Psychometric Properties of the Five Facet Mindfulness Questionnaire Short Version (FFMQ-SF) in the Colombian Population

Title

Psychometric Properties of the Five Facet Mindfulness Questionnaire Short Version (FFMQ-SF) in the Colombian Population

Complete Reference

Castañeda, D. F., Bianchi, J. M., Villalba-Garzón, J. A., & Ruiz, F. J. (2025). Psychometric Properties of the Five Facet Mindfulness Questionnaire Short Version (FFMQ-SF) in the Colombian Population. Mindfulness, 16, 918–932. https://doi.org/10.1007/s12671-025-02533-8

Accepted: January 30, 2025 | Published online: February 19, 2025

Study Type

Cross-sectional psychometric instrument validation study. Confirmatory factor analysis of the FFMQ-SF-24 structure in a Colombian sample, with examination of measurement invariance across sex and meditation practice, and analyses based on the Rasch Model.

---

CONTEXT AND OBJECTIVES

Mindfulness is conceptualized as awareness focused on the present moment, characterized by openness, receptivity, curiosity, and acceptance of ongoing experiences (Kabat-Zinn, 2003; Raul & Williams, 2016). In clinical settings, mindfulness is the core of mindfulness-based interventions, including Mindfulness-Based Stress Reduction (MBSR), Mindfulness-Based Cognitive Therapy (MBCT), Acceptance and Commitment Therapy (ACT), Behavioral Activation Therapy, and Mode Deactivation Therapy (Baer et al., 2006; Hayes et al., 1999; Jennings & Apsche, 2014).

The Five Facet Mindfulness Questionnaire (FFMQ) is one of the most widely used instruments for assessing mindfulness. Developed by Baer et al. (2006), it was designed on the basis of previously developed mindfulness scales: MAAS (Mindful Attention Awareness Scale), KIMS (Kentucky Inventory of Mindfulness Skills), CAMS (Cognitive and Affective Mindfulness Scale), and FMI (Freiburg Mindfulness Inventory). The FFMQ identified five facets: (a) Observing (attending to inner and outer experiences), (b) Describing (labeling experiences with words), (c) Acting with awareness (performing tasks with attention), (d) Non-judging of experience (adopting a non-evaluative stance), and (e) Non-reactivity to inner experience (allowing thoughts to flow without engaging with them).

The original FFMQ consists of 39 items with adequate to good internal consistency (α values ranging from 0.75 to 0.91). Bohlmeijer et al. (2011) developed the 24-item version (FFMQ-SF-24) by selecting items with better corrected item-total correlations and standardized factor loadings, reducing questionnaire length without compromising psychometric properties.

Previous FFMQ research has examined psychometric properties across diverse populations and languages: Spain (Cebolla et al., 2012); France (Heeren et al., 2011); China (Hou et al., 2014); Norway (Dundas et al., 2013); and Chile (Schmidt & Vinet, 2015). Sosa (2019) validated the FFMQ in Colombian samples, reporting a five-factor structure with reliability values ranging from 0.79 to 0.87. However, most research on the FFMQ-SF-24 has been limited, particularly regarding factorial invariance related to sex and meditation practice.

The present study aimed to examine the psychometric properties of the FFMQ-SF-24 in a Colombian sample of 582 adults, including analyses of factor structure, reliability, measurement invariance across sex and meditation practice, and relationships with related constructs (depression, anxiety, and stress symptoms; psychological inflexibility; environmental reward; and behavioral activation).

---

METHOD

Participants

Participants were 582 Colombian adults (age: 18–87 years; M = 34.69, SD = 14.51). The sample was predominantly female (74.3%, n = 432), single (68.0%, n = 396), and composed of university students (53.2%, n = 310). Participants were recruited from five regions of the country, primarily from Bogotá, D.C. (333, 57.22%). Regarding mental health, 241 participants (41.41%) had attended psychological or psychiatric counseling services at some point in their lives; 72 participants (12.37%) had received services in the previous month. Twenty-seven participants were excluded because they reported meditation practice of less than 6 months.

Instruments Assessed

Five Facet Mindfulness Questionnaire - Short Form (FFMQ-SF-24). The FFMQ was developed by Baer et al. (2006), and the Spanish version adapted by Sosa (2019) was used. The FFMQ is a 39-item questionnaire, although this study used the 24-item version. Items are responded to on a 5-point scale (1 = never or very rarely true; 5 = very often or always true) and assess five facets of mindfulness: Observing (attending to inner and outer experiences), Describing (labeling with words), Acting with awareness (performing tasks with focus), Non-judging of experience (non-evaluativity), and Non-reactivity to inner experience (allowing thoughts to flow). The Colombian version (Sosa, 2019) demonstrated adequate internal consistency (α = 0.86) and significant correlations with other variables. The FFMQ-SF-24 (Bohlmeijer et al., 2011) consists of 24 items selected for better corrected item-total correlations and standardized factor loadings, maintaining psychometric properties similar to the long version. In the present study, the FFMQ demonstrated Cronbach's alpha = 0.91 and McDonald's omega = 0.91, with a mean score of 128.46 (SD = 20.73).

Other Measurement Instruments

Mindful Attention Awareness Scale (MAAS). Developed by Brown and Ryan (2003), evaluated using the Spanish version by Ruiz et al. (2016b). The MAAS assesses the degree to which individuals pay attention during routine tasks using 15 items on a 6-point Likert scale (0 = almost never; 5 = almost always). Higher scores indicate greater levels of mindfulness. The Spanish version demonstrated adequate internal consistency (α = 0.92) in Colombian samples, as well as convergent and discriminant validity. In the present study, the MAAS demonstrated alpha = 0.93, omega = 0.93, with mean score of 64.12 (SD = 15.31).

Acceptance and Action Questionnaire – II (AAQ-II). Developed by Bond et al. (2011), evaluated using the Spanish version by Ruiz et al. (2013). The AAQ-II is a measure of psychological inflexibility consisting of 7 items responded to on a 7-point Likert scale (1 = never true; 7 = always true). Higher scores indicate greater psychological inflexibility. The Spanish version demonstrated adequate internal consistency (α = 0.91) in Colombian samples, as well as convergent and discriminant validity. In the present study, the AAQ-II demonstrated alpha = 0.91, omega = 0.95, with mean score of 21.98 (SD = 11.04).

Depression Anxiety Stress Scales – 21 (DASS-21). Developed by Lovibond and Lovibond (1995), evaluated using the Spanish version by Daza et al. (2002). The DASS-21 was designed to evaluate emotional symptoms of depression, anxiety, and stress using 21 items responded to on a 4-point Likert scale (0 = does not apply to me at all; 3 = applies to me very much or most of the time). The Spanish version demonstrated excellent internal consistency in Colombian samples (α = 0.96) and a hierarchical factor structure consisting of three first-order factors and a second-order factor. In the present study, the DASS-21 demonstrated alpha = 0.96, omega = 0.96, with mean score of 21.40 (SD = 14.97).

Behavioral Activation for Depression Scale – Short Form (BADS-SF). Developed by Manos et al. (2011), evaluated using the Spanish version by García (2019). The BADS-SF consists of 9 items responded to on a 7-point Likert scale (0 = not at all; 6 = completely) and measures two dimensions related to behavioral activation: Avoidance and Activation. The Spanish version demonstrated adequate internal consistency (α = 0.82) and convergent validity. In the present study, the BADS-SF demonstrated alpha = 0.82, omega = 0.81, with mean score of 32.75 (SD = 10.00).

Reward Probability Index (RPI). Developed by Carvalho et al. (2011), evaluated using the Spanish version by Reyes-Buitrago et al. (2023). The RPI is a self-report instrument designed to evaluate the magnitude of environmental reward as an approximation of contingent positive reinforcement response (RCPR) through two factors: Reward Probability and Environmental Suppressors. It consists of 20 items responded to on a 4-point Likert scale. The Spanish version demonstrated adequate internal consistency and convergent and discriminant validity. In the present study, the RPI demonstrated alpha = [verify in original article], with mean score of 56.38 (SD = 8.61).

Meditation Practice

The sociodemographic form included questions about experience with meditation practice. Participants were asked whether they engaged in some form of meditation practice. If they responded "yes," they were asked how long they had been practicing (response options: 1 month, 6 months, 1 year, 3 years, or more). Of the total sample, 22.6% (n = 132) reported engaging in some form of meditation practice: 24.2% were males and 75.70% were females; 21.2% reported formal practice for 6 months to 1 year, 41.7% between 1 and 3 years, and 37.1% for more than 3 years. Of these, 54.1% had attended mental health services at some point in their lives. Seventy-eight percent reported being a university or postgraduate student, 73.0% were single, followed by 15.0% in free union. Twenty-seven participants were excluded because they reported meditation practice less than 6 months.

Procedure

Permission to use the instrument was obtained via email from the corresponding author of the FFMQ-SF-24. Participants completed an anonymous online survey (Microsoft Forms) disseminated through social media networks (e.g., Facebook, WhatsApp, and Instagram), including meditation groups. A snowball sampling procedure was used in which researchers asked contacts and participants to share the survey post to reach more potential participants. The data collection period was from September to December 2021. Before accessing the questionnaire package, participants provided informed consent by accepting the conditions explained at the beginning of the survey, including that they were adults. The informed consent emphasized that participation was anonymous and that participants could stop participating at any time they wished. The median time to complete the survey was approximately 25 minutes. Participants were not compensated for their participation.

Data Analyses

A descriptive analysis of sociodemographic data and scale scores was performed to identify the type of distribution, frequencies, means, standard deviations, and ranges. Following the guidelines of the Standards for Psychological and Educational Testing by the American Educational Research Association, American Psychological Association, and National Council on Measurement in Education (AERA, APA & NCME, 2014), evidence was based on internal structure collected through confirmatory factor analysis (CFA).

CFA models were specified by comparing three models using the Diagonal Weighted Least Squares (DWLS) estimator: (a) a one-factor model, (b) a five-factor model, and (c) a five-factor model with a second-order factor. The following goodness-of-fit indices were compared: (a) the comparative fit index (CFI), (b) the Tucker-Lewis index (TLI), (c) the normed fit index (NFI), (d) the incremental fit index (IFI), (e) goodness of fit index (GFI), (f) root mean square error of approximation (RMSEA), and (g) the parsimonious normed fit index (PNFI). Models were evaluated based on χ²/df < 4; CFI ≥ 0.95; TLI, NFI, and IFI > 0.90; and RMSEA < 0.05. Additionally, higher PNFI values indicate more parsimonious fit (Riol et al., 2006).

Multigroup confirmatory factor analyses were conducted to examine configural, metric, scalar, and residual factorial invariance regarding sex and meditation practice condition. The analysis was progressive and sequential, beginning with a configural invariance model and then imposing restrictions until a model was rejected or the residual model was achieved. To reject models, criteria suggested by Lipke et al. (2007) (a p-value < 0.05) were used along with the criterion ΔCFI ≥ −0.010 and ΔARMESEA > 0.015 (Chen, 2007). For DIF detection, significance indices of ≥ 0.05 and Nagelkerke's pseudo-R² ≥ 0.010 between compared groups were used (Choi et al., 2011).

Internal consistency of instruments was estimated with Cronbach's alpha (α) and McDonald's omega (ω) coefficients. Values greater than 0.70 were considered adequate for both (George & Mallery, 2003; Ventura-León & Caycho-Rodríguez, 2017). Additionally, the discrimination index was calculated by correlating items with total scores. Item Response Theory (IRT) analyses were complemented with estimates of the Rasch Model of Item Response Theory, with INFIT and OUTFIT fit indices, and Test Information Functions (TIF).

To analyze relationships between the FFMQ-SF-24 and other variables, correlations between total and subfacet scores with other measures were estimated using Spearman's correlation coefficient (rho), where values < 0.10 were considered negligible, 0.10 to 0.30 weak correlations, > 0.30 moderate correlations, and > 0.50 strong correlations (Cohen, 2009). Student's t-test was used to identify differences between meditators and non-meditators, allowing data reporting even if not normally distributed, considering homogeneity of variances (Goss-Sampson, 2019).

Data analyses were performed with the ULLR Toolbox add-on version 4.0.2 for Windows RStudio (Hernández, 2017). The lavaan libraries (Version 0.6–7; Rosseel, 2012), semTools (Version 0.5–5; Jorgensen et al., 2019), semPlot (Version 1.1.2; Epskamp, 2015), eRm (Version 1.0–6; Mair & Hatzinger, 2007), and lordif (Version 0.3–3; Choi et al., 2011) were used.

---

RESULTS

Meditation Practice

Of the total sample, 22.6% (n = 132) reported engaging in some form of meditation practice (24.2% males and 75.70% females). Of these, 21.2% reported formal practice of 6 months to 1 year, 41.7% between 1 and 3 years, and 37.1% for more than 3 years. Of these, 54.1% had attended mental health services at some point in their lives. Seventy-eight percent reported being a university or postgraduate student, 73.0% were single, followed by 15.0% in free union.

Evidence of Validity Based on Internal Structure

Confirmatory Factor Analyses

Three models were analyzed: a one-factor model, the original five-factor model, and a five-factor model with a second-order factor. The five-factor model showed good goodness of fit: χ²/df = 2.302, RMSEA = 0.047 (90% CI [0.042, 0.053]), TLI = 0.962, CFI = 0.967, NFI = 0.943, PNFI = 0.826, IFI = 0.967, GFI = 0.971. The five-factor model with a second-order factor also showed good goodness of fit: χ²/df = 3.024, RMSEA = 0.059 (90% CI [0.054, 0.064]), TLI = 0.940, CFI = 0.946, NFI = 0.922, PNFI = 0.826, IFI = 0.947, GFI = 0.962.

All items presented factor loadings greater than 0.60, except for two items (9 and 10), whose loadings were less than 0.32. The highest correlation was obtained between the Non-reactivity to Inner Experience and Acting with Awareness factors, with a value of 0.64. The lowest correlation was obtained between Observing and Non-judging of experience, with a value of 0.05. The factor most correlated with others was Describing with a mean correlation of 0.49, while Observing had the lowest mean correlation with a mean correlation of 0.39.

Measurement Invariance

Factorial invariance analyses of the five-factor model regarding sex and meditation practice variables showed no statistically significant differences between males and females or regarding whether meditation is practiced regularly. No differences were found in the basic model configuration and factor weights (configural invariance and metric invariance), nor in their intercepts and residuals (scalar invariance and residual invariance). The results of the invariance analyses (Table 2) indicated that the differences found in comparisons were not attributable to differences in measurement precision (Putnik & Bornstein, 2016).

Differential Item Functioning (DIF) Detection

Items 10 (p = 0.00, 0.00, and 0.19 with Nagelkerke's R² = 0.01, 0.01, and 0.00) and 13 (p = 0.00, 0.00, and 0.03 with Nagelkerke's R² = 0.01, 0.01, and 0.00) presented differential functioning on the meditation practice variable with p = 0.42, 0.00, and 0.00 with Nagelkerke's R² = 0.00, 0.01, and 0.01. Item 10 presented uniform DIF with uniform probabilities for male scores across all four categories. In item 13, DIF was not uniform because females obtained lower probabilities in low categories (1 and 2) and males in high categories (3 and 4). This item had the same type of DIF with individuals who do not practice meditation because they obtained lower probabilities in categories (1, 2, and 3) but higher probabilities in category 4.

Reliability

Table 4 shows item-factor correlations, which ranged from 0.50 to 0.57 for Observing, 0.52 to 0.67 for Describing, 0.66 to 0.76 for Acting with Awareness, 0.43 to 0.70 for Non-judging of Experience, and 0.32 to 0.53 for Non-reactivity to Inner Experience. Thus, all items had item-factor correlations above 0.30 (Frías, 2021), which aligns with expectations. Alpha and omega values ranged from 0.68 (Non-reactivity to Inner Experience) to 0.87 (Acting with Awareness).

Relationships with Mindfulness, Present Moment Awareness, Emotional Symptoms, Psychological Inflexibility, Environmental Reward, and Activation

Table 5 shows that all correlations of the FFMQ-SF-24 with other variables were statistically significant (p < 0.001), except for the Observing facet. Table 5 indicates that the FFMQ-SF-24 correlated strongly with the original FFMQ. Likewise, the FFMQ-SF-24 facets showed medium and strong correlations with MAAS scores. The FFMQ-SF-24 showed medium and strong correlations with MAAS scores. The FFMQ-SF-24 correlated significantly and theoretically coherently negative with emotional symptoms and psychological inflexibility. Finally, the FFMQ-SF-24 showed theoretically coherent negative correlations with emotional and psychological symptoms and psychological inflexibility.

Relationships of FFMQ-SF-24 Scores with Meditation Practice

Independent samples Student's t-test was conducted after ensuring that test requirements were met, although variances were not normally distributed in all facets. Table 3 shows that the Observing facet (t(580) = 2.25, p = 0.023; d = 0.23) and Non-reactivity to Inner Experience (t(580) = 2.28, p = 0.02; d = 0.23) presented statistically significant differences in meditators and non-meditators groups with small effect sizes.

Psychometric Analyses Based on the Rasch Model

Compliance with the unidimensionality assumption required by the model was guaranteed when analyses were performed by subfacet. Parameters and goodness-of-fit indices were obtained because the Rasch model did not converge. According to the criteria of Hodge and Morgan (2017), the model did not fit items 1, 14, 19, and 22. For item 21, there was only a misfit in the OUTFIT, indicating that the model failed to predict person response with measures θ. Item b parameters were distributed in the following ranges: 7.64 logits (Describing, 13.42; Acting with Awareness), 9.90 logits (Non-judging of Experience, 13.42; Non-reactivity to Inner Experience, 6.53). For person θ parameters, ranges were 6.05 logits, 8.34, 6.48, and 6.23, respectively. When comparing the ranges of logit measures for items and persons, 8.42% of participants had b parameters below the FFMQ's b parameters, and 13.40% had outlier measures. Table 6 shows the averages of the b parameters obtained from the four categories, the average standard error of the parameters, and the Test Information Functions' OUTFIT indicating the precision of the fit indices.

---

DISCUSSION AND CONCLUSIONS

The present study examined the psychometric properties of the FFMQ-SF-24 in a large Colombian sample (N = 582). Specifically, the scale's reliability, internal structure, invariance analyses, and relationships with related constructs were analyzed.

Main Findings

The five-factor structure was confirmed with good fit. All items obtained adequate factor loadings on their dimensions (ranging from 0.305 to 0.917). Goodness-of-fit indices were excellent and comparable to those in Bohlmeijer et al. (2011) and Onate and Calvete (2018). The five-factor model with a second-order factor presented good fit indices, but the simple five-factor model showed better fit.

The FFMQ-SF-24 was invariant across sex and meditation practice. This is a relevant finding because it means the FFMQ-SF-24 assesses the same latent construct across groups and adds evidence to what has been reported in the literature. The five-factor model was found to be longitudinally invariant in studies with undergraduate students in Argentina (Correa et al., 2023) and US population samples (Haas & Akamatus, 2019). Other studies have also found evidence of measurement invariance across different FFMQ versions (Fong et al., 2021; Gu et al., 2016; Hu et al., 2024; Iqbal et al., 2023; Raphiphathana & Jose, 2020).

Some levels of DIF were found in all subfacets except Acting with Awareness. However, CFA modification indices analysis did not present evidence of DIF for any items in the subfacets. In summary, it appears necessary to continue investigating to determine whether this difference in functioning can really be explained by biases that may affect the fairness of measurement. The evidence collected in this study adds to that of the previous study by Fernández Curado et al. (2022), which detected.

Internal Structure and Reliability

In evidence of validity based on internal structure, the five-factor model was confirmed. All items obtained adequate factor loadings on their dimensions (ranging from 0.305 to 0.917). Goodness-of-fit indices were excellent and comparable to those in the original study (Bohlmeijer et al., 2011) and Onate and Calvete (2018). The five-factor model with a second-order factor presented good fit indices, but the simple five-factor model showed better fit.

Relationships with Related Constructs

Significant relationships were found between the FFMQ-SF-24 and the DASS-21 (emotional symptoms), its dimensions (depression, anxiety, and stress), the AAQ-II (psychological inflexibility), environmental avoidance, environmental suppressors, and clinical symptomatology activation, which is coherent with previous research (Bohlmeijer et al., 2011; Onate & Calvete, 2018; Vechof et al., 2011).

Correlations found between the level of attention paid when performing tasks (MAAS), the probability of reward (RPI), the level of activation (BADS-SF), psychological inflexibility (AAQ-II), emotional symptomatology (DASS-21) were of expected magnitude and direction.

Meditation Practice and Group Differences

Meditators presented significantly higher scores compared to the non-meditators group in Observing and Non-reactivity to Inner Experience, with small effect sizes. This result may be evidence of the discriminant validity of the FFMQ-SF-24.

Clinical and Measurement Considerations

Results from Rasch Model analysis revealed that some items in the Non-judging of Experience (14 and 19), Describing (1), and Acting with Awareness (22) dimensions did not fit the model, suggesting potential adjustments. However, these misfits should be interpreted with caution, considering that CFA modification indices analyses found no evidence of DIF for any items in the subfacets.

Significance and Contribution

This study makes a significant contribution to psychometric research on mindfulness by providing the first comprehensive validation of the FFMQ-SF-24 in Colombian populations. The findings demonstrate that the instrument possesses robust psychometric properties, with a confirmed five-facet factor structure and excellent fit indices. The demonstration of measurement invariance across gender and meditation practice is particularly important, validating that the instrument measures the mindfulness construct equivalently across different population subgroups. This study contributes to the field of mindfulness measurement in Latin America by providing a valid and reliable tool for assessing present-moment awareness and its differentiated facets in clinical, research, and well-being contexts within Spanish-speaking Colombian populations.

Limitations and Future Research

Some limitations of this study are worth mentioning. First, the FFMQ-SF-24 was evaluated in a general adult population sample, with a small subgroup of meditators. Therefore, it is recommended that future studies analyze the psychometric properties of the FFMQ-SF-24 in a larger and more homogeneous sample in terms of representativeness by gender, regions of the country, and meditation practice. Second, no systematic information was obtained regarding participants' clinical diagnoses, which prevented comparison of the clinical relevance of the instrument. Third, all variables were measured through self-reports, and for this type of study, direct measurement of a variable in the present study is recommended. Fourth, a convenience sample was recruited, which does not allow generalizing results to the general Colombian population. Therefore, results obtained in this study should be considered cautiously as they could be a function of sampling greater than the actual psychometric properties of the FFMQ-SF-24 in the Colombian population. In this regard, this sample was composed of a considerably higher percentage of females than males. Future studies with additional Colombian samples are suggested to confirm the results. Finally, it is necessary to explore the sensitivity of the FFMQ-SF-24 to treatment within scale and analyze whether the instrument's scores are sensitive to the effect of psychological interventions aimed at increasing mindfulness.

---