The severe impact of intimate partner violence (IPV), a form of interpersonal trauma, on survivors’ health and well-being is well documented, showing increased risk of depression, post-traumatic stress disorder (PTSD), anxiety, somatic symptoms, substance abuse, and suicidal ideation ^[1][2][3][4]. This suffering is associated with functional impairment, low sense of coherence (SOC), and substantially reduced quality of life ^[5][6][7], even years after leaving the abusive relationship ^[8][9].

Despite these adverse outcomes related to IPV, previous research has shown that help-seeking among survivors is low. Some never seek help, and those who do mainly choose informal sources of help, usually from their family or friends and are less likely to seek formal help, such as from shelters, healthcare services, or the police ^{[10][11][12][13]}. The IPV help-seeking literature is primarily focused on escaping the violence and attending to the immediate harm caused. While these often first steps are critical, there is a need for an increased focus on survivors’ pathways for trauma recovery ^[14][15][16]. Help-seeking after IPV is a complex journey involving a series of meaning-making judgments and socially engaged and culturally informed actions ^[17][18][19], and the road to recovery is often challenging ^[20][21].

Findings of low help-seeking rates are consistent with other studies reporting that IPV survivors are faced with a wide range of barriers on sociocultural, structural, interpersonal, and individual levels, e.g., normalization of violence, access challenges, fearing consequences of disclosure, and self-blame ^{[22][23][24][25]}. Moreover, studies have indicated that survivors with depression, PTSD, and low SOC face even more significant barriers to help-seeking, such as symptom burden, fearing mental illness stigmatization, and a weak sense of manageability and meaning, making it more challenging to take action ^{[21][26][27][28]}.

In a widely used definition based on a review of definitions, mixed methods research is defined as a “type of research in which a researcher or team of researchers combines elements of qualitative and quantitative research approaches (e.g., use of qualitative and quantitative viewpoints, data collection, analysis, inference techniques) for the broad purposes of breadth and depth of understanding and corroboration ^[29] (p. 123)”.

One of the earliest examples of using multiple research methods for validation dates to the 1950s, with Campbell and Fiske’s ^[30] framework giving rise to methodological triangulation and arguing that the convergence of findings derived from more than one method would strengthen the evidence of validity. However, as innovative and valuable their framework has been, it is first and foremost quantitative (Quan). To date, in the instrument development literature, construct validation is often conceived as mainly a Quan endeavor ^{[29][31][32][33][34]}. When qualitative (Qual) data are used, it is usually only granted a supplementary role to Quan data, and often the methods are utilized in isolation rather than fully integrated ^[32][34][35]. Still, there is a growing literature on mixed methods validation. A few frameworks have been developed that place equal value on Quan and Qual methods, focusing on validity and trustworthiness, and emphasizing the integration or “mixing” of findings from both databases to inform validation evidence for a measure ^{[31][33][34][36]}.

The term legitimation ^[37] has been recommended to refer to validity and quality in mixed methods studies, as it considers both Quan and Qual research paradigms ^[38][39]. The “fit” of data integration refers to the coherence of Quan and Qual findings ^[40]. Such assessment is likely to lead to four possible outcomes: Confirmation is when the findings are consistent with each other, supporting drawing the same conclusion from each. Complementarity is when the findings tell different but nonconflicting stories (reflecting different sides of the same coin). Expansion is when the findings diverge to a certain degree but, when combined, can expand insights. Discordance is when the findings are inconsistent, contradictory, or disagree with each other ^[40][41].

The Barriers to Help-Seeking for Trauma (BHS-TR) scale was developed from an existing mental health barriers measure ^[42] focusing on service use for mental disorders. Based on an international literature review about barriers to seeking help after trauma and findings from focus groups and individual interviews with American and Irish gender-based violence (GBV) survivors, the original scale was adapted for GBV survivorship ^[21][43]. New items about normalization, shame, mistrust, perceived rejection, being afraid of the consequences of disclosure, and feeling frozen were added, making the measure more trauma-specific and survivor-centered. The early work on the BHS-TR scale indicated that the barriers could be grouped into structural and internal dimensions ^[21], which was later confirmed in a psychometric study among American GBV survivors. Moreover, a seven-factor structure was revealed (Unavailable/Not Helpful; Financial Concerns; Discrimination; External Constraints; Shame; Frozen/Confused; and Problem Management Beliefs), and the scale was found to be reliable and valid ^[44].

The BHS-TR was translated and cross-culturally adapted into the Icelandic language and context ^[45] and initially validated in a mixed methods study among IPV survivors in Iceland ^[46], creating the first Icelandic trauma-specific measure that assesses help-seeking barriers. An essential part was qualitatively evaluating the scale through cognitive interviewing (n = 17), resulting in the development of new barrier items based on the survivors’ lived experiences. These new items represented barriers related to viewing help-seeking as a sign of weakness and the desire to safeguard oneself from re-traumatization. Using these findings, building was utilized to adapt the BHS-TR scale, and then a psychometric evaluation of the whole instrument with the additional items was carried out (n = 137). Both Qual and Quan phases provided evidence that the Icelandic BHS-TR is relevant, reliable, and valid ^[45][46]. Nevertheless, there was a noticeable mismatch between the Qual and Quan findings regarding several items on the scale. Primarily, items that were significant barriers to help-seeking in the survivors’ narratives were problematic in the exploratory factor analysis, mainly due to cross-loadings with different factors, indicating the removal of these items ^[46]. This mismatch between the participants’ narratives and the factor analysis results pointed to a legitimation issue, demonstrating the need for further systematic assessment of the coherence of barriers to help-seeking Qual and Quan findings.

A mixed methods legitimation strategy of integration was employed to evaluate the BHS-TR structure by merging the Qual and Quan data through a joint display analysis  ^[47][48] and examining the coherence of the findings.

The joint displays linking the Qual and Quan findings are shown in Table 1 (Structural Barriers) and Table 2 (Internal Barriers), revealing evidence of complementarity, expansion, and discordance. To illuminate the lived experiences of the barriers, exemplar quotations from the survivors were chosen and reported in the Qual columns. The items referred to (using their numbers) in the Quan columns can be found in Supplementary Tables S1 and S2.

Characteristics	Qual Phase (n = 17)	Quan Phase (n = 137)
Age
18–29	4 (23.5%)	24 (17.5%)
30–39	7 (41.2%)	34 (24.8%)
40–49	4 (23.5%)	38 (27.7%)
50–59	1 (5.9%)	18 (13.1%)
60+	1 (5.9%)	6 (4.4%)
Not stated	-	17 (12.4%)
Racial and ethnic background
Caucasian	17 (100%)	-
Iceland-born	16 (94.1%)	-
Foreign-born	1 (5.9%)	-
Level of education
High school or less	3 (17.6%)	11 (8.0%)
Technical or junior college degree	5 (29.4%)	29 (21.2%)
University degree	9 (52.9%)	82 (59.9%)
Not stated	-	15 (10.9%)
Employment status (not mutually exclusive)
Working	12 (70.6%)	88 (64.2%)
Unemployed or looking for work	2 (11.8%)	7 (5.1%)
Student	5 (29.4%)	26 (19.0%)
Homemaker	1 (5.9%)	3 (2.2%)
Unable to work due to sickness/disability	3 (17.6%)	20 (14.6%)
Other	-	24 (17.5%)
Number of children
None	5 (29.4%)	24 (17.5%)
One or two	9 (52.9%)	59 (43.1%)
Three or more	3 (17.6%)	46 (33.6%)
Not stated	-	8 (5.8%)
Years in the abusive relationship
1–5	4 (23.5%)	-
6–10	9 (52.9%)	-
11–15	2 (11.8%)	-
15+	2 (11.8%)	-
Years out of the abusive relationship
1–5	10 (58.8%)	-
6–10	6 (35.3%)	-
11–15	1 (5.9%)	-
Current medical diagnosis (mental and/or physical)
No	6 (35.3%)	44 (32.1%)
Yes	11 (64.7%)	93 (67.9%)
History of receiving mental healthcare
No	8 (47.1%)	24 (17.5%)
Yes	9 (52.9%)	112 (81.8%)
Not stated	-	1 (0.7%)

Conceptual Structure	Qualitative Phase	Quantitative Phase	Coherence of Findings
Structural Barriers	In the interviews, participants (14 of 17) generally made a specific distinction between structural and internal barriers to seeking help. When discussing structural barriers, the women mainly mentioned system-level barriers referring to healthcare and social services. Findings provided evidence of relevance, face validity, and content validity. “There are so many walls to climb over in our system, and when you are so shattered and exhausted, you just can’t.”	The “Structural Barriers Index” included Financial Concerns, Unavailable/Not Helpful, External Constraints, and Inconvenience factors. The index had good internal consistency (α = 0.75), and the results provided evidence of convergent, discriminant, and known-groups validity.	Complementarity: Not included in subsequent analysis.
Financial Concerns	A majority (12 of 17) of the participants agreed that the items about financial concerns were significant barriers, especially related to seeking professional psychological help, as the Icelandic Health Insurance covers not all mental healthcare. Findings provided evidence of relevance, face validity, and content validity. “Let’s be clear, getting professional help to work through your trauma is hardly part of our great welfare system, and I couldn’t even pay the bills, let alone go to a psychologist.”	The “Financial Concerns” factor comprised items #2, 19, and 18. All items had high factor loadings, and the internal consistency was good (α = 0.82). Results provided evidence of convergent, discriminant, and known-groups validity.	Complementarity: Not included in subsequent analysis.
Unavailable/Not Helpful	While these items did not represent the main barriers hindering the women from seeking help, more than half (11 of 17) said that the healthcare they needed had not been available to them. Findings provided evidence of relevance, face validity, and content validity. “I didn’t tick in the right boxes when I finally had the courage to go to the hospital. Like sure honey, we will stitch up your head … but you won’t get mental healthcare there.”	The “Unavailable/Not Helpful” factor comprised items #15, 16, and 17. All items had high factor loadings, and the internal consistency was good (α = 0.71). Furthermore, results provided evidence of convergent, discriminant, and known-groups validity.	Complementarity: Not included in subsequent analysis.
External Constraints	Many (11 of 17) participants were afraid of the consequences of seeking help, and the other external constraints impacted them as well. Findings provided evidence of relevance, face validity, and content validity. “You can lose so much … friends and family members who sided with him, and I knew he would use it against me in the custody battle … unfit mentally ill mother.”	The “External Constraints” factor comprised items #14, 34, and 25. All items had high factor loadings, and the internal consistency was good (α = 0.77). Furthermore, results provided evidence of convergent, discriminant, and known-groups validity.	Complementarity: Not included in subsequent analysis.
Inconvenience	The inconveniences barriers were not the foremost reasons stopping participants from seeking help. However, a majority (10 of 17) thought these barriers were part of the picture. The most mentioned was the time factor. Findings provided evidence of relevance, face validity, and content validity. “There was no time … being a single mom working a full-time job doesn’t give you a lot of space.”	The “Inconvenience” factor was comprised of items #5 and 8. The items had high factor loadings, but the internal consistency was poor (α = 0.52). Evidence of convergent, discriminant, and known-groups validity was provided. One inconvenience item (#9) about not getting time away from work or family needed to be dropped as it did not load significantly onto this or any other factor.	Expansion: Item #9 was included in the subsequent analysis.
Discrimination	The participants interpreted the prejudice and discrimination items as relating to race and ethnic background, which did not apply to them but recognized these items would be important for the survivor immigrants to Iceland. Yet, many (13 of 17) said they were worried about and experienced prejudice and discrimination for being an IPV survivor. These experiences centered around stereotyping and victim-blaming. “Take the risk of revealing myself as a victim … no. When people know your story, it’s like you become nothing else, the weak abused women stamp is burnt to your forehead.”	All the discrimination items (#20, 21, and 23) were identified as problematic due to cross-loadings onto different factors and needed thus to be dropped.	Expansion: Items #20, 21, and 23 were included in subsequent analysis.

Most of the Qual and Quan findings were congruent, reinforced one another, were deemed complementary, and strengthened the overall trustworthiness and validity evidence of the BHS-TR. Divergent (expansion and discordance) findings were critically evaluated in an iterative spiraling process ^[49][50] that illuminated core issues and guided potential refinements; that can help the scale better capture the significant hindrances faced and the immense amount of effort survivors often take to seek help.