the authors’ goals.

The third research questions asked if there is a significant difference in the discriminant validity for the nonclinical versus clinical families. The result of the study of discriminant validity of the FAD is that significant differences were observed between the parents in the clinical group and nonclinical group in five of the seven subscales of the original FAD and five of the eight in the Icelandic FAD. Over all, the clinical group had higher mean score (more dysfunctional) in all the subscales except BC of the original FAD and BH and DF subscales of the factor-based FAD. These findings are similar to other studies on the FAD (Barroilhet et al., 2009; Chen et al., 2003; Kazarian, 2009; Roncone et al., 1998; Shek, 2002; Wenniger et al., 1993). This means that the Icelandic FAD seems to be sensitive to family function in clinical and nonclinical families in Iceland.

The reason for the relatively small difference between the two groups in the discriminant validity analysis can be that the Icelandic nation as a whole is small, the population counting 321.857 (Statistics Iceland, 2014) and homogenous, where 84.5% of the population is of Icelandic decent (Statistics Iceland, 2014). So a conclusion can be drawn that there are some customs, rules, and behaviors within families that are the same throughout the country. There is also a timing factor to consider, since this study was done shortly after the biggest financial crisis the Icelandic nation has gone through and it is possible that the effects of that shock was having an impact on the way both groups answered (Gudmundsdottir, 2013).

Consideration should be given to the limitations of this study, which reside in the fact that no screening tools were used to measure the parents and children’s individual health, no questions were asked about other children in the household or the size of the family in general, and temporal stability was not measured. Future studies should also include CFA of the Icelandic FAD introduced in this study.

Social workers use assessment tools for their work frequently, but it is not common that they are using tools that are assessed by them, from their professional point of view, for their work. Family functioning is an interesting phenomenon to study for social workers, in clinical and nonclinical families alike, but few studies are available that focus on family functioning in Icelandic families that are done using a self-report instrument that measures family functioning. One way to overcome lack of research regarding family functioning in Iceland is to translate and validate instruments that have been created in other countries and have proved their usefulness in the field of clinical family therapy and research. That was the focus of this study, in order to have a validated instrument to use in future studies on families, for prevention purposes, assessment of service, and to obtain information regarding situations of certain groups within the society.

Another focus of this study was to contribute to the field of social work research in Iceland, since it is important, for social workers who want to work within an evidence-based framework, to use assessment tools that help them to evaluate process and outcome for their work. Evidence-based practice means that one relies on the best scientific evidence that is currently available (McNeece & Thyer, 2004). That includes using assessment tools like self-report measurements that have been evaluated regarding their psychometric properties in its relevant culture. In the spirit of evidence-based practice, it is also important that social workers prepare the intervention that has been proposed very well with careful assessment. They need to have clear and measurable treatment goals, develop a treatment plan based on those goals, and monitor progress toward them frequently, for example, by assessing the progress with a self-report instrument like FAD and modify or end treatment as needed (McNeece & Thyer, 2004). In this kind of work, it is important to have good assessment tools.

In order for a self-report instrument like FAD to be used by social workers in the field, it is important to point a few things out. When social workers are wondering if they should use a self-report questionnaire to assess family functioning they have to consider a few things. Self-report measurements of families provide important information about subjective things like feelings, perspective, and ability of the family members, from themselves, that will be hard to get with general questioning or observation. These kinds of measurements can thus help with the assessment of the family and help the social worker in deciding on a treatment plan for the family. In addition, these measurements can assess outcome of treatment, for example, family therapy or parental training and thus validate the decision of the implementation to provide further funding for that treatment (Corcoran, 2009). As was mentioned earlier in this article, this kind of self-report measurement, which evaluates family functioning, would be useful for social workers working with families in child and adolescent psychiatric settings, adult psychiatric settings, in child protection services, or with families that are experiencing any kind of medical illnesses where one in the family is sick, a child or an adult. It has also been used in settings where there has been trauma in the family like divorce (Brown, Eichenberg, Portes, & Christensen, 1991) or when a family has lived through major disasters like war or natural disasters (Kiliç et al., 2003).

The FAD is a widely used and validated instrument with vast appeal to many dimensions of family studies, which focuses on family functioning from a multidimensional focus. In this article, the results of the first adaptation of the Icelandic FAD were presented. The results of this study are promising, as they imply that the Icelandic FAD is both valid and reliable and thus merit continued application.

	Nonclinical Group	Clinical Group	p
Gender	.819
Female	187	77.9	218	75.4
Male	48	20	53	18.3
Missing data	5	2.1	18	6.2
Age	.000*
25–30 years	0	0	35	12.1
30–35 years	9	3.8	59	20.4
35–40 years	48	17.1	69	23.9
40–45 years	61	28.3	63	21.8
45–50 years	101	42.1	46	15.9
Average age range	40–45 years		35–40 years
Missing data	21	8.8	17	5.9
Education in years
Parent A	.230
10 years or less	14	5.6	25	10.6
11–14 years	74	30	52	22.1
15–18 years	91	36.8	105	44.6
19 years or more	68	27.5	53	22.5
Average education in years (SD)	17 years (4.1)		16 years (3.5)
Missing data	12		55
Parent B	.016*
10 years or less	14	7.4	30	15.5
11–14 years	53	28	49	25.3
15–18 years	80	42.3	74	38.3
19 years or more	42	22.2	40	20.7
Average education in years (SD)	17 years (3.9)		15 years (3.6)
Missing data	35		94

Sample	New FAD Subscales
Nonclinical sample	0.85	0.79	0.74	0.68	0.54	0.79	0.62	0.39
Clinical sample	0.88	0.73	0.75	0.73	0.59	0.68	0.55	0.26
Total sample	0.87	0.76	0.75	0.71	0.57	0.64	0.59	0.31

Author	Sample	n	FAD Subscales	Total Scale
Epstein, Baldwin, and Bishop (1983)	Community sample	503	0.74	0.75	0.72	0.83	0.78	0.72	0.92
Present study	Nonclinical sample	240	0.76	0.79	0.69	0.73	0.73	0.70	0.86	0.94
Clinical sample	289	0.76	0.81	0.77	0.70	0.72	0.68	0.88	0.95
Total sample	529	0.76	0.81	0.75	0.71	0.73	0.68	0.87	0.95

FAD Subscales	Clinical Sample (n = 289)	Nonclinical Sample (n = 240)	p	Cohen’s d
Problem solving	1.87	0.43	1.76	0.39	.014*	.27
Communication	1.95	0.44	1.81	0.62	.002*	.22
Roles	2.17	0.44	2.03	0.35	.001***	.35
Affective responsiveness	1.71	0.42	1.69	0.46	.486	.05
Affective involvement	1.97	0.43	1.80	0.41	.000***	.41
Behavior control	1.74	0.40	1.75	0.43	.483	.02
General functioning	1.74	0.46	1.60	0.41	.007*	.32
Communicate	1.85	0.42	1.75	0.35	.025*	.26
Expressing emotions	1.80	0.44	1.71	0.46	.052**	.20
Needs and understanding	1.89	0.48	1.71	0.43	.01**	.40
Cooperation	2.44	0.54	2.25	0.48	.000***	.40
Behavior	1.66	0.58	1.74	0.57	.065	.14
Family roles	2.1	0.53	2.02	0.45	.145	.16
Loyalty and respect	2.12	0.52	2.00	0.45	.001***	.25
Danger in family	1.67	0.53	1.70	0.59	.389	.05

	PS	CM	RL	AR	AI	BC
CM	0.741/0.693
RL	0.601/0.503	0.656/0.536
AR	0.614/0.534	0.653/0.662	0.489/0.459
AI	0.473/0.390	0.565/0.590	0.589/0.540	0.535/0.633
BC	0.609/0.503	0.616/0.484	0.509/0.478	0.489/0.495	0.452/0.516
GF	0.752/0.667	0.807/0.774	0.698/0.633	0.749/0.761	0.638/0.646	0.593/0.546

	CO	CC	EE	NU	BH	FR	LR
CC	0.519/504
EE	0.438/0.462	0.644/0.646
NU	0.485/0.549	0.504/0.539	0.618/0.611
BH	0.217/0.237	0.324/0.399	0.415/0.370	0.368/0.294
FR	0.380/0.466	0.449/0.626	0.215/0.357	0.156/0.376*	0.290/0.322
LR	0.481/0.387	0.508/0.344	0.481/0.423	0.514/0.405	0.274/0.232	0.191/0.241
DF	0.201*/0.125	0.353/0.452	0.243/0.325	0.262/0.250	0.362/0.327	0.232/0.337	0.128/0.210**

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported in part by a grant from the Icelandic Association of Social Workers (Félagsrá㦪afafélag Íslands). It also received assistance from Specialisterne, a work program for adults with autism in Iceland.

References

Addington

Collins

McCleery

Addington

(2005). The role of family work in early psychosis. Schizophrenia Research, 79, 77–83. doi:10.1016/j.schres.2005.01.013

Al-Krenawi

Graham

J. R.

Slonim-Nevo

(2002). Mental health aspects of Arab-Israeli adolescents from polygamous versus monogamous families. The Journal of Social Psychology, 142, 446–460. doi:10.1080/00224540209603911

Al-Krenawi

Lev-Wiesel

Sehwail

A. M.

(2007). Psychological symptomatology among Palestinian male and female adolescents living under political violence 2004-2005. Community Mental Health Journal, 43, 49–56.

Barroilhet

Cano-Prous

Cervera-Enguix

Forjaz

M. J.

Guillén-Grima

(2009). A Spanish version of the family assessment device. Social Psychiatry and Psychiatric Epidemiology, 44, 1051–1065. doi:10.1007/s00127-009-0022-8

Bhui

Mohamud

Warfa

Craig

T. J.

Stansfeld

S. A.

(2003). Cultural adaptation of mental health measures: Improving the quality of clinical practice and research (editorial). The British Journal of Psychiatry, 183, 184–186.

Brown

J. H.

Eichenberg

S. A.

Portes

P. R.

Christensen

D. N.

(1991). Family functioning factors associated with the adjustment of children of divorce. Journal of Divorce and Remarriage, 17, 81–95.

Celikel

F. C.

Cumurcu

B. E.

Koc

Etikan

Yucel

(2008). Psychologic correlates of eating attitudes in Turkish female college students. Comprehensive Psychiatry, 49, 188–194. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/j.comppsych.2007.09.003

Chen

J. L.

Kennedy

Kools

Slaughter

R. E.

Franck

Kong

S. K. F.

Wong

T. K. S.

(2003). Culturally appropriate family assessment: Analysis of the family assessment device in a pediatric Chinese population. Journal of Nursing Measurement, 11, 41–59.

Chen

J. Y.

Clark

M. J.

(2007). Family function in families of children with Duchenne muscular dystrophy. Family and Community Health, 30, 296–304. doi:10.1097/01.FCH.0000290542.10458.f8

10.

Corcoran

(2009). Using standardized tests and instruments in family assessment. In Roberts

R. A.

(Ed.), Social work desk reference (2nd ed., pp. 390–394). New York, NY: Oxford university press.

11.

Costello

A. B.

Osborne

J. W.

(2005). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Practical Assessment, Research & Evaluation, 10. Retrieved from http://pareonline.net/getvn.asp?v=10&n=7

12.

Cronbach

L. J.

(1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297–334.

13.

Derisley

Libby

Clark

Reynolds

(2005). Mental health, coping and family-functioning in parents of young people with obsessive-compulsive disorder and with anxiety disorders. The British Journal of Clinical Psychology, 44, 439–444.

14.

Epstein

N. B.

Baldwin

L. M.

Bishop

D. S.

(1983). The McMaster family assessment device. Journal of Marital and Family Therapy, 9, 171–186. doi:10.1111/j.1752-0606.1983.tb01497.x

15.

Erol

Yazici

Toprak

(2007). Family functioning of patients with an eating disorder compared with that of patients with obsessive compulsive disorder. Comprehensive Psychiatry, 48, 47–50.

16.

Fabrigar

L. R.

Wegener

D. T.

MacCallum

R. C.

Strahan

E. J.

(1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4, 272–299. doi:1082-989X/99/S3.00

17.

Goldenberg

(2008). Family therapy, an overview (7th ed.). Belmont, CA: Brooks/Cole, Cengage learning.

18.

Gu㡲andsdóttir

Björgvinsdóttir

S. D.

(2011). Próffræ㨬egir eiginleikar íslenskrar útgáfu Family Adaptability and Cohesion Evaluation Scale-IV. [Psychometric properties of an Icelandic version of FACES-IV] (master’s thesis). Retrieved from www.skemman.is; http://hdl.handle.net/1946/8534

19.

Gudjonsdottir

(Ed.). (2012). Iceland in figures 2012. Statistic Iceland (vol. 17). Retrieved October 22, 2012, from https://hagstofa.is/lisalib/getfile.aspx?ItemID=13913

20.

Gudmundsdottir

D. G.

(2013). The impact of economic crisis on happiness. Social Indicators Research, 110, 1083–1101. doi:10.1007/s11205-011-9973-8

21.

Hanks

R. A.

Rapport

L. J.

Vangel

(2007). Caregiving appraisal after traumatic brain injury: The effects of functional status, coping style, social support and family functioning. Neuro Rehabilitalion, 22, 43–52.

22.

Houtzager

B. A.

Oort

F. J.

Hoekstra-Weebers

Caron

H. N.

Grootenhuis

M. A.

Last

B. F.

(2004). Coping and family functioning predict longitudinal psychological adaptation of siblings of childhood cancer patients. Journal of Pediatric Psychology, 29, 591–605. doi:10.1093/jpepsy/jsh061

23.

Kabacoff

R. I.

Miller

I. W.

Bishop

S. D.

Epstein

N. B.

Keitner

I. G.

(1990). A psychometric study of the McMaster family assessment device in psychiatric, medical and nonclinical samples. Journal of Family Psychology, 3, 431–439.

24.

Kaiser

H. F.

Goffrey

(1965). Alpha factor analysis. Psychometrika, 30, 1–14.

25.

Kankaraš

Moors

(2010). Researching measurement equivalence in cross-cultural studies. Psihologija, 43, 212–136. doi:10.2298/PSI1002121K

26.

Kazarian

S. S.

(2009). Cultural appropriateness of the Family Assessment Device (FAD) in the case of Ethnic Armenian adolescents in Lebanon. International Journal of Social Psychiatry, 56, 230. doi:10.1177/0020764009105646

27.

Keitner

G. I.

Fodor

Ryan

C. E.

Miller

I. W.

Epstein

N. B.

Bishop

D. S.

(1991). A cross-cultural study of major depression and family functioning. Canadian Journal of Psychiatry, 36, 254–258.

28.

Keitner

G. I.

Ryan

C. E.

Miller

I. W.

Kohn

Bishop

D. S.

Epstein

N. B.

(1995). Role of the family in recovery and major depression. American Journal of Psychiatry, 152, 1002–1008.

29.

Keitner

G. I.

Ryan

C. E.

Miller

I. W.

Zlotnick

(1997). Psychosocial factors and the long-term course of major depression. Journal of Affective Disorders, 44, 57–67.

30.

Kiliç

E. Z.

Ozguven

H. D.

Sayil

(2003). The psychological effects of parental mental health on children experience of Bolu earthquake in Turkey. Family Process, 42, 485–495.

31.

Koyama

Akyama

Miyake

Kurita

(2004). Family functioning perceived by patients and their family members in three diagnostic and statistical manual-IV diagnostic groups. Psychiatry and Clinical Neuroscience, 58, 495–500.

32.

Lauth

Magnússon

Ferrari

Pétursson

(2008). An Icelandic version of the Kiddie-SADS-PL: Translation, cross-cultural adaptation and inter-rater reliability. Nordic Journal of Psychiatry, 62, 379–385. doi:10.1080/08039480801984214

33.

Lieb

Wittchen

H. U.

Hofler

Fuetsch

Stein

M. B.

Merikangas

K. R.

(2000). Parental psychopathology, parenting styles and the risk of social phobia in offspring; a prospective-longitudinal community study. Archives of General Psychiatry, 57, 859–866.

34.

MacCallum

R. C.

Widaman

K. F.

Zhang

Sehee

(1999). Sample size in factor analysis. Psychological Methods, 4, 84–99. doi:1082-989X/99/S3.00

35.

Magill-Evans

Darrah

Pain

Adkins

Kratochvil

(2001). Are families with adolescents and young adults with cerebral palsy the same as other families? Developmental Medicine and Child Neurology, 43, 466–472. doi:10.1111/j.1469-8749.2001.tb00744.x

36.

Malec

J. F.

Testa

J. A.

Rush

B. K.

Brown

A. W.

Moessner

A. M.

(2007). Self-assessment of impairment, impaired self-awareness and depression after traumatic brain injury. Journal of Head and Trauma Rehabilitation, 22, 156–166. doi:10.1097/01.HTR.0000271116.12028.af

37.

Matsunaga

(2010). How to factor-analyze your data right: Do’s, Dont’s and How-to’s. International Journal of Psychological Research, 3, 97–110. Retrieved from http://mvint.usbmed.edu.co:8002/ojs/index.php/web/article/view/464

38.

McNeece

A. C.

Thyer

B. A.

(2004). Evidence-based practice and social work. Journal of Evidence Based Social Work, 1, 7–25. doi:10.1300/J394v01n01_02

39.

Miller

I. W.

Bishop

D. S.

Herman

D. S.

Stein

M. D.

(2007). Relationship quality among HIV patients and their caregivers. AIDS Care, 19, 203–211.

40.

Miller

I. W.

Keitner

G. I.

Bishop

D. S.

Epstein

N. B.

(1985). The McMaster family assessment device; reliability and validity. Journal of Marital and Family Therapy, 11, 345–356.

41.

Miller

I. W.

McDermut

Gordon

K. C.

Keitner

G. I.

Ryan

C. E.

(2000). Personality and family functioning in families of depressed patients. Journal of Abnormal Psychology, 109, 539–545.

42.

Miller

I. W.

Ryan

C. E.

Keitner

G. I.

Bishop

D. S.

Epstein

N. B.

(2000a). Commentary: “Factor analysis of the family assessment device”, by Ridenour, Daley, & Reich. Family Process, 39, 141–144.

43.

Miller

I. W.

Ryan

C. E.

Keitner

G. I.

Bishop

D. S.

Epstein

N. B.

(2000b). Why fix what isn’t broken? A rejoinder to Ridenour, Daley, & Reich. Family Process, 39, 381–384. doi:10.1111/j.1545-5300.2000.39309.x

44.

Olson

D. H.

(2011). FACES-IV & the circumplex model: Validation study. Journal of Marital & Family therapy, 37, 64–80. doi:10.1111/j.1752-0606.2009.00175.x

45.

Prince

(2008). Measurement validity in cross-cultural comparative research. Epidemiologia e Psichiatria Sociale, 17, 211–220.

46.

Ridenour

T. A.

Daley

Reich

(1999). Factor analyses of the Family Assessment Device. Family process, 38, 497–510. doi:10.1111/j.1545-5300.1999.00497.x

47.

Roncone

Rossi

Muiere

Impallomeni

Matteucci

Giacomelli

… Casacchia

(1998). The Italian version of the Family Assessment Device. Social Psychiatry and Psychiatric Epidemiology, 33, 451–461.

48.

Russel

D.W.

(2002). In search of underlying dimensions: The use (and abuse) of factor analysis in personality and social psychology bulletin. Personality and Social Psychology Bulletin, 28, 629.

49.

Ryan

C. E.

Epstein

N. B.

Keitner

G. I.

Miller

I. W.

Bishop

D. S.

(2005). Evaluating and treating families: The McMaster approach. New York, NY: S. Routledge.

50.

Schmitt

Piha

Helenius

Baldus

Kienbacher

Steck

… Romer

(2008). Multinational study of cancer patients and their children; Factors associated with family functioning. Journal of Clinical Oncology, 26, 5877–5833. doi:10.1200/JCO.2007.12.8132

51.

Shek

D. T. L.

(2002). Assessment of family functioning in Chinese adolescents: The Chinese version of the family assessment device. Research in Social Work Practice, 12, 502–524. doi:10.1177/1049731502012004003

52.

Sourander

Pihlakoski

Aromaa

Rautava

Helenius

Sillanpää

(2006). Early predictors of parent-and self-reported perceived global psychological difficulties among adolescents; a prospective cohort study from age 3 to age 15. Social Psychiatry and Psychiatric Epidemiology, 41, 173–182. doi:10.1007/s00127-005-0013-3

53.

Statistics Iceland. (2014). Population overview. Retrieved from http://www.statice.is/Statistics/Population

54.

Stein

Williamson

D. E.

Birmaher

Brent

D. A.

Kaufman

Dahl

R. E.

… Ryan

N. D.

(2000). Parent-child bonding and family functioning in depressed children and children at high risk and low risk for future depression. Journal of American Academy of Child and Adolescent Psychiatry, 39, 1387–1395. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1097/00004583-200011000-00013

55.

Stevenson-Hinde

Akister

. (1995). The McMaster model of family functioning: Observer and parental ratings in a non-clinical sample. Family Process, 34, 337–347. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/S0165-0327(97)00105-5

56.

Tamplin

Gooyer

I. M.

(2001). Family functioning in adolescents at high and low risk for major depressive disorder. European Child & Adolescent Psychiatry, 10, 170–179. doi:10.1007/s007870170023

57.

Tiffin

P. A.

Pearce

Kaplan

Fundudis

Parker

(2007). The impact of socio-economic status and mobility on perceived family functioning. Journal of Family and Economic Issues, 28, 653–667. doi:10.1007/s10834-007-9077-4

58.

Wenniger

W. F. M. D. B.

Hageman

W. J. J. M.

Arrindell

W. A.

(1993). Cross-national validity of dimensions of family functioning: First experiences with the Dutch version of the McMaster Family Assessment Device (FAD). Personality and Individual Differences, 14, 769–781. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/0191-8869(93)90090-P

	PS	CM	RL	AR	AI	BC
	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clin/nonclinical
CM	0.741/0.693
RL	0.601/0.503	0.656/0.536
AR	0.614/0.534	0.653/0.662	0.489/0.459
AI	0.473/0.390	0.565/0.590	0.589/0.540	0.535/0.633
BC	0.609/0.503	0.616/0.484	0.509/0.478	0.489/0.495	0.452/0.516
GF	0.752/0.667	0.807/0.774	0.698/0.633	0.749/0.761	0.638/0.646	0.593/0.546

	CO	CC	EE	NU	BH	FR	LR
	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical	Clinical/nonclinical
CC	0.519/504
EE	0.438/0.462	0.644/0.646
NU	0.485/0.549	0.504/0.539	0.618/0.611
BH	0.217/0.237	0.324/0.399	0.415/0.370	0.368/0.294
FR	0.380/0.466	0.449/0.626	0.215/0.357	0.156/0.376*	0.290/0.322
LR	0.481/0.387	0.508/0.344	0.481/0.423	0.514/0.405	0.274/0.232	0.191/0.241
DF	0.201*/0.125	0.353/0.452	0.243/0.325	0.262/0.250	0.362/0.327	0.232/0.337	0.128/0.210**

	Nonclinical Group		Clinical Group		p
	(n = 240)	%	(n = 289)	%	p
Gender					.819
Female	187	77.9	218	75.4
Male	48	20	53	18.3
Missing data	5	2.1	18	6.2
Age					.000*
25–30 years	0	0	35	12.1
30–35 years	9	3.8	59	20.4
35–40 years	48	17.1	69	23.9
40–45 years	61	28.3	63	21.8
45–50 years	101	42.1	46	15.9
Average age range	40–45 years		35–40 years
Missing data	21	8.8	17	5.9
Education in years
Parent A					.230
10 years or less	14	5.6	25	10.6
11–14 years	74	30	52	22.1
15–18 years	91	36.8	105	44.6
19 years or more	68	27.5	53	22.5
Average education in years (SD)	17 years (4.1)		16 years (3.5)
Missing data	12		55
Parent B					.016*
10 years or less	14	7.4	30	15.5
11–14 years	53	28	49	25.3
15–18 years	80	42.3	74	38.3
19 years or more	42	22.2	40	20.7
Average education in years (SD)	17 years (3.9)		15 years (3.6)
Missing data	35		94

Items Number	Item Name	FAD Scale	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Factor 6	Factor 7	Factor 8
Items Number	Item Name	FAD Scale	Communicate	Expressing Emotions	Needs and Understanding	Cooperation	Behavior	Family Roles	Loyalty and Respect	Danger in Family
59	When we don’t like what som, has done, we talk	CM	0.868
60	We try to think of diff. ways to solve problems	PS	0.838
18	People say things instead of hinting	CM	0.700
43	We are frank w.each oth.	CM	0.637	0.201
50	We confront problems involving feelings	PS	0.592	0.242
29	We talk to people dir. rather than go betweens	CM	0.580
38	We resolve most em. upsets that come up.	PS	0.507
57	We cry openly	AR	0.487	0.463	−0.250		−0.251
3	When someone is upset the others know why	CM	0.457		0.211
52	We don’t talk to each other when we are angry	CM	0.423						0.219	−0.264
20	We know what to do in an emergency	BC	0.408		0.276
2	We resolve most everyday probl. in house	PS	0.403
24	After fam. tries to solve probl., we disc. if worked	PS	0.365			0.306	0.282		−0.346
12	We usually act on our decisions reg. problems	PS	0.330			0.223
9	We are reluctant to show affection for each other	AR		0.842					−0.302
28	We don’t show our love for each other	AR		0.830
49	We express tenderness	AR	0.296	0.681
33	We only get inv. with each oth. when interested	AI		0.429			0.245
37	Interest when we can get s.th out of it personally	AI	−0.218	0.388	0.304
19	Some of us don’t respond emotionally	AR		0.377				−0.234
22	It’s diffic. to talk to each oth ab. tender feelings	CM	0.281	0.371		0.211		−0.291
39	Tenderness takes second place to other th. in fam.	AR		0.364	−0.302	0.311				−0.333
8	We somet. run out of things we need	RL			0.730
23	We have trouble meeting our bills	RL			0.716		−0.263
58	We don’t have reasonab. transport	RL			0.679
34	There’s little time for pers. interests	RL		0.339	0.409
42	Fam. shows int. in e. other wh. get someth. out of it	AI		0.302	0.367
35	We often don’t say what we mean	CM	0.275		0.354
14	You can’t tell how a pers. is feeling from what they say	CM	0.256		0.323		0.247	−0.266		0.249
5	If someone is in trouble, others get too involved	AI			0.315	0.227	−0.217		0.262
45	If people are asked to do someth., they need reminding	RL				0.796		0.282
4	When you ask someone to do someth., you have to check if done	RL				0.691	−0.209
15	Family tasks don’t get spread around enough	RL				0.493		0.367
13	You only get the interest of others when someth. is important to them	AI		0.212		0.459
17	You can easily get away with breaking the rules	BC				0.332	0.260	0.246	0.253
44	We don’t hold to any rules or standards	BC			−0.214		0.922			0.208
27	We have no clear expect. about toilet habits	BC	−0.201				0.753			0.382
48	Anything goes in our family	BC				−0.371	0.575		0.277
10	We make sure people meet their fam. responsib	RL				0.216		0.624
30	Each of us have particul. duties and responsib.	RL						0.621
40	We discuss who is to do househ. jobs	RL	0.225					0.586
54	Even though we mean well, we intr too much into each other’s lives	AI							0.779
25	We are too self-centered	AI							0.581
53	We are generally dissatisfied with family duties assigned to us	RL		−0.347					0.564
47	If the rules are broken, we don’t know what to expect	BC				0.276	0.215		0.334
32	We have rules about hitting people	BC					0.343			0.749
55	There are rules about dangerous situations	BC				−0.216	0.373	0.233		0.387
7	We don’t know what to do when an emergency comes up	BC			0.253				0.259	0.357