Psychometric Properties of a Scale to Measure Family Inclusion in Mexican Adults

María Montero-López Lena ^1,*, Livia Sánchez Carrasco ¹ , Montserrat Celdrán Castro ²

1. Faculty of Psychology, National Autonomous University of Mexico, Mexico

2. Faculty of Psychology, University of Barcelona, Spain

* Correspondence: María Montero-López Lena

Academic Editor: Mats Niklasson

Special Issue: Psycho-social Wellbeing in Aging

Received: June 20, 2025 | Accepted: December 07, 2025 | Published: December 17, 2025

OBM Geriatrics 2025, Volume 9, Issue 4, doi:10.21926/obm.geriatr.2504332

Recommended citation: Montero López Lena M, Sánchez Carrasco L, Celdrán Castro M. Psychometric Properties of a Scale to Measure Family Inclusion in Mexican Adults. OBM Geriatrics 2025; 9(4): 332; doi:10.21926/obm.geriatr.2504332.

© 2025 by the authors. This is an open access article distributed under the conditions of the Creative Commons by Attribution License, which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original work is correctly cited.

1. Introduction

The literature has widely documented the importance of the emotional bond [1,2,3] as a critical variable in promoting mental health throughout life. According to Bowlby [2], the quality of the affective bond that is created in childhood is a precursor of the type of interpersonal relationship that is established in later stages of human development. In this regard, several authors have developed questionnaires to assess the specific characteristics of Bowlby's attachment styles [4,5,6,7,8], which persist throughout life. However, efforts to measure emotional components of attachment styles have mainly been documented in children or young adults, with few explicitly described for the older population, and research in this regard carried out with Latin samples is particularly scarce [9]. This may be due to the social bias that, as people age, their emotional connections deteriorate as older individuals experience social isolation [10]. While these questionnaires document essential aspects of the emotional bond, they present two characteristics that make them difficult to apply to older adults. On the one hand, they have been generated and tested on samples of young adults, mostly university students from European, American, or Asian cultures, which may undermine the conception of emotional interaction as ageing. On the other hand, the extension of these makes it difficult to apply them quickly and parsimoniously, with consequent impact on participants' responses, especially in samples of older adults.

For humans, emotional bonds are a critical factor in mental health [11,12]. The ability to build meaningful, stable, and lasting relationships becomes crucial as one passes through the different stages of life and presents special challenges with ageing. In particular, the quality and intensity of the affective ties that older people form within their circle of close interaction, as family and friends, are essential to preserving both mental health and quality of life [13]. Feasible evidence of the emotional bond is the perception of family inclusion that older people express. Suppose the older person perceives that his/her presence is identified, considered, and valued by the family members relevant to him/her; family inclusion becomes a crucial behavioral pattern that accounts for the effectiveness of said emotional bond. Since the emotional component is essential to mental health, it is strategic to use valid and reliable measures that assess it. In this regard, significant efforts have been made to measure specific aspects of the emotional bond in older individuals. In the literature, only one Latin American study [14] has documented the psychometric properties of the Late Adulthood Attachment Scale (LAAS), whose purpose was to evaluate safe attachment in older adults and identified two factors: (1) Fearful Avoidance and (2) Safe Commitment (Secure Engagement). Although this scale documented the reliability of both factors after three months, it seems that the affective component was evaluated generally, overlooking the peculiarities of family interactions. It is precisely in the scale of family inclusion (FI) documented in this article that these interactions are incorporated.

The concept of family inclusion was derived from previous work [15], which explored how family dynamics facilitated or hindered the manifestation of socio-affective and functional behaviors in children. Subsequently, Montero [16] adapted this questionnaire for older adults by replacing the notion of functionality with family inclusion, as this was the emotional component on which emphasis was placed. The concept of family inclusion is considered vital because it is assumed that the adaptive success in psychological terms that the person achieves throughout his or her life will depend, in part, on the efficiency with which the affective interactions characteristic of the different stages of life are consolidated, and on the significance attributed to the quality of such interactions. Thus, the efficiency and quality of the emotional bond were considered in the family inclusion questionnaire, which was empirically tested.

Based on the evidence presented on potential cultural biases and the cognitive processing difficulties that older individuals may encounter when asked to complete self-administered questionnaires with scalable options, a questionnaire was specifically designed to address these two possible contingencies. In agreement, this study aimed to document the psychometric properties of a scale to measure family inclusion in adults, which was relevant to Mexican samples and parsimonious in its application.

2. Methods

2.1 Participants

University students participated, and they invited their maternal or paternal grandparents to answer the questionnaire as well. The questionnaire was implemented face-to-face and digitally. Three hundred ninety-six responses were collected using the instrument, forming the initial database. After pre-processing, we generated a database with complete records.

2.2 Instruments

The questionnaire under evaluation was part of a broader project [17] aimed at documenting the effects of loneliness perception on mental health in adults. It was composed of a sociodemographic section and five scales: Depression, Family Inclusion, Loneliness, Loneliness Coping Behaviors, and Stress Perception. Only the scales analyzed in this paper are described below.

Sociodemographic section, such as age, gender, educational level, or civil status.

2.2.1 Family Inclusion Scale (FIS)

It was defined as the perception that adults have of being valued, recognized, and considered essential members of the family group. This scale has ten Likert-type scalar items with four response options, whose psychometric characteristics are documented in the present study. For the full version of the FIS questionnaire, the first author could be contacted to provide a copy.

2.2.2 Loneliness (S)

It evaluated a potentially stressful experience resulting from actual or subjective emotional deficiencies [18]. Montero and Rivera [19] documented the psychometric properties of this scale, consisting of 20 Likert-type items with five response options, in which a higher score indicates a higher occurrence of the lonely experience. This questionnaire has two sub-scales: (1) Personal-PD dissatisfaction with nine items (alpha = 0.89) and Familial Alienation-FA with 11 items (alpha = 0.93). FA was defined as the absence of affection within the nuclear family, while PD was defined as the perception of dissatisfaction and lack of purpose in life.

2.2.3 Perception of Stress (PS)

This scale evaluates reactions to emotionally stressful situations. Initially, Cohen et al. [20] proposed a 14-item Likert-type questionnaire with 5-point scales ranging from 0 (never) to 4 (very often) to assess the degree to which everyday life events were perceived as stressful. The authors documented the psychometric criteria using university student samples, using the test-retest method to demonstrate response stability over time, and reported alpha indices of 0.83. They also noted that “the coefficient alpha reliability estimate for the four-item PS was 0.72. The test-retest reliability of the four items over the four-item scale over a two-month interval was 0.55” ([20], p. 392). In Mexico, previous studies conducted with adults from the general population have confirmed the psychometric stability of this scale [15,16]. Therefore, to reduce response bias due to fatigue, this study used the four-item version.

2.3 Procedure

The field application lasted 15 days. Teachers from two areas of knowledge, psychology and sciences, were contacted and trained to apply the questionnaire to their students. To homogenize the administration conditions, each teacher received an instruction sheet. The classroom application lasted, on average, 15 minutes. Each student was asked to apply the same questionnaire to his or her grandparents (maternal and/or paternal); if he or she did not have them, the corresponding questionnaire was left blank. The questionnaire was sent electronically, and the results were processed using RStudio [21] and R 4.3.2. [22]. Likewise, the packages: diverse [23], psych [24], caravan [25], stools [26], and tidy SEM [27].

The instrument was applied in the schools of Psychology, Sciences, and Higher Studies Aragon of the National Autonomous University of Mexico between October and December 2023 through Google Forms. Of the 396 questionnaires applied, 48 were omitted because participants answered more than once, did not answer, or indicated they had no grandparents, leaving the remaining answers blank. To avoid any potential order effects between the five scales mentioned in the Instruments section, they were presented in five different sequences determined by a Latin square. Participants were randomly assigned to each sequence. The distribution of participants in the various versions of the questionnaire was as follows: 1 (n = 82), 2 (n = 70), 3 (n = 67), 4 (n = 56), and 5 (n = 72). To analyze the characteristics of the participants by age group, three groups were considered: those under 29 were deemed young people, those over 60 were considered older adults, and one person was not classified since his age was between 50 and 54. Thus, the sample consisted of 257 young people and 90 older adults.

2.4 Data Analysis

The database was thoroughly cleaned by eliminating records with multiple responses, no responses, or inconsistent information. After this cleaning process, we generated the socio-demographic profile of the sample. Subsequently, an exploratory factorial analysis (EFA) was conducted to analyze the psychometric properties of the Family Inclusion Scale (FIS). To obtain a preliminary assessment of the instrument's convergent and discriminant validity, we conducted a Confirmatory Factor Analysis (CFA) using data from the FIS and the Loneliness and Perception of Stress Scales, all administered to the same participants. The results of this analysis enable us to evaluate item loadings, as well as the correlation between factors, to determine convergent and discriminant validity.

2.5 Sample

The final sample consisted of 347 participants: 90 older adults over 60 years of age and 257 younger adults under 29. Of the older adults who responded to the questionnaire, 47 were maternal grandmothers, 13 were paternal grandmothers, 17 were maternal grandfathers, and 13 were paternal grandfathers. Of them, 30 were men, and 60 were women. Most of these participants were married, in a committed relationship, or in a free union (n = 33), and the rest were widows (n = 43), separated (n = 9), or single (n = 3). In terms of degree of study, responses were distributed across all levels, with non-study participants identified (n = 11), participants with secondary or primary, partial or complete (n = 60), and some with preparatory, bachelor's, or postgraduate degrees, complete or incomplete (n = 19). Regarding occupation, 34 participants were retired and contributing to family expenditure, 21 lived independently of the family, 20 were not working and dependent on the family, and 15 worked in paid employment.

Of the young adults, 105 were males, and 152 were females. Most were single (n = 253), and four were engaged, married, or living with a partner. The majority reported having an incomplete bachelor's degree (n = 188). In contrast, the remaining participants reported having a full bachelor's degree (n = 19), a high school diploma (n = 47), an incomplete high school diploma (n = 22), or a postgraduate degree (n = 1). Regarding work activities, 209 indicated not working and living with their families, 45 indicated working on a paid basis, and 3, probably by mistake, indicated being pensioners. The data from these three participants were evaluated individually, and it was observed that the responses were consistent with those of the young people.

2.6 Psychometric Criteria of the Family Inclusion Scale (FIS)

Distribution of answers. The FI scale consists of 10 Likert-type items whose response options were evaluated from 1 (little or almost no) to 4 (almost daily). The responses were placed in the “at least once a week or almost daily” options for eight of the ten items, corresponding to the large-scale values. For the FI5 item (“Speak to your family about the fears you have”), most of the answers are in the low values, little or almost nothing, and at least once. Finally, the FI2 item (“Receives flattery comments and/or recognition from your family”) has the most responses in intermediate value options, i.e., at least once a month and at least once a week.

3. Results

3.1 Exploratory Factor Analysis-EFA

Exploratory Factor Analysis-EFA. An initial exploratory analysis was conducted across young and older adults; since no differences were observed among item correlations, AFE was carried out with data from the entire sample. For this analysis, polychoric correlations were used [28]. The Kaiser-Meyer-Olkin (KMO) and Barlett's tests were carried out to determine whether the data were suitable for factor analysis. The KMO test assesses the suitability of the sample for factor analysis. The test measures sample adequacy (MSA) for each variable observed in the model and the complete model. The test showed adequate overall MSA = 0.83 and MSA values (>0.8) for most items, except FI3 and FI9, which had values <0.7 and were excluded. In addition, when conducting the exploratory factor analysis (EFA) with these items included, the factor loadings were 0.36 for item IF3 and 0.41 for item IF9. Moreover, it was determined that these items evaluated the ability of adults to provide support to their families, rather than how families integrated them into their dynamics. Barlette's test was significant $\chi _{45}^{2} $ = 1900, suggesting substantial correlations between the items.

Eigenvalues obtained from the exploratory analysis indicated that the organization was a single factor. Thus, AFE was carried out considering a single factor ($\chi _{20}^{2} $ = 1510.16), and an RSMR = 0.07 was obtained, which is regarded as acceptable. However, problems were detected with the model adjustment; the RMSEA index = 0.157 [90% Confidence Interval 0.137-0.178] was higher than the desirable value of 0.08, and the Tucker-Lewis index = 0.839 was slightly below the recommended value of 0.90. However, a 53% explained variance was achieved, which substantially supported the statistical consideration of the constructed factor, with loadings >0.4 (Table 1), making it feasible to obtain a reasonably acceptable reliability of α = 0.86 with eight items.

Table 1 Exploratory Factor Analysis with Varimax Rotation for the Familiar Inclusion Scale [FIS] (n = 347).

3.2 Convergent and Divergent Validity

To conduct a preliminary assessment of convergent and divergent validity, we performed a confirmatory factor analysis (CFA) on the same sample. It is important to note that this approach implies that the CFA does not provide an independent test of the model. Therefore, the results should be interpreted as preliminary and will require future cross-validation. To determine the divergent and convergent validity of the Family Inclusion Scale, the Family Alienation-FA subscale, part of the Loneliness Scale in Older Adults IMSOL-AM [19], and the Perceived Stress Scale-PS [20] were used. The FA scale measures people's perception of the possible emotional exclusion in their family circle. Thus, this scale was considered to evaluate the divergent validity of the Family Inclusion scale. The FA subscale has 10 Likert-type items of 5 points, with the highest value being the most excellent perception of family alienation. The perceived stress subscale [20] was used to evaluate convergent validity. This scale consisted of four 5-point Likert-type items; since initially, two of the four items had a positive direction of perceived stress, we recoded them in the negative direction. Thus, the highest values indicated less stress. Therefore, the PS scale was used to evaluate convergent validity since it is expected to correlate positively with the FI scale, given that family inclusion should produce less stress. Furthermore, the sub-scales' reliability was assessed using Cronbach's alpha. The reliability indices obtained were for the FI scale, α = 0.86, whereas for the FA subscale, α = 0.92, and for the PS subscale, α = 0.71.

3.3 Confirmatory Factor Analysis CFA

To confirm the factor structure of the Family Inclusion-FI scale, a diagonally weighted least squares estimate (DWLS) was utilized, as recommended for polychoric correlations with ordered data (i.e., data from Likert scales). The measurement model considered covariance between items of the same scale with modification indices (mi) greater than 14. Thus, the items L5~~L9, L2~~L4, L4~~L15, and L8~~L9 were considered correlated in the Loneliness scale. On the family inclusion scale, the correlated items were FI4~~FI10, FI7~~FI8, and FI1~~FI10. It obtained a $\chi _{220}^{2} $ = 433.544, an RMSEA robust = 0.078, and a SRMR = 0.06, which suggest a proper model fit (See Figure 1 and Table 2 for details). The Tucker Lewis Robust Index (TLI) = 0.901, which should be >0.90, while the Robust Comparative Adjustment Index (CFI) was 0.914, which should also be >0.90. No unfavorable variance was observed for any of the parameters considered. The α.ord values for the scales were 0.94, 0.90, and 0.75 for Loneliness, Family Inclusion, and Stress Perception, respectively. Also, the ω values were 0.92, 0.84, and 0.72 for the Loneliness, Family Inclusion, and Stress Perception scales, demonstrating the reliability of the used scales.

Figure 1 Confirmatory Factor Analysis with the raw loadings and the covariance estimates.

Table 2 Confirmatory Factor Analysis, Reliability, and Validity Indicators for the Familiar Inclusion, Loneliness, and Perception of Stress Scales.

Figure 1 shows that the standardized factor loadings for most items were above 0.5 and significant p < 0.05, except for item PS2 of the perceived stress scale, which had a factor loading of 0.483 and was not important. The variances drawn for the Loneliness, Family Inclusion, and Stress Perception scales were 0.61, 0.53, and 0.45, respectively. The Loneliness and Family Inclusion scales exhibit adequate convergent validity metrics. A positive correlation (+0.4465, [0.3433, 0.5497]) between the Family Inclusion and Low-Stress Perception scales was observed for convergent validity. That is, those people with high scores on the Family Inclusion Scale show high scores on the Lower Stress Perception scale. In addition, the factor loadings for the Stress Perception and Family Inclusion scales were ≥0.7 for almost all items, except for item PS2 on the Perception of Stress scale and items FI8 and FI10 on the Family Inclusion scale. AVE values ≥0.5 for the FIS indicate that the construct explains more than 50% of the variance. However, the AVE for the stress perception scale was lower than the expected value of 0.5, likely due to the small number of items in that scale. In contrast, the Loneliness Scale is negatively correlated with the Family Inclusion scale (-0.6729, [-0.7404, -0.6051]) and with the Low-Stress Perception scale (-0.7468, [-0.8061, -0.6875]). The Fornell-Lacker criterion indicated discriminant validity between the FIS and the Loneliness Scale. The square root of AVE was 0.73, which is less than the square of the correlation between the FIS scale and loneliness (0.445). This finding suggests that the construct evaluated through FIS accounts for more variance among its indicators than among the other constructs in the measure model.

4. Discussion

Based on the results, it is assumed that the objective of this study was fulfilled, as the psychometric properties of a psychological instrument measuring family inclusion among Mexican adults were documented. It now has a valid and reliable questionnaire that reports on the peculiarities associated with emotional inclusion perceived by adults in relation to their nuclear family. It is important to note that this instrument, in addition to meeting the psychometric criteria established by Nunnally & Bernstein [29], is distinguished by its simplicity in the conceptualization of family inclusion and by its synthetic extension, features that allow it to be used with adults and older adults without the risk of prejudices in the responses for fatigue. It is worth noting that the elimination of the fact that only two items from the initial 10 demonstrates a conceptual approach consistent with the emphasis on the emotional component and a parsimonious representation of the concepts to be represented. According to Nunnally & Bernstein [29], the psychometric ideal is to have a reliable measure with few items that show temporal stability across several samples. The evidence obtained in this study shows both conceptual coherence and response stability.

This instrument is valid, on the one hand, to document the quality of the affective relationship that adults assume, as they perceive themselves as essential members within their family nucleus. The relevance of this instrument is also confirmed by documenting some of the actions that adults carry out within their family interactions (e.g., offering help, giving their opinion), based on which they can perceive that they play a relevant role within their families.

It is important to note that the scale of family inclusion reflects both the efficiency and significance of the affective interactions established as a means for a person to feel valued within his/her family. Family inclusion as a process can be generated in two directions: (a) active, where the focal subject offers help, support, or advice to another member of his or her family, and (b) receptive, where a focal person can receive social or affective recognition from his/her family. Thus, in the process of family inclusion, the focal person can act as a source or recipient of affection. This is essential for adjusting possible interventions based on the assessment obtained with this scale. If the person, as a recipient, does not perceive they are sufficiently included in family dynamics, interventions should be oriented towards examining the context. In contrast, if the focal person is the one who initiates the action by offering support, help, or collaboration in his/her family, and in the diagnosis, they do not perceive reciprocity. The intervention strategy should be oriented to the emotional modulation the person offers or to the meaning it gives to the affective exchanges it establishes. In any case, the evaluation derived from this scale offers valuable data to place the focal person within family dynamics and thus to assess the possible risk of depression due to emotional exclusion or to ratify the emotional functionality when evaluating a fluid and explicit inclusion of the person within their family nucleus.

In psychometric terms, it is pertinent to note that the family inclusion scale (FIS) meets two strategic characteristics proposed by Nunnally & Bernstein [29]: it is clear in its conceptualization and easy to apply. The conceptualization of family inclusion is based primarily on the consideration of the emotional component, which, as has been widely documented, is essential in the psychosocial development of the individual. At the same time, the application is optimal because with few items, a valuation is obtained of the affective quality perceived by the focal subject in their family interactions.

In statistical terms, the family inclusion scale met both the criteria of discrimination and factor grouping. The factorial weights obtained ratified the integration of family inclusion into a single factor, which makes the interpretation of the scores received simple. In the Confirmatory Factor Analysis section, we included the covariance between items within the same scale in the measurement model specification. This inclusion led to improved fit measures compared to those observed in the Exploratory Factor Analysis (EFA), particularly in the Root Mean Square Error of Approximation (RMSEA). However, these findings indicate a need for a more detailed evaluation of the correlations between these items. Since this result could suggest a similar wording or participants perceiving the items as identical. Additionally, the adjustment values observed in both the confirmatory factor analysis and the tests for convergent and divergent validity clearly demonstrate the conceptual and empirical relevance of the developed questionnaire. However, we acknowledge the methodological limitation of using the same sample in the EFA as in the CFA. Further studies are needed to address these shortcomings.

While it is assumed that the family inclusion questionnaire is valid and reliable, and can be used with Mexican samples of adults and older adults, it is pertinent to note that, like any indirect measurement instrument, it still has limitations. In this regard, in conceptual terms, it would be appropriate to broaden the notion of family inclusion, exploring both affective, cognitive, and even behavioral dimensions, to enrich the possible forms of inclusion within families. In empirical terms, it would be strategic, in future studies, to expand the sample to include more older adults to make comparisons by age and gender in order to detect possible bias and differences in family inclusion based on these two variables.

5. Conclusions

In conclusion, by its conceptual relevance, by emphasizing the emotional peculiarities established by adult persons within their family, by the statistical robustness with which their psychometric properties were documented, for the simplicity of its application, and for the clarity in the interpretation of the results, it is assumed that the Family Inclusion Scale-FIS is a valid and reliable instrument that can be used efficiently with samples of adult Mexican persons.

Acknowledgments

We thank to the Academic Personnel Affairs Department of the National Autonomous University of Mexico-UNAM for the support granted to Dr. Montero to carry out a sabbatical stay at the University of Barcelona (September-December 2023).

Author Contributions

María Montero-López Lena: Design of the instrument under evaluation. Conceptual framework. Coordination and supervision of data collection in the field. Writing the first draft. Revision of the article based on reviewer feedback. Livia Sánchez Carrasco: Development of the field application strategy. Execution of statistical analyses. Collaboration on the first draft of the article. Adjustments to the text based on the reviewers' comments. Montserrat Celdrán Castro: Participation in the initial planning of this study. Discussion of the conceptual dimensions to be evaluated. Participation in the first draft. Analysis of the comments received from the reviewers.

Funding

This work was supported by UNAM-DGAPA through a sabbatical grant for the first author at the University of Barcelona and by the UNAM-PAPIIT project IN303925.

Competing Interests

The authors have declared that no competing interests exist.