The Cognitive Triad Inventory (CTI) psychometric properties on a Romanian sample

Abstract

Background

This study aims to develop the Romanian version of the Cognitive Triad Inventory (CTI) by documenting the Romanian translation of the instrument items and its psychometric characteristics. We were concerned with the items’ adaptation, analyzed the scale’s reliability and the three subscales’ and assessed the construct validity of the Romanian sample.

Methods

The instrument contains 36 items that assess the cognitive triad, both positive and negative, and was administered to a nonclinical sample (N = 675). To maintain the conceptual equivalence of the original instrument, translation was executed according to literature standards, including forward and backward translation, expert reviews, to ensure cultural and contextual relevance.

Results

The confirmatory factor analysis on the Romanian scale items showed very good factorial validity in the present sample. The psychometric evaluation highlighted strong internal consistency with the global scale and the three dimensions: view of self, view of world, and view of future, supporting its robustness for research purposes. Convergent and discriminant validity were supported through correlations with a similar instrument. Three-month test–retest results indicated good stability over time. Cross-cultural configural invariance was preliminarily supported by significant correlations between Romanian factor loadings and those from German and Polish samples.

Conclusion

The findings underscored the CTI’s utility in assessing cognitive patterns within diverse cultural contexts.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-025-03581-4.

Introduction

Understanding the cognitive perspectives of mental health has been a longstanding focus in psychological research, particularly in the context of depression and well-being. One of the most influential theoretical models in this domain is the cognitive triad, which provides a framework for analyzing how individuals perceive themselves, their environment, and their future. The negative cognitive triad was conceptualized by Beck, Rush, and Shaw (1979) [1] in the study of depression, but the items reflect both the negative and the positive cognitive frameworks, being constructed for both directions [2]. While the negative cognitive triad is associated with depression, the positive cognitive triad is based on positive emotions that, together with satisfaction with life, contribute to individual happiness [3]. The positive cognitive triad is linked to well-being and resilience, involving optimistic perceptions of oneself, the surrounding world, and prospects. The negative cognitive triad is a hallmark of depression: individuals with this thinking pattern tend to harbor feelings of worthlessness, view the world as overwhelming or unjust, and expect negative outcomes or failures [1]. This pessimistic outlook often reinforces itself, deepening the cycle of depressive symptoms. If a negative self-view is related to low self-esteem and underestimation of abilities [2] a positive self-view has the opposite effects. The cognitive triad impacts quality of life, mental health, and well-being. The present research aims to adapt an instrument that measures the cognitive triad for the Romanian-speaking population. The results of this study can contribute to developing future interventions to reduce depression and maximize resilience and well-being. The cognitive triad indicates in which of the three domains—perspective on the self/world/future—there is a greater need for therapeutic interventions.

We study the cognitive triad to understand how individuals can participate in what is positive and expect the best. People tend to remember negative things and expect the worst to happen [4]. It is important to investigate how we can re-educate thinking away from the negative, towards the positive. The positive cognitive triad focuses on a positive self-evaluation, an optimistic view of world, and hope for the future, while the negative cognitive triad focuses on a negative self-image, a pessimistic view of world, and a lack of hope. The negative cognitive triad has been more studied, especially in research related to the theory of depression, starting with Beck (1987) [5], while the effects of the positive cognitive triad, including effects on well-being and resilience, are understudied.

The objective of this paper is to adapt the instrument The Cognitive Triad Inventory (CTI), originally developed by Beckham, Leber, Watkins, Boyer, and Cook in 1986 [6], to the Romanian population and to guarantee the psychometric tool’s accuracy in measuring cognitive patterns, considering cultural relevance. By evaluating the local applicability of CTI, researchers can gain valuable knowledge on the impact of cognitive patterns on well-being, resilience, and depression within Romanian cultural contexts, which can enhance the quality of mental health assessments and interventions.

State of science

The negative cognitive triad

Depression is a disabling factor that affects a significant number of people worldwide. During their lifetime, 25% of women and 12% of men experience at least one major depressive episode [7]. Depression is one of the most common causes of disability for people aged 15–44 years worldwide [8], explaining a percentage of 12% of disability cases [9]. Nearly 80% of the population suffering from moderate or severe depression in the US reported that their work, family life, and social functioning were affected [10].

Cognitive theories claim that specific cognitive frameworks known in literature, as self-schemas or cognitive schemas that emerge during childhood, are susceptible to the onset of depression, due to their adverse impact on memory, interpretation, and attention. Such maladaptive schemas arise from adverse experiences and negative interactions with caregiving figures during childhood and they can be identified in individuals with depression, through feelings of hopelessness and helplessness [11]. In the study of depression, in 1987, Beck brought a new concept, the cognitive triad, which included negative thoughts about oneself (the perception of oneself as worthless and inadequate), about the world (others are unfair and prevent you from achieving your goals) and about the future (lack of hope and optimism about the future) [5]. Beck’s theory is based on the stress-vulnerability model which posits that individuals with a predisposition to certain cognitive patterns are more susceptible to depression when exposed to stress or negative activating events, while stress-activating depressive schemas about oneself can remain inactive under other conditions [5]. Beck’s theory presents negative automatic thoughts and cognitive errors, as defining factors for the development and maintenance of depression [12]. Automatic thoughts and dysfunctional beliefs underlie the development of depression [13], so a negative cognitive style is a vulnerability factor for depression [14]. Newer cognitive models consider the process or strategy of thinking more important than the content of thoughts [7]. Depressive states amplify negative thoughts and ruminations, creating a spiral in which the two influence each other. Rumination increases the effect of depressive states on thinking, making the use of negative thoughts and memories more likely [15]. Mor and Winquist (2002) concluded in a meta-analysis that rumination correlates most strongly with depressive symptoms [16]. It is important to recognize the symptoms associated with depression and seek professional help so that rumination and negative maladaptive cognitions do not further potentiate the state of depression.

The positive cognitive triad

Mak, Ng, and Wong (2011) use the opposite of the negative cognitive triad: people with a high level of resilience have a thinking style called the positive cognitive triad, which makes them feel good about themselves, protecting them from depression [17]. Mak et al. (2011) used CTI to measure the positive cognitive triad, with its three components: view of self, view of world, view of future [18]. Studies by Pössel (2009) on the German-speaking population [2], Mak et al. (2011) in China [17], Śliwerski (2014) on the Polish population [14], Mehta et al. (2018) in the US [19], and Erarslan and Işikli (2019) on the Turkish-speaking population [20] show that the theory of the positive cognitive triad is valid in several countries. Resilient people have a positive perspective on themselves, world, and future, and implicitly a higher level of well-being [17]. Humans adjust to environmental stimuli and revert to their former level of well-being after both bad and positive experiences [21]. Caprara et al. (2009) define positive self-view, positive view of others, and positive view of future with the term “positivity”, and this is what the positive cognitive triad also captures [22]. Positivity acts as a psychological buffer, protecting individuals from the negative effects of stress by fostering adaptive coping mechanisms and a hopeful outlook. People with a positive view of the world are more efficient at solving difficulties and recognizing opportunities in trying circumstances [23]. They also tend to engage in healthier interpersonal relationships, creating a reciprocal emotional support and positivity cycle. These relationships are an additional resource during adversity, providing emotional and practical assistance. Individuals who maintain a positive view of future are less likely to experience chronic stress and depression. Anticipation of a better future fosters hope, a critical component of resilience. Hopeful individuals tend to set realistic goals and engage in proactive behaviors that contribute to their well-being [24]. As captured by the positive cognitive triad, future-oriented optimism strengthens an individual’s capacity to recover from setbacks, enabling them to maintain well-being when facing significant challenges. The above mentioned studies show that the positive cognitive triad theory works across cultures - individuals with a positive cognitive triad are better equipped to navigate life’s challenges.

This study’s results are relevant for clinical practice in screening depressive and negative thoughts. It may also be useful in guiding psychotherapy interventions by addressing the three cognitive structures to show where restructuring is required or monitoring the progress of (cognitive) therapy. It also draws attention to the cognitive triad’s role in a better quality of life and increased well-being.

Aims of the study

The proposed study aims to translate, culturally adapt, and validate CTI for the Romanian population. This research contributes to the literature by providing psychologists with a scientifically validated tool that measures cognitions about oneself, the world, and the future, assessing cognitive patterns associated with depression, anxiety or well-being. The adaptation process ensures the tool is linguistically appropriate and culturally relevant for Romanian individuals.

Method

Participants

The study, based on non-probability sampling, was conducted on Romanian adults (N = 675). The participants’ ages ranged from 20 to 68 years (M = 39.633, SD = 10.379). Women were overrepresented (75%), as they were more willing than men to respond to questions.

A different group of participants (N = 62) was assessed afterwards at two points in time to evaluate test–retest reliability. The administration was conducted three months after the initial measurement. The participants’ ages ranged from 19 to 55 years (M = 32.911, SD = 9.150).

According to Wolf et al. (2013), regarding the convenience lot size, for a CFA analysis with a minimum statistical power of 0.80, for the smallest estimated parameter of interest and a correlation between factors of 0.30, and a probability of error (p) of 0.05, the minimum required sample size is 160 participants for eight items per factor, in a model with three factors [25].

Procedure

Participants were recruited through social media platforms, email invitations, and online forums, using a form of convenience sampling (or snowball sampling, if participants also forwarded the invitation). This method was chosen due to its efficiency in reaching a broad and diverse sample within the target population, especially given the constraints of online data collection. Participants were informed about the study’s purpose and relevance through a personalized email or social media message containing a link to the questionnaire. Data was collected online, adhering to the deontological norms. Prior to completing the questionnaire, participants were assured of confidentiality and anonymity, and their informed consent was obtained. They were also made aware of their right to withdraw from the study at any point without consequence. The questionnaire scale was administered via Google Forms over two months. A cross-sectional design was used. None of the participants received any financial reward, and they were all guaranteed confidentiality when collecting the data. All data were securely stored, with access restricted to authorized members of the research team. The study received ethical approval, confirming compliance with established research ethics standards.

Measure

The 36-item Cognitive Triad Inventory (CTI), developed by Beckham, Leber, Watkins, Boyer, & Cook in 1986 contains three subscales, each with 10 items (see Supplementary Material 2) [6]. Items number 1, 2, 4, 7, 14, 22 are designated as fillers and, as specified in the scoring guidelines, are not included in the calculation of the final score. The answers for each item are on a 7-point Likert scale, from 1 (= total agreement), to 7 (= total disagreement) [18].

The most easily obtainable form of reliability is internal consistency, which represents the extent to which a construct is concentrated within the items that compose it [26]. CTI has very good internal consistency with Cronbach Alpha index`s values ​​between 0.81 and 0.95 on a Polish sample, as reported by Śliwerski [14]. On an American sample, the Cronbach Alpha index values are high: 0.91 for self-perspective, 0.81 for world-perspective, and 0.93 for future-perspective, 0.95 for the entire scale [18].

The American version of the CTI was translated from English into Romanian, considering the recommendations of the International Test Commission Guidelines for Translating and Adapting Tests [27]. The author, an authorized translator of the English language, three Romanian psychologists, conversant in the English language, and an American citizen (a native English speaker) participated in the translation. During this process, minor ambiguities emerged related to culturally embedded expressions (e.g., interpretations of “failure”, ”hostile place”, ”worthwhile human being” or “adequate”), which were resolved through discussion and iterative revisions to preserve the original meaning while ensuring natural phrasing in Romanian. Item 4 was changed, from a negative to a positive formulation, to eliminate double negation, which could hinder respondent comprehension. This does not affect the scoring, because item 4 is a filler. The translators worked independently. The method of translation and retroversion was used to check whether the two versions are linguistically equivalent. The process prioritized semantic and idiomatic accuracy, ensuring that the translated items preserved the meaning of the original content while remaining comprehensible to the target population. The translated versions were compared and refined until the final version was reached.

We have no information about the construct validity of the CTI original English form. Pössel (2009) [2] examined the psychometric properties and factor structure of the German version based on the original CTI in a sample of 796 participants. The study reported satisfactory internal consistencies for the three subscales (α = 0.77–0.86) and good test–retest reliability over four weeks (r =.67–0.76), as well as meaningful correlations of CTI subscales with depressive symptoms (e.g., negative CTI scores were strongly associated with higher depression scores), supporting its convergent validity. These findings provide evidence that the CTI captures theoretically coherent constructs consistent with Beck’s cognitive triad model, thereby supporting its construct validity.

Ryff Carol’s (1989) Scale of Psychological Well–Being (SPWB) with 42 items was used to assess six constructs: self-acceptance, autonomy, personal development, environmental control, positive relationships with others, and purpose in life [28]. We used the scale translated into Romanian by Miron and colleagues (2011) [29] and Tudorel and colleagues (2013) [30].

During the confirmatory factor analysis, two items (CTI_3 and CTI_16) were removed due to unsatisfactory psychometric performance. CTI_3 showed consistently low standardized loadings on the view of world factor, contributing to the reduced average variance extracted (AVE) for this dimension, while CTI_16 displayed problematic cross-loadings and modification indices suggesting poor model fit. Eliminating these items improved the factorial validity and internal consistency of the Romanian version, without altering the theoretical meaning of the constructs. Similar practices are common in cross-cultural validation studies when individual items underperform psychometrically (Hair et al., 2010).

Statistical data analysis

The factorial validity of the Romanian CTI version was assessed using Confirmatory Factor Analysis (CFA) in R version-4.4.1 [31]. The diagonally weighted least squares (DWLS) method was used to enhance the robustness of the model calibration. DWLS provides more accurate estimations of interfactor interaction and factor loadings compared to maximum likelihood (ML). Structural coefficient estimates obtained by DWLS tend to outperform ML, particularly when dealing with ordinal data [32]. Although 7-point Likert-type scales are sometimes treated as continuous and may justify the use of robust maximum likelihood (MLR), we chose the DWLS estimator given both the ordinal nature of the data and recent findings that caution against using MLR with sample sizes below 1000 [33]. We considered p-values < 0.05 to be statistically significant for the data analyses.

We examined the psychometric properties of the Romanian CTI version. The convergent and discriminant validity were calculated using similar psychological constructs measured with the Psychological Well-Being Scale [28] administered concurrently. The six components of psychological well-being measure content similar to the three components of the CTI as follows: personal growth and self–acceptance with the view of self, environmental mastery and positive relations with others with the view of the world, and purpose in life and environmental mastery with the view of the future. Discriminant validity is supported through correlations between CTI components and unrelated well-being dimensions, indicating that each CTI subscale measures distinct cognitive patterns aligned with its theoretical domain.

By confirming its reliability and validity, the study helps ensure this measure can be used in clinical practice and research settings in Romania. By providing an adapted instrument for assessing the cognitive triad, the study contributes to understanding the role of cognitive patterns in mental health.

Results

Descriptive statistics for the cognitive triad dimensions are presented in Table 1. The subscales exhibited sufficient variability, with slightly negative skewness and moderate to high kurtosis.

Table 1.

Descriptive statistics

Factor	M	SD	Median	Min	Max	Skewness	Kurtosis
View of self	60.10	8.24	62	20	70	−1.49	2.44
View of world	51.04	8.02	53	19	63	−1.03	1.05
View of future	54.78	8.24	57	11	63	−1.66	3.37

Reliability for the three factors that make up the scale was very good (see Table 2). These high internal consistency values show the reliability of each CTI subscale.

Table 2.

Reliability values

Factor	Cronbach’s α (95% CI)	McDonald’s ω (95% CI)
View of self	0.86 [0.83–0.88]	0.92 [0.91–0.93]
View of future	0.91 [0.90–0.92]	0.95 [0.94–0.95]
View of world	0.80 [0.78–0.82]	0.86 [0.85–0.88]

We ran a Shapiro–Wilk normality test to check if the dataset followed a normal distribution (see Table 3) [34]. The test suggested that all three variables were nonparametrically distributed.

Table 3.

Shapiro-Wilk normality test

	W	p	Mean	SD
View of self	0.869	< 0.001	60.102	8.240
View of future	0.840	< 0.001	54.784	8.243
View of world	0.932	< 0.001	51.041	8.243

Nonparametric correlations were performed because the assumption of normality was not met for the cognitive triad factor scores (see Table 4).

Table 4.

Correlation matrix of the cognitive triad dimensions - Spearman’s correlations

			Spearman’s rho (latent factor correlation)	p	Lower 95% CI	Upper 95% CI
View of self	-	View of world	0.67 (0.84)	< 0.001	0.63	0.71
View of self	-	View of future	0.63 (0.84)	< 0.001	0.58	0.67
View of world	-	View of future	0.61 (79)	< 0.001	0.56	0.66

The Correlation Matrix in Table 4 shows significant correlations between the scales, with a large effect size, a necessary condition for CFA. Strong positive correlations between how individuals view themselves, the world, and the future suggest internal coherence while maintaining sufficient distinctiveness to warrant multidimensional measurement.

Confirmatory Factor Analysis (conducted with DWLS estimator) confirmed the instrument’s reliability and factorial validity, revealing an acceptable fit. We used three categories of fitness: incremental, parsimonious, and absolute fit, and we selected one fit index for each category: TLI (Tucker – Lewis Index) and CFI (Comparative Normed Fixed Index) for incremental fit, CMIN (Chi-square)/df (degrees of freedom) (the minimum discrepancy) for the parsimonious fit, RMSEA (mean square error of approximation) for absolute fit, and SRMR (standardized root mean square residual). Fit indices indicate good model fit: chi – square =.

1919.75, p <.001, CMIN/df = 5.55, CFI = 0.983, TLI = 0.98, RMSEA = 0.08, and SRMR = 0.07, meeting conventional cutoffs for comparative fit and residuals. Standardized loadings were statistically significant. The view of world items loaded between 0.57 and 0.74, and in the second-order specification, the general factor (PCT) loaded strongly on the three first-order factors (λ = 0.94 for view of self, 0.88 for view of world, 0.89 for view of future). Composite reliability was high (CR = 0.92 for view of self, 0.87 for view of world, 0.95 for view of future) and the average variance extracted was 0.53, 0.42, and 0.67, providing further evidence of construct reliability and convergent validity.

The standardized second-order CFA model is represented in Fig. 1. The higher-order positive cognitive triad (PCT) accounts for the common variance among the three first-order factors: view of self, view of world, and view of future, with strong second-order loadings. First-order item loadings remained positive and statistically significant, indicating adequate representation of each construct.

Fig. 1.

Confirmatory Factor Analysis for CTI

Discriminant validity (Fornell–Larcker, 1981) [35]. Table 5 reports √AVE on the diagonal and latent factor correlations off the diagonal. √AVE values were 0.730 for view of self, 0.647 for view of world, and 0.815 for view of future. Latent correlations ranged from 0.789 to 0.840 (view of self – view of world = 0.839; view of self – view of future = 0.840; view of world – view of future = 0.789). According to the Fornell–Larcker criterion [35], discriminant validity is supported when each construct’s √AVE exceeds its correlations with the other constructs. This condition was not fully met: for example, view of self – view of future (0.840) exceeded √AVE for view of self (0.730) and was comparable to √AVE for view of future (0.815), and view of world – view of future (0.789) exceeded √AVE for view of world (0.647). Thus, while convergent validity was adequate for view of self (AVE = 0.533) and view of future (AVE = 0.665) and lower for view of world (AVE = 0.418), discriminant validity among the three first-order factors is limited. This pattern is theoretically coherent with a strong common higher-order component. Accordingly, we modeled a second-order positive cognitive triad factor that accounted for the shared variance and showed strong second-order loadings (PCT → view of self 0.95; PCT → view of world 0.89; PCT → view of future 0.89), while retaining the first-order factors for interpretive specificity.

Table 5.

Fornell–Larcker discriminant validity matrix for the CTI. Diagonal = √AVE; off-diagonals = latent factor correlations (from the three-factor CFA, DWLS)

	View of self	View of world	View of future
View of self	0.730	0.839	0.840
View of world	0.839	0.647	0.789
View of future	0.840	0.789	0.815

Preliminary configural invariance was tested by comparing factor loadings from the Romanian sample with those from German (Pössel, 2009) [2] and Polish (Śliwerski, 2014) populations [14]. The German sample consisted of university students, faculty, and staff (N = 796, 80% female, M age = 23.71, SD = 6.57), while the Polish sample included a smaller, more heterogeneous group of adults (N = 86, 58.1% female, M age = 32.94, SD = 12.58), including students, and clinical participants. Despite differences in composition, both samples comprise adults assessed using the same self-report measures under standardized conditions, allowing for a meaningful comparison at the configural invariance level. Spearman correlations indicated a moderate similarity between Romanian and German samples (rho = 0.568, p <.001) and Romanian and Polish samples (rho = 0.700, p <.001). These findings provide initial support for configural invariance across these cultural contexts.

Construct validity was also assessed through Average Variance Extracted (AVE) and Composite Reliability (CR) indicators as shown in Table 6. The results confirmed adequate convergent validity and high internal consistency. Specifically, AVE values were above the recommended threshold of 0.50 for view of self (AVE = 0.533) and view of future (AVE = 0.665), while view of world showed a slightly lower AVE (0.418), suggesting limited convergent validity for this dimension. All three factors demonstrated high composite reliability: CR = 0.916 for view of self, CR = 0.947 for view of future, and CR = 0.866 for view of world, exceeding the minimum standard of 0.70 [36]. These results support the internal consistency and construct validity of the CTI subscales.

Table 6.

Average variance extracted (AVE) and composite reliability (CR)

Factor	Average Variance Extracted	Composite Reliability
View of self	0.533	0.916
View of future	0.665	0.947
View of world	0.418	0.866

Within view of world, standardized loadings ranged from 0.57 to 0.74, with CR = 0.866 and AVE = 0.418. While reliability was good, convergent validity remained below the 0.50 benchmark for this factor.

We did a Test-Retest Analysis to evaluate the results’ reliability or consistency over time. We administered CTI to a smaller group of participants (N = 62), at two different points in time (T0 and T1). Spearman correlations were performed: view of self (rho = 0.88, 95% CI [0.82 − 0.92], p <.001), view of world (rho = 0.81, 95%CI [0.69 − 0.89], p <.001), view of future (rho = 0.84, 95% CI [0.74 − 0.90], p <.001). The correlations indicated a strong consistency between T0 (initial measurement) and T1 (second measurement after three months) for all assessed variables.

Intraclass Correlation Coefficient (ICC) was also performed with a single score ICC(A,1) = 0.756, 95%CI [0.69 − 0.89], p <.001. The ICC of 0.756 confirmed good reliability, indicating the measurements were consistent across the two time points.

These results demonstrate that the variables: view of self, view of world, and view of future remain consistent over time, supporting the reliability of measurements. The strong correlations and the good ICC value validate the temporal stability of the instrument used in this study. This reliability ensures that the tool consistently measures the intended constructs and is suitable for longitudinal research.

Although the sample was not demographically representative and included a higher proportion of female participants (75%), we examined the measurement invariance of the instrument across gender and age groups (See table in Supplementary Material 3). These analyses, while exploratory due to sample imbalance, offered preliminary evidence regarding the stability of the instrument’s factorial structure across key demographic categories (Table 7).

Table 7.

Measurement invariance of the CTI across sex and age groups

Group	Model	χ²(df)	CFI	RMSEA	SRMR	ΔCFI	ΔRMSEA
Sex	Configural	731.84(694)	0.998	0.013	0.066	—	—
	Metric	799.36(719)	0.995	0.018	0.069	−0.003	0.006
	Scalar	834.54(744)	0.995	0.019	0.07	−0.001	0.001
	Strict	855.35(772)	0.995	0.018	0.071	< 0.001	−0.001
Age	Configural	684.18(694)	1	< 0.001	0.066	—	—
	Metric	761.24(719)	0.997	0.013	0.07	−0.003	0.013
	Scalar	827.76(744)	0.995	0.018	0.071	−0.002	0.005
	Strict	861.77(772)	0.995	0.019	0.075	< 0.001	< 0.00

ΔCFI and ΔRMSEA are computed relative to the previous model

Criteria for invariance: ΔCFI ≤ .010 and ΔRMSEA ≤ .015

Measurement invariance was tested across gender and age groups using multi-group CFA with the DWLS estimator (Table 7). For gender (N₁ = 509 women; N₂ = 164 men), the configural model demonstrated good fit (CFI = 0.998, RMSEA = 0.013, SRMR = 0.066). Constraining factor loadings (metric), intercepts (scalar), and residuals (strict) resulted in minimal changes in fit (all ΔCFI ≤ 0.003; ΔRMSEA ≤ 0.006; see Table 7), indicating full measurement invariance across gender.

For age, participants were classified as young adults (20–40 years; N = 321) and middle-aged adults (41–65 years; N = 350) following the lifespan development framework proposed by Papalia and Martorell [37]. Four participants outside this range were excluded. The configural model showed excellent fit (CFI = 1.000, RMSEA = 0.000, SRMR = 0.066). Adding equality constraints on loadings (metric), intercepts (scalar), and residuals (strict) produced only small changes (all ΔCFI ≤ 0.003; ΔRMSEA ≤ 0.013; see Table 7). These results supported full measurement invariance across age groups as well.

Convergent validity was tested, providing evidence of its accuracy in capturing the concept. Strong correlations between CTI factors and theoretically related psychological constructs of Ryff Carol’s (1989) Psychological Well Being Scale [28] support the convergent validity. The Scale of Psychological Well–Being (SPWB) with 42 items was used to assess six constructs: self-acceptance, autonomy, personal development, environmental control, positive relationships with others, and purpose in life. In the present sample, the six factors’ reliability was very good, showing good fidelity: McDonald’s ω = 0.68, 95% CI [0.64 − 0.71] for personal growth, McDonald’s ω = 0.69, 95% CI [0.65 − 0.72] for autonomy, McDonald’s ω = 0.83, 95% CI [0.80 − 0.85] for self-acceptance, McDonald’s ω = 0.80, 95% CI [0.79 − 0.83] for positive relationships with others, McDonald’s ω = 0.81, 95% CI [0.79 − 0.84] for environmental mastery, and McDonald’s ω = 0.76, 95% CI [0.73 − 0.79] for purpose in life.

View of self correlated with self-acceptance rho = 0.69, p <.001, 95% CI [0.64 − 0.73] and personal growth rho = 0.50, p <.001, 95% CI [0.44 − 0.55]. The good associations highlight that self-perception aligns with self-worth and personal acceptance. Individuals with a positive view of themselves have higher levels of self-acceptance. Conversely, people with a negative view of themselves are likely to have lower self-acceptance. A positive view of self is moderately associated with greater personal growth.

View of future correlated with environmental mastery rho = 0.67, p <.001, 95% CI [0.62 − 0.71] and purpose in life rho = 0.54, p <.001, 95% CI [0.47 − 0.58]. The associations indicate that a positive outlook on the future is associated with positively seeing the environment; individuals perceive themselves as being in control of and capable of effectively managing their environment and life circumstances. On the contrary, a negative view of future is related to lower environmental mastery and a hostile, negative environment. People with a positive view of future tend to have a stronger sense of purpose or meaning in life.

View of world correlated with environmental mastery rho = 0.55, p <.001, 95% CI [0.49 − 0.61] and positive relations rho = 0.49, p <.001, 95% CI [0.43 − 0.55]. People with a positive, favorable view of world can manage their environment and life circumstances while people with a negative view of world are less likely to experience environmental mastery. The good association indicates that positively perceiving the world fosters healthy interpersonal relationships. Discriminant validity was supported by the lower correlations between CTI dimensions and the theoretically less-related construct of autonomy. While the CTI subscales showed strong correlations with their respective psychological well-being components (e.g., self-acceptance, environmental mastery), their correlations with autonomy were consistently lower: view of self (rho = 0.40, 95% CI [0.34–0.46]), view of world (rho = 0.34, 95% CI [0.27–0.41]), and view of future (rho = 0.33, 95% CI [0.26–0.40]). These weaker associations suggest that each CTI factor captures specific cognitive patterns distinct from general autonomy, thus supporting the scale’s discriminant validity.

Meaningful relationships with external indicators of psychological well-being supported validity. The results confirmed the validity of the CTI Scale across its dimensions.

Results affirm the CTI structural integrity and construct validity. This model validates that the instrument appropriately encompasses the three fundamental aspects of cognitive perceptions: self, future, and world.

Discussions

The results of this study confirm the validity and reliability of CTI on a Romanian sample. The reliability of the three components of the scale was excellent, comparable to that of the American sample. The convergent validity tests revealed that the translated items accurately captured the intended concepts. Reliability ensures minimal error in measurements, which enhances the test’s statistical power and the likelihood of confirming research hypotheses [38]. Additional investigation is necessary to enhance the reliability of supporting evidence for concept and criterion validity.

By analyzing the construct through CFA, we found the instrument maintains its robust psychometric properties and configural invariance across a culturally distinct population. Specifically, the three-factor structure was replicated with high-reliability coefficients and significant factor loadings. The high standardized loadings across the three subscales confirm that the items effectively measure their respective constructs. Loadings ranged from moderate to high across all three dimensions, indicating strong relationships between the latent variables and their indicators.

The model fit indices provide strong evidence for the factorial structure of the CTI Romanian version. The goodness-of-fit indices (CFI, TLI, RMSEA, SRMR) meet the thresholds commonly recommended in the literature for a well-fitting model [39]. This supports the applicability of the cognitive triad theory, as originally conceptualized by Beck, in Romanian cultural contexts. The replication of the triadic structure suggests that CTI captures positive and negative cognitive patterns predictive of mental health outcomes such as depression, well-being, and resilience.

We achieved the goal of getting a Romanian version of CTI. The validation of this instrument in Romania holds significant implications for both clinical and research settings. Therapists can more accurately identify maladaptive thinking or positive cognitive patterns, which can inform resilience-building interventions. Given the correlation between positive cognitive schemas and well-being, future interventions could focus on reducing negative thoughts and fostering positive cognitions about the self, the world, and the future. The findings suggest that cognitive patterns associated with depression and well-being, as measured by CTI, are applicable across cultures, including in Romania.

Future directions

Employing this instrument in varied settings will gradually accumulate evidence regarding its strengths and weaknesses. Cross-cultural comparisons may reveal important differences in cognitive patterns. More research is required to enhance the quantity of supporting evidence for construct validity. Future studies should also examine the predictive validity of the Romanian CTI version in longitudinal research. Additionally, research could explore the effectiveness of cognitive restructuring interventions based on the CTI results, assessing how shifts in cognitive patterns about the self, the world, and the future impact the reduction of depressive symptoms and the enhancement of well-being. Alternative models or extensions, such as multi-group CFA, could further validate the scale’s robustness across different subgroups within the Romanian population.

Limitations

The limitations of this study must be acknowledged. The instrument adaptation could not be done on a calibrated normative sample, because financial and time resources did not allow the researcher access to a nationally representative sample. The researched batch is sufficient to perform psychometric analyses; the lack of norms does not prevent the use of this instrument for research purposes. Subjectivity can compromise the reliability of the gathered data: responses may be influenced by the situations of the participants at the time of their response. While this model provided good fit indices, alternative models or extensions, such as multi-group confirmatory factor analysis, could further validate the scale’s robustness across different subgroups within the Romanian population.

Although view of world achieved good composite reliability (CR = 0.866), its AVE = 0.418 indicates limited convergent validity. Future work should revisit the lowest-loading items (reported in the supplement) for potential rewording or removal, using cross-validation to avoid capitalization on chance.

Conclusion

This study successfully adapted CTI for use in the Romanian-speaking population, demonstrating its reliability and validity in assessing cognitive patterns related to mental health. CTI continues to be a valuable tool for clinical and research purposes, enabling practitioners to assess both negative and positive cognitive schemas. Its use can support early identification of maladaptive thinking patterns and guide targeted psychological interventions.

Abbreviations

CFA

Confirmatory Factor Analysis

CFI

Comparative Normed Fixed Index

CMIN

Chi-square Minimum Discrepancy

CTI

Cognitive Triad Inventory

df

degrees of freedom

DWLS

The diagonally weighted least squares

ICC

Intraclass Correlation Coefficient

M

mean

ML

maximum likelihood

MLR

Maximum Likelihood Estimation with Robust Standard Errors

RMSEA

mean square error of approximation

SD

standard deviation

SRMR

standardized root mean square residual

TLI

Tucker – Lewis Index

Authors’ contributions

A. M. P.: Conceptualization, Design, Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Projectadministration, Methodology, Investigation, Formal analysis, Data curation.A. M.: Conceptualization, Design, Writing – review & editing, Visualization, Validation, Software, Methodology, Formal analysis, Data curation.V. Ș. R.: Design, Visualization, Supervision, Project administration.All authors read and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and publication of this article, including no funds, grants, or other support.

Data availability

The datasets analyzed during the current study are available from the corresponding author upon request. All relevant data supporting the conclusions of this study have been included within the article and its supplementary information files. Additional datasets, including raw data, are available from the corresponding author upon reasonable request. The data, methods used in the analysis, code, and materials used to conduct the research will be made available to any researcher for purposes of reproducing the results or replicating the procedure. The data are not publicly available because they contain sensitive information that could compromise the privacy of research participants. Data and Code availabilityThe code used to generate results reported in this paper and other materials can be accessed in the OSF repository or it is available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

Ethical clearance was received from the Ethics Committee of the University of Bucharest, Bucharest, Romania (registration number 59/12.04.2024).

The study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki and its subsequent amendments.

Informed consent was obtained from all individual participants included in the study. No identifying information (identifying images or other personal or clinical details) about the participants is included in this article.

Consent for publication

Not applicable. The manuscript does not contain individual person’s data in any form.

Conflicting interests

The authors declare no competing interests.