The Brief Family Relationship Scale

Abstract

The Relationship dimension of the Family Environment Scale, which consists of the Cohesion, Expressiveness, and Conflict subscales, measures a person’s perception of the quality of his or her family relationship functioning. This study investigates an adaptation of the Relationship dimension of the Family Environment Scale for Alaska Native youth. The authors tested the adapted measure, the Brief Family Relationship Scale, for psychometric properties and internal structure with 284 12- to 18-year-old predominately Yup’ik Eskimo Alaska Native adolescents from rural, remote communities. This non-Western cultural group is hypothesized to display higher levels of collectivism traditionally organized around an extended kinship family structure. Results demonstrate a subset of the adapted items function satisfactorily, a three-response alternative format provided meaningful information, and the subscale’s underlying structure is best described through three distinct first-order factors, organized under one higher order factor. Convergent and discriminant validity of the Brief Family Relationship Scale was assessed through correlational analysis.

Keywords

Alaska Native cohesion expressiveness conflict family relationships family climate

There is need for valid measures of family functioning in research with youth on a number of topics across diverse cultural settings, including family relations, resilience, and protection from substance abuse and psychopathology. We report here on a cultural adaptation of the Relationship dimension of the Family Environment Scale (FES; Moos & Moos, 1994) with a sample of rural Yup’ik Alaska Native youth. The FES has been criticized for low reliabilities when used with youth and people from non-Western cultural backgrounds (e.g., Ma & Leung, 1990; Roosa & Beals, 1990; Sanford, Raymond, & Zucker, 1999). The objectives of this study were to (a) assess the scale structure of the Brief Family Relationship Scale (BFRS), (b) investigate the item characteristics of the item pool for their functioning and for optimal response-level calibration for Alaska Native youth, and (c) assess the evidence for validity of the BFRS score interpretations.

Method

Participants

Participants were 284 12- to 18-year-old Alaska Natives from rural, remote Alaska communities. Most (194) were recruited from a boarding school in Southeast Alaska, with the remainder coming from a predominately Yup’ik regional hub community in Southwestern Alaska. The sample was 57.7% female, and the mean age was 15.5 years (SD = 1.5 years).

Measures

Brief Family Relationship Scale

The BFRS is adapted from the 27-item Relationship dimension of the FES (Moos & Moos, 1994), consisting of Cohesion, Expressiveness, and Conflict subscales (9 items each). These subscales measure support, expression of opinions, and angry conflict within a family.

Communal Mastery Family Scale (CMFS; α = .76)

The CMFS is a 4-item measure adapted for Alaska Native youth from the 10-item Communal Mastery Scale (Jackson, McKenzie, & Hobfoll, 2000), measuring resources provided by the family for coping with stress.

Reasons for Life Scale (RFLS; α = .83)

The RFLS is a new 13-item scale designed for Alaska Native adolescents, adapted from Osman et al. (1996) Brief Reasons for Living–Adolescent scale. The RFLS explores beliefs and experiences that contribute to sense of meaning in life and is hypothesized inversely related to suicidal ideation.

Youth Community Protective Factors Scale (YCPFS; α = .77)

The YCPFS is a 6-item scale adapted from the Yup’ik Protective Factors Scale for adults (Allen et al., 2006). Support and Opportunities subscales tap youth perceptions regarding the extent of these protective factors available to young people in their community.

Alaska Native Cultural Identification Scale (ANCIS)

The ANCIS is an 8-item scale adapted from the Orthogonal Cultural Identification Scale (Oetting & Beauvais, 1990-1991) that measures home and dominant cultural identification in Alaska Native youth. The ANCIS has two 4-item subscales, Alaska Native Cultural Identification (ANCI; α = .77) and White American Cultural Identification (WACI; α = .63).

Procedures

After rewriting items with the assistance of focus groups, we pilot tested and removed poorly functioning items, leaving an 8-item Cohesion subscale, 4-item Expressiveness subscale, and 7-item Conflict subscale, totaling 19 items (see the appendix).

Participants were recruited through active parental consent and youth assent procedures approved by the University of Alaska Fairbanks (UAF) Institutional Review Board. They were paid $15 for completing measures, which were administered in school computer labs via a secure web server based at UAF. Responses on all measures were via a continuous analog scale with a pointer in the shape of a salmon with three semantic anchors placed below the scale (Not at all, Somewhat, and A lot).

Results and Discussion

Prior to analysis, we converted the continuous “slider” scale into 20 intervals and then grouped the intervals in a manner that approximated a normal distribution: 1-7 = 1, 8-11 = 2, 12-15 = 3, 16-18 = 4, 19-20 = 5. Principal components analysis suggested each subscale was unidimensional.

Objective 1: Assessing the Internal Structure of the BFRS

Comparing unidimensional, orthogonal, and second-order three-factor structures (see Figure 1) using essentially tau-equivalent confirmatory factor analysis (CFA) models, we found that the second-order three-factor model fit better than either the unidimensional model, Δχ²(5) = 150.6 and ΔBIC = 122.32, or the first-order three orthogonal factor model, Δχ²(3) = 335.8 and ΔBIC = 318.78 (Graham, 2006; Lee, Dunbar, & Frisbie, 2001). Freeing the item loadings to locate poorly fitting items and items with correlated unique variances led us to eliminate one item from the final scale. The resulting second-order three-factor model was an acceptable fit to the data, χ²(132) = 261.3, χ²/df = 1.98, goodness-of-fit index (GFI) = .91, comparative fit index (CFI) = .93, and root mean square error of approximation (RMSEA) = .06 (see Table 1).

Figure 1.

Confirmatory factor analyses comparison between the unidimensional, first-order three orthogonal factor and second-order three-factor models.

Table 1.

Model Fit Indexes for the Unidimensional, First-Order Three Orthogonal Factor, Second-Order Three-Factor, and Other Models.

Model	Number of items	χ² (df)	χ²/df	BIC	GFI	CFI	RMSEA
Unidimensional	19	605.3 (170)	3.56	718.28	.78	.79	.10
First-order three orthogonal factor	19	790.5 (168)	4.71	914.74	.78	.70	.11
Second-order three-factor	19	454.7 (165)	2.76	595.96	.85	.86	.08
Second-order three-factor three-category calibration	16	164.9 (101)	1.63	362.60	.93	.95	.05

Note. CFA = confirmatory factor analysis; BIC = Bayesian information criterion; GFI = goodness-of-fit index; CFI = comparative fit index; RMSEA = root mean square error of approximation.

Objective 2: Evaluate Item Functioning With Alternative Response Scales

We used Samejima’s (1996) graded response item response theory (IRT) model, through the ltm package in R (Rizopoulos, 2006), to examine the relation between responses on the items to the underlying latent construct measured by each subscale. We follow the procedures of Marshall, Orlando, and Jaycox (2002) in reporting the IRT results.

Results of the item parameter estimates showed that the 18 items overall provided satisfactory discrimination. Item 9 provided the highest discrimination index, and Item 17 was the least discriminating. Examining the option probability curves led us to recode the data to three options by collapsing Options 2 to 4. The amount of overlap between Categories 2 and 4 suggested that these three categories might be providing redundant information. Table 2 reports the IRT results for this three-category calibration, including the discrimination or slope parameter (a) and the location parameters (b_s) for each item.

Table 2.

Item Parameters for Three-Category Calibration for the 18-Item BFRS.

Item	a	b ₁	b ₂
Cohesion
1	2.20	−2.31	0.34
3	1.56	−1.09	1.84
6	1.74	−2.14	1.16
7	2.90	−1.41	0.85
10	1.27	−1.48	2.12
12	2.06	−2.20	0.43
14	2.35	−2.43	−0.83
16	2.11	−1.77	0.75
Expressiveness
4	1.67	−1.23	0.92
8	1.88	−1.05	1.28
17	0.70	−2.45	2.29
18	1.71	−1.12	1.38
Conflict
2	1.57	−1.66	1.57
5	1.88	−2.42	0.39
9	3.91	−1.31	0.69
11	2.11	−1.93	0.25
13	1.56	−2.02	−0.16
19	1.77	−0.57	1.82

Note. BFRS = Brief Family Relationship Scale.

Location parameter values in Table 2 show that for Cohesion Items 1, 12, and 14 and Conflict Resolution Items 5, 11, and 13, endorsing a higher response option occurred at relatively low levels of the latent trait attribute tapped by the subscale, indicating these items provided more information for individuals at low levels of these attributes. To investigate the item information further, we examined the item information functions for each subscale for the three-category calibration. Item 9 provided the greatest amount of information, especially at slightly below and slightly above-average levels, and Items 10 and 17 provided the least information. Items 1, 7, 11, and 16 supplied moderate to high information at below or slightly above-average levels. The items that provided low to moderate information across most ranges were Items 2, 4, 5, 6, 8, 10, 12, 13, and 18.

Final CFA results

We then reran the CFA analysis, adjusted to correspond with these IRT findings, testing the 18-item second-order, three-factor model using three-category calibration of the rating scale against the original five-category calibration. The difference in BICs was 31.9, a substantial improvement in fit for the three anchor point calibration (Raftery, 1993). Items 10 and 17 provided little unique information according to the IRT information curves and were removed from the final scale. The improvement in fit after removing Items 10 and 17 was significant, Δχ²(31) = 64.6, p < .01, ΔBIC = 87.2. The final 16-item second-order three-factor CFA using the three-category calibration was a nearly acceptable fit, χ²(101) = 164.9, χ²/df = 1.63, GFI = .93, CFI = .95, and RMSEA = .05. Internal consistency was acceptable for Cohesion (M = 15.73, SD = 2.70, α = .83) and Conflict (M = 13.06, SD = 2.51, α = .80) and for the full-scale BFRS (M = 34.76, SD = 5.53, α = .88) but weaker for Expressiveness (M = 5.97, SD = 1.47, α = .65).

These internal consistency values were slightly higher than those reported in the manual on the Cohesion and Conflict subscales and slightly lower on the Expressiveness subscale. The comparatively low internal consistency of the Expressiveness subscale found in this study may stem from poor fit of the construct of expressiveness in this non-Western cultural group. We explored the BFRS with a non-Western North American indigenous cultural group in part because we hypothesized the group was representative of an orientation with alignments toward collectivism (Oyserman, Coon, & Kemmelmeier, 2002), possessing somewhat different value orientations to family relationships. The Expressiveness subscale may not perform as well with this and other non-Western groups, in comparison with mainstream American cultural groups more aligned with individualistic orientations to family functioning. This echoes Ma and Leung’s (1990) similar findings of low reliabilities with the Expressiveness subscale in Hong Kong and stated concerns that some items may not have equivalent meaning. Future research will be needed to determine whether the Expressiveness subscale can be adapted for use with non-Western cultural groups.

Objective 3: Convergent and Discriminant Evidence for Validity of BFRS Score Interpretations

The BRFS scores correlated in the expected direction with the CMFS scores (r = .51, p < .01), the RFLS scores (r = .48, p < .01), and the YCPFS scores (r = .44, p < .01; see Table 3). As expected, the weakest convergent validity correlation, between the BFRS and the YCPFS scores (r = .44), was significantly greater in magnitude than the strongest discriminant validity correlation, between BFRS and ANCI scores (r = .18; t = −4.87, p < .01).

Table 3.

Correlation Among the BFRS Scale, Subscales, and Validity Measures.

	1	2	3	4	5	6	7	8
1. BFRS
2. Cohesion	.86^**
3. Expressiveness	.82^**	.60^**
4. Conflict^a	.76^**	.55^**	.36^**
5. CMFS	.51^**	.56^**	.39^**	.31^**
6. RFLS	.48^**	.53^**	.34^**	.31^**	.55^**
7. YCPFS	.44^**	.44^**	.37^**	.26^**	.39^**	.45^**
8. ANCI	.18^**	.26^**	.03	.19^**	.17^**	.30^**	.21^**
9. WACI	−.10	−.13^*	−.00	−.14^*	−.05	−.09	−.02	−.43^**

Note. BFRS = Brief Family Relationship Scale; CMFS = Communal Mastery Family Scale; RFLS = Reasons for Life Scale; YCPFS = Youth Community Protective Factors Scale; ANCI = Alaska Native Cultural Identification; WACI = White American Cultural Identification.

The Conflict subscale score is reverse-keyed, such that a higher Conflict score represents lower conflict in the family, or more positive family relationships.

p < .05. **p < .01.

In summary, the BFRS provides an assessment of youth perceptions of family functioning by measuring three aspects of family relationships. In addition to Alaska Native populations, the BFRS may be suitable for use with other American Indian groups and with other non-Western and collectivist cultural groups, such as individuals from East and South Asian cultures and individuals from Western cultural backgrounds. Further research is needed to test BFRS operating characteristics and to validate the interpretation of scores within these distinct cultural groups.

Footnotes

Appendix

Brief Family Relationship Scale

Cohesion

1. In our family we really help and support each other.

3. In our family we spend a lot of time doing things together at home.

6. In our family we work hard at what we do in our home.

7. In our family there is a feeling of togetherness.

12. My family members really support each other.

14. I am proud to be a part of our family.

16. In our family we really get along well with each other.

Expressiveness

4. In our family we can talk openly in our home.

8. In our family we sometimes tell each other about our personal problems.

18. In our family we begin discussions easily.

Conflict

2. In our family we argue a lot. (R)

5. In our family we are really mad at each other a lot. (R)

9. In our family we lose our tempers a lot. (R)

11. In our family we often put down each other. (R)

13. My family members sometimes are violent. (R)

19. In our family we raise our voice when we are mad. (R)

Note. Items with (R) are reverse-keyed. Only the final 16 items are shown here. Three items, 10 (In our family we do things for each other without being asked), 15 (In our family we work out our problems), and 17 (In our family we are usually careful about what we say to each other), were removed in the process of our analyses.

Acknowledgements

We acknowledge the invaluable contributions of James A. Walsh. We also want to thank the People Awakening Team including participants, community coresearchers, our Coordinating Council, and our project staff for their assistance in completing this research.

Authors’ Note

The People Awakening (PA) Team includes the Yupiucimta Asvairtuumallerkaa Councils, the Ellangneq Councils, the Yup’ik Regional Coordinating Council, the Ellangneq Advisory Group, and the Ellangneq, Yupiucimta Asvairtuumallerkaa, and Cuqyun Project Staff. The Yupiucimta Asvairtuumallerkaa Councils included Sophie Agimuk, Harry Asuluk, Thomas Asuluk, T. J. Bentley, John Carl, Mary Carl, Emily Chagluk, James Charlie Sr., Lizzie Chimiugak, Ruth Jimmie, Jolene John, Paul John, Simeon John, Aaron Moses, Phillip Moses, Harry Tulik, and Cecelia White. The Ellangneq Councils includes Catherine Agayar, Fred Augustine, Mary Augustine, Paula Ayunerak, Theresa Damian, Lawrence Edmund Sr., Barbara Joe, Lucy Joseph, Joe Joseph, Placide Joseph, Zacheus Paul, Charlotte Phillp, Henry Phillip, Joe Phillip, Penny Alstrom, Fred Augustine, Mary Augustine, Paula Ayunerak, Theresa Damian, Shelby Edmund, Flora Patrick, Dennis Sheldon, Isidore Shelton, Catherine Agayar, Theresa Damian, Freddie Edmund, Shelby Edmund, Josie Edmund, and Flora Patrick. The Yup’ik Regional Coordinating Council includes Martha Simon, Moses Tulim, Ed Adams, Tammy Aguchak, Paula Ayunerak, Sebastian Cowboy, Lawrence Edmunds, Margaret Harpak, Charles Moses, and Raymond Oney. The Ellangneq Advisory Group includes Walkie Charles, Richard Katz, Mary Sexton, Lisa Rey Thomas, Beti Thompson, and Edison Trickett. The Ellangneq Project Staff includes Debbie Alstrom, Carl Blackhurst, Rebekah Burkett, Diana Campbell, Arthur Chikigak, Gunnar Ebbesson, Aaron Fortner, John Gonzalez, Scarlett Hopkins, Nick Hubalik, Joseph Klejka, Charles Moses, Dora Nicholai, Eliza Orr, Marvin Paul, Michelle Dondanville, Jonghan Kim, Rebecca Koskela, Johanna Herron, and Stacy Rasmus.

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 research was funded by the National Institute of Alcohol Abuse and Alcoholism and the National Center for Minority Health Disparities (R21AA0016098, National Institute on Alcohol Abuse and Alcoholism, PI: James Allen; R24MD001626, National Center on Minority Health and Health Disparities, PI: James Allen; R21AA015541, National Institute on Alcohol Abuse and Alcoholism, PI: Gerald V. Mohatt; R01AA11446, National Institute on Alcohol Abuse and Alcoholism & National Center on Minority Health and Health Disparities, PI: Gerald V. Mohatt; P20RR061430, National Center for Research Resources, PI: Gerald V. Mohatt) and a University of Alaska International Polar Year Postdoctoral Fellowship award to the first author.

References

Allen

Mohatt

G. V.

Hazel

Rasmus

Thomas

Lindley

(2006). The tools to understand: Community as co-researcher on culture specific protective factors for Alaska natives. Journal of Prevention & Intervention in the Community, 32, 41-64. doi:10.1300/J005v32n01_04

Graham

J. M.

(2006). Congeneric and (essentially) tau-equivalent estimates of score reliability: What they are and how to use them. Educational and Psychological Measurement, 66,930-944.

Jackson

McKenzie

Hobfoll

S. E.

(2000). Communal aspects of self-regulation. In Boekaerts

Pintrich

P. R.

Zeider

(Eds.), Handbook of self-regulation (pp. 275-300). San Diego, CA: Academic Press.

Lee

Dunbar

S. B.

Frisbie

D. A.

(2001). The relative appropriateness of eight measurement models for analyzing scores from tests composed of testlets. Educational and Psychological Measurement, 61, 958-975. doi:10.1177/00131640121971590

H. K.

Leung

M. C.

(1990). The adaptation of the Family Environment Scale to Chinese children and adolescents in Hong Kong. International Journal of Psychology, 25, 545-555. doi:10.1080/00207599008247882

Marshall

G. N.

Orlando

Jaycox

L. H.

(2002). Development and validation of a modified version of the peritraumatic dissociative experiences questionnaire. Psychological Assessment, 14, 123-134. doi:10.1037/1040-3590.14.2.123

Moos

R. H.

Moos

B. S.

(1994). Family Environment Scale manual (3rd ed.). Palo Alto, CA: Consulting Psychologists.

Oetting

E. R.

Beauvais

(1990-1991). Orthogonal cultural identification theory: The cultural identification of minority adolescents. International Journal of the Addictions, 25,655-685.

Osman

Kopper

B. A.

Barrios

F. X.

Osman

J. R.

Besett

Linehan

M. M.

(1996). The Brief Reasons for Living Inventory for Adolescents (BRFL-A). Journal of Abnormal Child Psychology, 24, 433-443. doi:10.1007/BF01441566

10.

Oyserman

Coon

H. M.

Kemmelmeier

(2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological Bulletin, 128, 3-72. doi:10.1037/0033-2909.128.1.3

11.

Raftery

A. E.

(1993). Bayesian model selection in structural equation models. In Bollen

K. A.

Long

J. S.

(Eds.), Testing structural equation models (pp. 63-180). Newbury Park, CA: SAGE.

12.

Rizopoulos

(2006). ltm: An R package for latent variable modeling and item response theory analyses. Journal of Statistical Software, 17(5), 1-25.

13.

Roosa

M. W.

Beals

(1990). Measurement issues in family assessment: The case of the Family Environment Scale. Family Processes, 29, 191-198. doi:10.1111/j.15455300.1990.00191.x

14.

Samejima

(1996). Graded response model. In van der Linden

W. J.

Hambleton

R. K.

(Eds.), Handbook of modern item response theory (pp. 85-100). New York, NY: Springer.

15.

Sanford

Raymond

Zucker

R. A.

(1999). Validity issues with the Family Environment Scale: Psychometric resolution and research application with alcoholic families. Psychological Assessment, 11, 315-325. doi:10.1037/1040-3590.11.3.315