Further Validation of the Couple Relationship Scale

ABSTRACT

This study aimed to build on previous research by S. R. Anderson et al. in 2022 and further validate the Couple Relationship Scale (CRS), a routine outcome measure for use in couple therapy. Using data from a community sample (N = 300, Mean age = 41.61, 49% females), confirmatory factor analysis supported a robust single‐factor structure with no evidence of gender‐based measurement invariance. The CRS demonstrated excellent internal consistency reliability. Strong concurrent validity was evidenced by strong correlations with established measures of couple functioning and moderate correlations with mental health and well‐being scales. CRS items showed expected associations with corresponding subscales of other measures of relational functioning and well‐being scales. Scores were largely independent of demographic variables. CRS shows promise as a psychometrically sound tool to bridge the research‐practice gap in couple therapy, although further validation in clinical populations is warranted.

1. Introduction

Routine outcome measures (ROMs) are validated, structured tools used to measure client progress and provide session‐by‐session feedback on the impact of psychosocial interventions, including couple and family therapy (Anker et al. 2009; Barkham et al. 2023; Börjesson and Boström 2020). ROMs provide practice‐based evidence for the efficacy of interventions and give therapists and clients with regular feedback on progress. Not only are individuals who receive feedback more likely to experience clinically noteworthy change, implementing formal client feedback into clinical practice is an ethical imperative (Lappan et al. 2018; Reese et al. 2010). Moreover, data gathered through ROMs are very useful in monitoring a therapist's facilitative interpersonal skills development (T. Anderson et al. 2016).

Research conducted by Halford et al. (2016) revealed that while meta‐analyses of randomized controlled trials of couple therapy find large improvements in couple adjustment, real‐world effectiveness studies report only weak to moderate effects. The use of clinically robust ROMs such as the Couple Relationship Scale (CRS, S. R. Anderson et al. 2022) may serve to close the research‐efficacy to practice‐effectiveness gap.

The validation of ROMs for assessing couple relationship distress should show that they correlate with demographic and psychosocial factors that are associated with the quality and stability of couple relationships (Amato 2014; Bradbury et al. 2000; Karney 2020; Whisman and Robustelli 2016). For example, couples with high socioeconomic status (SES) postpone marriage and childbirth to invest in higher education and career development, but have higher rates of marriage and lower divorce rates compared with their low SES counterparts (Karney 2020). In the psychosocial domain, there is a bidirectional relationship between mental health and relationship distress, with psychopathology being a risk factor for greater relationship distress, and couples with relationship problems, having poorer mental health (Whisman and Robustelli 2016). In the current study, correlations were computed between CRS and a range of demographic variables and indices of mental health.

1.1. The CRS

In recent years, there has been a move toward enhancing the clinical utility of ROMs for evaluating relationship functioning. While psychometric measures are widely recognized as valuable tools in couple therapy, many existing instruments fail to adequately integrate clinical utility with strong psychometric properties, leaving a gap for the development of brief, valid, and reliable ROMs for couple functioning (Barkham et al. 2023). CRS was developed to address this gap and contribute to the slowly growing body of brief, clinically useful measures of couple functioning (S. R. Anderson et al. 2022). CRS is a 10‐item measure of relational functioning that assesses emotional intimacy, commitment, trust, safety, cohesion, acceptance, conflict, physical intimacy, overall happiness, and personal well‐being. These components were identified by researchers, clinicians and couples as key facets of the relationship to monitor during couple therapy. Participants rate each dimension using the 100 mm visual analog scales (VAS) anchored by descriptors representing the upper and lower levels of each dimension. CRS total scores range from 0 to 100, with higher scores indicating better relationship functioning. A total score is calculated by averaging all item responses from 0 to 100. It is designed to routinely monitor relationship functioning over time. A clinical cut‐off of 70.9 identifies the point at which participants are more likely to experience clinical levels of relationship distress, with lower scores indicating more distress. The measure can be completed in 1–2 min and requires a fifth‐grade reading level.

1.2. The Current Study

Building on the work of S. R. Anderson et al. (2022), the aim of the current study was to provide an independent and extended examination of the validity of the CRS. This was achieved by addressing the following six research questions:

• 1.In the current sample does CRS have a single‐factor structure?
• 2.Is there evidence of any gender‐based measurement invariance?
• 3.In the current sample does CRS have high internal consistency reliability?
• 4.
  Does CRS have construct/concurrent validity shown by the following pattern of correlations:

  • Strong significant correlations with other brief psychometric measures of couple relational functioning (Clinical Outcome Routine Evaluation for Couples (C‐SCORE; Carr et al. 2025); Revised Dyadic Adjustment Scale (RDAS; Busby et al. 1995);
  • Moderate significant correlations with brief psychometric measures of mental health (Mental Health Inventory‐5 (MHI‐5; Berwick et al. 1991) and well‐being (WHO‐5; World Health Organization, Regional Office for Europe 1998);
  • Moderate significant correlations with self‐ratings on 10‐point scales of the severity of respondents' main relationship problem or challenge; their management of this problem or challenge; how safe they feel in their relationship; and how helpful they think couple therapy would be?

• 5.
  Do the 10 CRS items have concurrent validity shown by strong significant correlations with relevant factor scales of other brief psychometric measures of couple relational functioning, and well‐being and mental health scales? Specifically, are there strong significant correlations between the following:

  • CRS emotional intimacy item and C‐SCORE sex and intimacy subscale;
  • CRS commitment item and both C‐SCORE commitment, trust and respect, and RDAS satisfaction subscales;
  • CRS trust item and C‐SCORE commitment, trust, and respect subscale;
  • CRS safety item and C‐SCORE safety subscale;
  • CRS cohesion item and RDAS cohesion subscale;
  • CRS acceptance item and C‐SCORE communication and problem‐solving subscale;
  • CRS conflict item and both C‐SCORE consensus on care, housework, and money and RDAS consensus subscales;
  • CRS physical intimacy item and C‐SCORE sex and intimacy subscale.
  • CRS overall happiness item and RDAS satisfaction subscale;
  • CRS personal well‐being item and both WHO‐5 and MHI‐5.

• 6.Does CRS have significant associations with the following demographic variables: age, partner's age, age difference between partners, differing partner's ethnicity, relationship duration, educational level, SES as indexed by main occupation; gender, ethnicity, partner's ethnicity, country of residence, sexual orientation, relationship status, and couple therapy status?

2. Method

2.1. Ethics and Pre‐Registration

The study was conducted with ethical approval from University College Dublin Humanities Research Ethics Committee and informed consent of participants. It involved secondary analysis of data from Carr et al. (2025). The previous study that produced the data set re‐analyzed in this study (https://doi.org/10.17605/OSF.IO/FKD48), and the current study (https://osf.io/3h9ab/?view_only=6f7e68fb154046b6b60f72ae3a3846e1) were preregistered on Open Science Framework before data analysis.

2.2. Design

This study employed a cross‐sectional correlational design. The following scales were used: the CRS, C‐SCORE, RDAS, MHI‐5, and WHO‐5. These are described below, along with self‐reporting and demographic items.

2.3. Data Set

The data set used in this study is described in Carr et al. (2025). It was collected from a group of 300 community‐based participants who were (a) over 18 years; (b) fluent and literate in English; (c) in a cohabiting or marital heterosexual or nonheterosexual couple relationship for at least 1 year; and (d) their partner was not a study participant. Partners of study participants were excluded because the data analytic methods used assumed data points were independent. Within the constraints of the demographic profiles of potential participants available through the Prolific (https://www.prolific.com/) platform a convenience sample was recruited stratified by age (under 40, [50%] or over 40 years [50%]), gender (female [49%], male [47%], Nonbinary [4%]), and country (United Kingdom [36%], United States [34%], Australia [15%], Europe [15%]). Participants were paid £4.50 for participating and completed online questionnaires on the Pavlovia (https://pavlovia.org/) platform.

2.4. Sample Size

A review of sample size requirements by Kyriazos (2018) found that at least 5–10 cases per item are required to conduct reliability and factor analysis of multi‐item scales. As the CRS contains 10 items, 50–100 cases were required for the study. The available data set contained 300 cases and was therefore adequately powered for planned analyses.

2.5. Sample

Table 1 shows that the community‐based sample who provided data for this study included 300 participants (M age = 41.6 years), most of whom were married (85%), heterosexual (85%), and White/Caucasian (81%). The majority resided in the United Kingdom, United States, or Australia, held at least a bachelor's degree, and worked in professional or managerial roles. Most were not currently in couple therapy (92%), though a small proportion had previously attended (7%).

Table 1.

Demographic characteristics.

Variable	Categories	f or M	% or SD
Age a		41.61	12.39
Partner's agea		40.82	13.753
Age difference between partners a		4.37	4.33
Genderb	Female	147	49.0
	Male	141	47.0
	Trans	2	0.7
	Nonbinary	9	3.0
	Other	1	0.3
Ethnicity b	White/Caucasian	242	80.7
	Black	20	6.7
	Asian	18	6.0
	Hispanic	7	2.3
	Indigenous minority e.g. Aboriginal, Roma	1	0.3
	Mixed‐race	10	3.3
	Missing	2	0.7
Partner's ethnicity b	White/Caucasian	239	79.7
	Black	20	6.7
	Asian	21	7.0
	Hispanic	9	3.0
	Indigenous minority e.g. Aboriginal, Roma	1	0.3
	Mixed‐race	8	2.7
	Missing	2	0.7
Differing partner's ethnicity b	Same	238	79.3
	Different	61	20.3
Country of residence b	UK	109	36.3
	USA	102	34.0
	Australia	44	14.7
	Belgium	6	2.0
	Germany	12	4.0
	Greece,	11	3.7
	Ireland	5	1.7
	Netherlands	11	3.7
Sexual orientation b	Heterosexual	255	85.0
	Lesbian	5	1.7
	Gay	6	2.0
	Bisexual	23	7.7
	Pansexual	9	3.0
	Other or prefer not to say	2	0.7
Relationship status b	Single	13	4.3
	Married or civil partnered	255	85.0
	Other or prefer not to say	32	10.7
Relationship duration a		15.33	11.55
Educational level b	Primary/grade school	2	.7
	High school	46	15.3
	Certificate or diploma	60	20.0
	Bachelor's degree	131	43.7
	Masters' degree	50	16.7
	Doctoral degree	11	3.7
Socioeconomic status b	Higher professional or managerial	67	22.3
	Lower professional or managerial	113	37.7
	Nonmanual (or student)	27	9.0
	Skilled manual	23	7.7
	Semi‐skilled manual	6	2.0
	Unskilled manual	6	2.0
	Unemployed	58	19.3
Couple therapy status b	Not in couple therapy	275	91.7
	In couple therapy now	5	1.7
	Previously in couple therapy	20	6.7

Abbreviations: f, frequency; M, mean; SD, standard deviation.

Note: This table is reproduced with permission from Carr et al. (2025). N = 300. The data set (Carr et al. 2025) comprised 300 community‐based adults who were at least 18, fluent in English, and in a cohabiting or marital relationship of at least one year. Partners were excluded to maintain data independence.

^a Values are means and standard deviations.

^b Values are frequencies and percentages.

2.6. Instruments

2.6.1. CRS

CRS is a 10‐item measure of relational functioning that assesses emotional intimacy, commitment, trust, safety, cohesion, acceptance, conflict, physical intimacy, overall happiness, and personal well‐being (S. R. Anderson et al. 2022). For all 10 items, participants were asked to evaluate their relationship with their partner over the past week by responding on a visual analogue scale, for example, “Distant‐Close” to assess intimacy, and “Like giving up‐Completely committed” to assess commitment. In the initial validation study, CRS showed strong concurrent and construct validity, and internal consistency (α = 0.93), and test‐retest (r = 0.75) reliability. In the current study, the internal consistency reliability of CRS was α = 0.96.

2.6.2. Systemic Clinical Outcome Routine Evaluation for Couples (C‐SCORE)

C‐SCORE is a recently developed ROM for use in couple therapy that assesses significant domains of couple relational functioning (Carr et al. 2025). It is the couple version of the widely used family SCORE assessment scale (Carr and Stratton 2017). There are 22 and 26‐item versions of C‐SCORE, and in the current study, the factor‐analytically based 22‐item version was used. C‐SCORE yields an overall couple relational functioning index, and scores on five factor scales that assess (1) commitment, respect, and trust, (2) communication and problem‐solving, (3) consensus on care, housework, and money, (4) safety, and (5) sex and intimacy. Respondents indicate how well each item describes their current view of their relationship on 5‐point Likert‐type scales ranging from 0 = strongly disagree to 4 = strongly agree. High scores indicate better functioning. The following are example C‐SCORE items: “We are both committed to this relationship” and “I am listened to by my partner.” Carr et al. (2025) found that C‐SCORE had good internal consistency (α = 0.95) and test–retest (r = 0.95) reliability, good construct validity, and negligible correlations with demographic variables. In the current study, the C‐SCORE internal consistency reliability was α = 0.95.

2.6.3. Revised Dyadic Adjustment Scale (RDAS)

RDAS is a 14‐item short version (Busby et al. 1995) of the 32‐item Dyadic Adjustment Scale (Spanier 1976). It yields an overall relational functioning score and scores for relationship satisfaction, consensus, and cohesion subscales. For 13 items, responses are given on 6‐point Likert‐type scales, and for the remaining item, a 5‐point Likert‐type scale is used. High scores indicate better relational functioning. The following are example RDAS items: “Do you ever regret that you married (or lived together?)” and “How often do you and your mate get on each other's nerves?” RDAS has been very widely used in couple therapy research. It has excellent psychometric properties including a replicated 3‐factor structure, reliability coefficients ranging from 0.85 to 0.90, construct validity, and sensitivity to change (S. R. Anderson et al. 2014; MacIntosh and Butters 2014). In the current study, the internal consistency reliability of RDAS was α = 0.86.

2.6.4. Mental Health Inventory‐5 (MHI‐5)

MHI‐5 is a 5‐item scale widely used to screen for common mental health problems (Berwick et al. 1991). It was developed from a 38‐item version (Ware et al. 1979). MHI‐5 contains three items that measure depressive symptoms and two that assess anxiety symptoms. Respondents indicate on 6‐point Likert‐type scales how much of the time in the past month they have felt in the ways described in each of the item statements, ranging from 0 = all of the time, to 5 = none of the time. The following are example MHI‐5 items: “Have you been a very nervous person?” and “Have you felt down hearted and blue?”. Positively worded items are reverse scored before summing items to give the scale total, which is multiplied by 4 to give an MHI‐5 score range from 0 to 100 with high scores indicating greater well‐being. MHI‐5 has good internal consistency reliability coefficients ranging from 0.85 to 0.90, and construct and criterion validity (Shields et al. 2021). In the current study, the MHI‐5 internal consistency reliability was α = 0.89.

2.6.5. World Health Organization 5‐item Well‐being Scale (WHO‐5)

WHO‐5 is a 5‐item scale which is among the most widely used for assessing well‐being (World Health Organization, Regional Office for Europe 1998). Respondents indicate on 6‐point scales how much of the time in the preceding 2 weeks they have felt in the ways described in each of the item statements ranging from 0 = none of the time, to 5 = all of the time. The following are example WHO‐5 items: “I have felt cheerful and in good spirits” and “I have felt active and vigorous”. The sum of five item scores is multiplied by four to give a WHO‐5 total ranging from 0 to 100 with high scores indicating greater well‐being. The WHO‐5 has been translated into over 30 languages. It has good construct and criterion validity, and also has good internal consistency reliability with studies yielding reliability coefficients ranging from 0.82 to 0.89 (Topp et al. 2015). In the current study, the WHO‐5 internal consistency reliability was α = 0.91.

2.6.6. Self‐Ratings and Demographic Items

In this study self‐ratings of the following items were used:

• Severity of respondents' main relationship problem or challenge;

• Their management of this problem or challenge;

• How safe they feel in their relationship; and

• How helpful they think couple therapy would be.

The following demographic items were used: age, partner's age, gender, ethnicity, partner's ethnicity, country of residence, sexual orientation, relationship status, relationship duration, educational level, main occupation, and currently in couple therapy. Table 1 presents demographic characteristics of the sample, and shows the response options for each of the demographic items.

2.7. Data Analytic Plan

To address the first research question about the factor structure of the CRS, a confirmatory factor analysis (CFA) was conducted to determine if the data fit a single factor structure. The adequacy of the model fit was assessed using the following indices: comparative fit index (CFI) and Tucker–Lewis index (TLI) with values > 0.90 and > 0.95 indicating good and excellent fits respectively; root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR) with values < 0.08 and < 0.09 indicating good and excellent model fits respectively (Hu and Bentler 1999; Kline 2016).

To address the second research question of measurement invariance, a multiple indicators multiple causes (MIMIC) model was conducted to examine sex differences in item performance, that is, to test for differential item functioning (DIF) as proposed by Finch (2005). The MIMIC model extends the CFA model by adding the gender grouping variable (Males = 0, Females = 1) and regressing the latent variable on this binary variable. A significant regression coefficient indicates significant mean differences at the level of the latent variable between males and females. The presence of DIF was examined by inspecting the modification indices (MIs) for the direct paths from the gender variable to the CRS items. MIs indicate which path could be added to the model that would significantly improve model fit if freely estimated, that is, reduce the χ ² by 3.84 or more (which is the critical value for the χ ² for 1 degree of freedom). A process was followed whereby the path with the largest MI was freely estimated in the model, and the model was re‐estimated. This continued until there were no MIs greater than 5. Any significant paths would be interpreted, with the regression coefficient indicating the degree to which females would score significantly higher or lower on that item compared with males, given the same level of the latent variable. All analyses were conducted in Mplus 8.11 (Muthén and Muthén 2017).

To address the third research question about the internal consistency reliability of the CRS, Cronbach (1951) alpha and McDonald (1999) omega coefficients were computed.

To address the fourth question about construct and concurrent validity, Pearson product–moment correlations were calculated between CRS and RDAS, C‐SCORE, MHI‐5, and WHO‐5 totals; as well as ratings on 10‐point scales of relationship challenge severity, management of this challenge, relationship safety, and possible helpfulness of couple therapy.

To address the fifth research question about the concurrent validity of CRS items, correlations were computed between the 10 CRS items and corresponding factor scales of other brief psychometric measures of couple relational functioning, and well‐being and mental health scales.

To address the sixth research question about the statistical significance of the relationship between CRS and demographic variables, correlations were computed between CRS and the following continuous variables: age, partner's age, age difference between partners, differing partner's ethnicity, relationship duration, educational level and SES as indexed by main occupation.

Analysis of variance (ANOVA), with Tukey–Kramer post‐hoc comparisons for groups of unequal sizes were used to assess the significance of differences in mean CRS scores for subgroups on the following categorical demographic variables: gender, ethnicity, partner's ethnicity, country of residence, sexual orientation, relationship status, and couple therapy status.

3. Results

3.1. Factor Structure

In answer to the first research question about the factor structure of the CRS, a CFA of all 10 CRS items was conducted. This showed that a single factor structure fit the data when two correlated errors were included in the model shown in Figure 1 (χ ² = 88.34, df = 33, p < 0.001; CFI = 0.97; TLI = 0.95; RMSEA = 0.075; SRMR 0.026). All item loadings were quite strong ranging from β = 0.67 to β = 0.96. This indicates that the single latent factor is well captured by all items, thus providing further support for the appropriateness of the one‐factor solution.

Figure 1.

Confirmatory factor analysis of the Couple Relationship Scale. [Color figure can be viewed at wileyonlinelibrary.com]

3.2. Measurement Invariance

In answer to the second research question, the binary gender variable did not predict the CRS Relationship Functioning latent variable (β = −0.07, p = 0.226), and this indicated that there was no mean difference between males and females on the latent variable. The MIs for the direct effect from the gender variable to each of the CRS items ranged from 0.150 to 3.308, with none reaching the critical value of 3.84. This indicated that there was no evidence for gender‐based measurement invariance.

3.3. Internal Consistency Reliability

In answer to the third research question, Cronbach's α and McDonald's omega reliability analyses showed the CRS had excellent internal consistency reliability with alpha and omega values of 0.96. Table 2 shows the reliability coefficients for other scales and subscales used in the study ranged from 0.75 (acceptable) to 0.95 (excellent).

Table 2.

Correlations between CRS and measures of relational functioning, mental health and wellbeing, self‐rated couple relationship issues, and demographic variables; descriptive statistics for all variables; and reliability coefficients for all scales.

			Internal consistency reliability
Variables	M	SD	α	Ω	Correlation with CRS
Couple Relationship Scale	79.83	19.26	0.96	0.96	—
Relational functioning scales
C‐SCORE Total	81.51	16.15	0.95	0.95	0.80***
C‐SCORE Commitment, trust and respect	24.24	4.63	0.95	0.93	0.78***
C‐SCORE Communication and problem‐solving	15.98	3.49	0.88	0.87	0.76***
C‐SCORE Consensus on care, housework, and money	8.77	2.36	0.75	0.76	0.58***
C‐SCORE Safety	15.80	3.81	0.79	0.80	0.60***
C‐SCORE Sex and intimacy	5.19	2.19	0.88	—	0.61***
RDAS Total	49.28	8.50	0.86	0.86	0.70***
RDAS Consensus	22.72	4.49	0.79	0.80	0.58***
RDAS Satisfaction	15.33	2.72	0.82	0.82	0.70***
RDAS Cohesion	11.23	3.33	0.77	0.78	0.51***
Mental health and well‐being scales
MHI‐5‐5	68.60	20.59	0.89	0.88	0.51***
WHO‐5	57.88	21.72	0.91	0.91	0.42***
Self‐rated couple relationship issues
Severity of main relationship problem/challenge	5.72	2.57	—	—	0.46***
How are you managing the problem/challenge together with your partner?	6.69	2.28	—	—	0.67***
How safe do you feel in this relationship?	8.82	1.78	—	—	0.65***
Do you think couple or family therapy could be helpful to your relationship?	5.01	2.63	—	—	0.26***
Demographic variables
Age	41.61	12.39	—	—	0.01
Partner's age	40.97	13.58	—	—	0.01
Age difference between partners	4.37	4.33	—	—	−0.06
Differing partner's ethnicity	—	—	—	—	−0.05
Relationship duration	15.33	11.55	—	—	0.05
Educational level	3.71	1.05	—	—	−0.01
Socioeconomic status	3.87	2.17	—	—	−0.01

Abbreviations: CRS, Couple Relationship Scale. C‐SCORE, Systemic Clinical Outcome Routine Evaluation for Couples; M, mean; MHI‐5, Mental Health Inventory‐5; RDAS, Revised Dyadic Adjustment Scale; WHO‐5, World Health Organization 5‐item Wellbeing Scale; SD, standard deviation; α, Cronbach's alpha; Ω, McDonald's omega.

Note: N, 300. α and Ω values of 0.7 and 0.9 are considered acceptable and excellent, respectively. Differing partner's ethnicity ranges from Different = 1 to Same = 2 Educational level ranges from Primary school = 1 to Doctoral level = 6. Socioeconomic status ranges from Unemployed = 0 to 6 = higher professional or senior manager. Correlations of 0.1, 0.3, and 0.5 are considered weak, medium and large, respectively (Cohen 1988).

^***For 2‐tailed correlations, p < 0.001.

3.4. Concurrent and Construct Validity of CRS

In answer to the fourth research question, Table 2 shows the CRS had strong (r > 0.5) significant (p < 0.001) correlations with total and subscale scores of other couple relational functioning measures (C‐SCORE and RDAS) ranging from 0.70 to 0.80. There were strong (r = 0.51) and moderate (r = 0.42) significant (p < 0.001) correlations between the CRS and mental health (MHI‐5) and wellbeing scales (WHO‐5), respectively. There were weak to strong significant (p < 0.001) correlations between the CRS and self‐rated couple issues, ranging from 0.26 to 0.67. Specifically, the correlation between CRS and the item “How are you managing the problems/challenges together with your partner?” was the strongest (r = 0.67), whereas the weakest correlation (r = 0.26) was between CRS and the item “Do you think couple or family therapy could be helpful to your relationship?”

3.5. Concurrent Validity of CRS Items

In answer to the fifth research question about the concurrent validity of CRS items, Table 3 shows CRS items and corresponding factor scales of other couple relational functioning measures (C‐SCORE and RDAS) had significant (p < 0.001) moderate (r > 0.3) to strong (r > 0.5) correlations ranging from 0.43 to 0.79. The strongest correlations were between the CRS commitment item and the C‐SCORE commitment, trust and respect subscale (r = 0.79); and between the CRS physical intimacy item and the C‐SCORE sex and intimacy subscale (r = 0.78). The weakest correlation was between the CRS conflict item and the RDAS consensus subscale (r = 0.43). There were strong significant correlations between the 10th item of the CRS scale, which assesses personal wellbeing and mental health (MHI‐5, r = 0.58) and wellbeing (WHO‐5, r = 0.51) scales.

Table 3.

Correlations between CRS items and corresponding RDAS and C‐SCORE subscales and well‐being scales.

CRS Items	Relevant RDAS and C‐SCORE subscales, and wellbeing and mental health scales	r
Emotional intimacy Distant – Close	C‐SCORE Sex and intimacy I am satisfied with the intimacy that we share I am satisfied with the kind of sex life that we have	0.49***
Commitment Giving up—Completely committed	C‐SCORE Commitment, trust and respect We are both committed to this relationship I often think about leaving the relationship (R) It feels miserable in our relationship (R) We respect each other I trust my partner My partner trusts me Being in the relationship is important to me	0.79***
	RDAS Satisfaction We never discuss or have considered divorce, separation, or terminating our relationship We never quarrel I never regret that I married (or lived together) We never get on each other's nerves	0.69***
Trust Suspicious—Trusting	C‐SCORE Commitment, trust and respect I trust my partner My partner trusts me We are both committed to this relationship I often think about leaving the relationship It feels miserable in our relationship We respect each other Being in the relationship is important to me	0.65***
Safety Not at all safe – Perfectly safe	C‐SCORE Safety My partner becomes aggressive when they are upset (R) It feels risky to disagree in our relationship (R) My partner controls too much in our lives (R) My partner blames me when things go wrong (R) Our relationship is affected too much by prejudice or discrimination we experience (R)	0.47***
Cohesion All alone—Part of a team	RDAS Cohesion We engage in outside interests together We have a stimulating exchange of ideas We work together on a project We calmly discuss something	0.49***
Acceptance Criticized—Accepted	C‐SCORE Communication and problem‐solving My partner shows that they care I am listened to by my partner In our relationship, we feel safe to show the full range of how we feel We are able to find new ways to deal with things that are difficult We talk to each other about the things that matter to us	0.67***
Conflict Always Fighting – Get along perfectly	C‐SCORE Consensus on care, housework, and money We agree on how we carry out our caring responsibilities for others We agree on how we share household tasks We agree about how we spend money	0.50***
	RDAS Consensus We agree on religious matters We agree on demonstrations of affection We agree about making major decisions We agree on sex relations We agree on conventionality (correct or proper behavior) We agree on career decisions	0.43***
Physical intimacy No needs met – All needs met	C‐SCORE Sex and intimacy I am satisfied with the kind of sex life that we have I am satisfied with the intimacy that we share	0.78***
Overall happiness Extremely unhappy –Perfectly happy	RDAS Satisfaction We never discuss or have considered divorce, separation, or terminating our relationship We never quarrel I never regret that I married (or lived together) We never get on each other's nerves	0.66***
Personal well‐being Worst I've ever felt – Best I ever felt	WHO‐5 I have felt cheerful in good spirits I have felt calm and relaxed I have felt active and vigorous I woke up feeling fresh and rested My daily life has been filled with things that interest me	0.51***
	MHI‐5 I have been a very nervous person I felt calm and peaceful (R) I felt downhearted and blue I have been a happy person (R) I felt down in the dumps that nothing could cheer me up	0.58***

Abbreviations: CRS, Couple Relationship Scale; C‐SCORE, Systemic Clinical Outcome Routine Evaluation for Couples; MHI‐5, Mental Health Inventory‐5; N, 300; r, Pearson product‐moment correlation; R, reverse score this item; RDAS, Revised Dyadic Adjustment Scale; WHO‐5, World Health Organization 5‐item Wellbeing Scale.

^***For 2‐tailed correlations, p < 0.001.

3.6. Associations With Demographic Variables

Tables 2 and 4 present results relevant to the sixth research question about the relationship between the CRS and demographic variables. The final section of Table 2 contains correlations between CRS and continuous demographic variables. Table 2 shows that correlations between CRS and age, partner's age, age difference between partners, differing partner's ethnicity, relationship duration, educational level, and SES were weak and nonsignificant (p > 0.05), ranging from r = −0.06 to 0.01. Table 4 presents ANOVA and post‐hoc test results assessing the significance of differences in mean CRS scores for subgroups on categorical demographic variables. Table 4 shows that relationship status and partner ethnicity were the only two demographic variables that had significant associations with CRS scores. When asked about their relationship status, participants who were married, civil partnered, or single, although in a couple relationship for longer than a year, had lower mean CRS scores than those who ticked the “other” or “prefer not to say” options. Participants whose partners were Asian had higher mean CRS scores than those whose partners were white, black, Hispanic, mixed race or from indigenous minorities, for example, Aboriginal. There were no significant associations between CRS and gender, ethnicity, country of residence, sexual orientation, or couple therapy status.

Table 4.

ANOVA results for comparison of mean CRS scores of subgroups based on categorical demographic variables.

	N	M	SD	F	p	Group differences
Gender
1. Female	147	78.48	20.41	0.841	0.50	1 = 2 = 3 = 4
2. Male	141	80.55	18.50
3. Trans	2	95.50	6.36
4. Nonbinary or other	9	86.22	10.41
5. Other	1	87.00	—
Ethnicity
1. White/Caucasian	242	80.28	18.36	1.80	0.112	1 = 2 = 3 = 4 = 5 = 6
2. Black	20	68.50	27.78
3. Asian	18	84.28	16.36
4. Hispanic	7	85.29	14.33
5. Indigenous minority, e.g., Aboriginal	1	90.00	—
6. Mixed race	10	78.50	24.167
Partner's ethnicity
1. White/Caucasian	239	80.28	18.43	3.481	< 0.05	1 = 2 = 4 < 3
2. Black	20	67.55	28.49
3. Asian	21	85.71	14.88
4. Hispanic/Indigenous minority/Mixed race	18	80.22	18.88
Country of residence
UK	109	78.53	18.70	0.518	0.82	1 = 2 = 3 = 4
USA	102	82.25	20.29
Australia	44	78.05	20.95
Europe	45	79.30	6.64
Sexual orientation
1. Heterosexual	255	79.99	19.38	0.52	0.76	1 = 2 = 3 = 4 = 5 = 6
2. Lesbian	5	82.60	14.86
3. Gay	6	84.50	10.88
4. Bisexual	23	74.61	23.62
5. Pansexual	9	82.22	9.81
6. Other or prefer not to say	2	87.50	7.78
Relationship status
1. Single	13	78.15	21.73	4.85	< 0.05	1 = 2 < 3
2. Married or civil partnered	255	78.67	19.77
3. Other or prefer not to say	32	89.72	8.93
Couple therapy status
1. Not in couple therapy	275	79.98	19.23	1.00	0.367	1 = 2 = 3
2. In couple therapy now	5	67.80	28.46
3. Previously in couple therapy	20	80.75	17.27

Abbreviations: ANOVA, analysis of variance; CRS, Couple Relationship Scale; F, F ratio from one‐way ANOVA; M, mean; N, Number of cases; p, probability level; SD, standard deviation.

Note: Group differences are from the Tukey–Kramer post hoc test for groups of unequal sizes.

4. Discussion

The aim of the present study was to build on previous research by S. R. Anderson et al. (2022) and further validate the CRS by addressing six specific research questions. In answer to the first research question about the factor structure of the CRS, a CFA of CRS items showed a single factor structure fit the data. This finding is consistent with the results of the exploratory factor analysis reported in S. R. Anderson et al. (2022). The confirmation of the CRS's single‐factor structure addresses a key research priority identified by S. R. Anderson et al. (2022). It indicates that all 10 items load onto a common underlying construct, enhancing the scale's utility in applied settings. As a unidimensional measure, the CRS permits the aggregation of item responses into a single composite score, enabling clinicians to obtain meaningful insights into their clients' perspectives on their relationships efficiently and with minimal time investment. Furthermore, this structural coherence offers additional support for the scale's strong construct validity.

In answer to the second research question, there was no evidence for gender‐based measurement invariance. Thus, CRS scores of males and females may be interpreted in the same way and validly compared.

In answer to the third research question about the CRS's reliability, consistent with findings from the original CRS validation study where an alpha of 0.93 was obtained (S. R. Anderson et al. 2022), we found that the CRS had excellent internal consistency reliability, with an alpha value of 0.96.

In answer to the fourth research question about the concurrent and construct validity of the scale, CRS was strongly correlated with other measures of relational functioning and had strong and moderate correlations with mental health and well‐being scales, respectively. These results are consistent with those of S. R. Anderson et al. (2022) who found a strong correlation (r = 0.83) between CRS and the Couple Satisfaction Index (Funk and Rogge 2007).

In answer to the fifth research question about the concurrent validity of CRS items, we found that CRS items had moderate to strong correlations with corresponding factor scales of other couple relational functioning measures ranging from 0.43 to 0.79 and strong correlations between the CRS item that assesses personal well‐being and mental health (r = 0.58) and well‐being (r = 0.51) scales. This is consistent with S. R. Anderson et al. (2022) report of strong correlations ranging from absolute values of 0.50 to 0.82 between 9 out of 10 CRS items and scales that assessed corresponding constructs in a community sample. However, S. R. Anderson et al. (2022) found that, in a community sample, the CRS acceptance item had low correlations with each of the two‐items in the Perceived Criticism Measure (PCM; Hooley and Teasdale 1989) which assesses how critical the participant is of their partner (r = 0.01) and how critical the partner is of the participant (r = −0.29). In our study we found that the CRS acceptance item had a strong correlation with the C‐SCORE communication and problem‐solving subscale (r = 0.67) which contained two items that indicate acceptance by one's partner (My partner shows that they care; I am listened to by my partner) and three items that indicate mutual acceptance (In our relationship, we feel safe to show the full range of how we feel; We are able to find new ways to deal with things that are difficult; We talk to each other about the things that matter to us).

In answer to the sixth research question about associations between CRS and demographic variables, partner ethnicity and relationship status were the only two (of 14) demographic variables that had significant associations with CRS. Participants whose partners were Asian, and participants who ticked the “other” or "prefer not to say” options, when asked about their relationship status obtained higher CRS scores, indicating better relationship functioning. This finding on relationship status is difficult to interpret because it is not clear what “other” or “prefer not to say” options refer to, apart from the fact that these participants were in a cohabiting or marital heterosexual or nonheterosexual couple relationship for at least 1 year, since this was an inclusion criterion that all participants met. The absence of associations between CRS and 12 other demographic variables was a particularly noteworthy finding. CRS did not have significant associations with age, partner's age, age difference between partners, differing partner's ethnicity, relationship duration, educational level, SES, gender, ethnicity, country of residence, sexual orientation, and couple therapy status. Consistent with the findings of S. R. Anderson et al. (2022), we found no significant relationship between gender and CRS scores.

These results collectively add validation support to the CRS measure of relational functioning. The primary contribution of this study lies in its extension of S. R. Anderson et al. (2022) work, providing further empirical support for the CRS as a reliable, valid, acceptable, and relevant measure of relational functioning.

One limitation of this study was the reliance on a community sample only, without inclusion of a clinically distressed sample. Additionally, the majority of participants were from Western countries, such as the United States and United Kingdom. Most individuals in the sample were highly educated and belonged to a higher socioeconomic group, as indicated by their occupations. These characteristics limit the generalizability of the findings to non‐western countries and lower socioeconomic groups.

A key strength of the study was its use of validated measures of relational functioning and mental health. This enhanced the reliability and validity of results and supported robust conclusions (Barry et al. 2014).

4.1. Clinical Implications and Future Research

The CRS demonstrates strong psychometric properties and brevity, making it an acceptable ROM to use across treatment sessions in couple therapy. ROM's, such as the CRS, provide immediate and quantifiable feedback on relational functioning. These measures enable therapists to monitor client progress, adjust interventions in real‐time, and identify early indicators of relational difficulties (Anker et al. 2009; Lambert 2016). Incorporating such feedback into clinical practice has been shown to enhance therapeutic outcomes, strengthen client engagement, and improve the therapeutic alliance (Reese et al. 2010; Shimokawa et al. 2010).

Item‐level CRS scores can further support targeted case formulation. For example, low emotional intimacy may signal patterns of attachment withdrawal, low perceived safety may indicate the need for stabilization, and elevation in conflict accompanied by low acceptance may reflect escalation patterns that can be addressed through behavioral or integrative approaches (Christensen et al. 2010). By mapping directly onto recognized therapeutic constructs, CRS data can be systematically incorporated into clinician' conceptual models.

ROMs also contribute to narrowing the research‐practice gap by generating practice‐based evidence, compliments findings of controlled trials, and supports service‐level decision making and quality improvement initiatives (Barkham et al. 2023; Halford et al. 2016). Moreover, sharing ROM feedback with couples fosters collaboration, strengthens the therapeutic alliance, and facilitates shared goal‐setting (Bickman et al. 2012; Knobloch‐Fedders et al. 2015).

Further research is required to establish the validity of the CRS in diverse clinical populations, assess its sensitivity to change, and evaluate its applicability across varying relationship types, cultural contexts and sexual orientations. Incorporating data from both partners in empirical studies would allow actor‐partner interdependence analysts (Bodenmann 2005; Funk and Rogge 2007). Overall, the CRS is a brief, reliable, and clinically meaningful ROM that supports evidence‐informed practice.

Ethics Statement

This study was conducted with ethical approval of the Human Research Ethics Committee, University College Dublin, and informed consent of participants.

Starken, P. , Stratton P., Kieran A., et al. 2025. “Further Validation of the Couple Relationship Scale.” Journal of Marital and Family Therapy 52: 1–13. 10.1111/jmft.70100.