Abstract
The study aimed to determine the dimensionality, internal consistency, and nomological validity of the Sexual Orientation Experiences of Discrimination (SOEOD-9) among lesbian, gay, bisexual, and queer people in Santa Marta, Colombia. The study included 303 individuals between 18 and 70 years old. Participants completed the SOEOD-9. Dimensionality was explored using confirmatory (CFA) factor analysis; internal consistency with Cronbach’s alpha and McDonald’s omega; and nomological validity with Student’s test. The SOEOD-9 presented two factors that explained 51.6% of the variance and high internal consistency. Nevertheless, the CFA showed poor indicators of goodness-of-fit for the two-dimensional solution. A five-item version (SOEOD-5) improved overall performance (dimensionality, internal consistency, and nomological validity). In conclusion, the SOEOD-9 shows a two-factor internal structure with poor goodness-of-fit indicators and acceptable internal consistency and nomological validity. The SOEOD-5 presents better global indicators of dimensionality, internal consistency, and nomological validity.
Stigma, prejudice, stereotype, and discrimination are phenomena studied by psychology, anthropology, and sociology (Hendry, 2016; Stangor, 2016), mainly based on attribution theory (Hernández & López, 2011). In recent years, these terms were unified in a theoretical construct called the “stigma-discrimination complex,” a complex, indivisible, spiral sequence that begins with stigma, continues as prejudice, becomes a stereotype, and ends in discrimination (Campo-Arias & Herazo, 2014).
In this complex, social stigma is understood as a character with a negative connotation, which appears as a result of social interactions when the individual does not satisfy social expectations (Campo-Arias & Herazo, 2013; Goffman, 1963). Stigma gives way to prejudice when valuation is marked as undesirable; it is a value judgment with a solid emotional base that supports unfavorable attitudes towards a social group, which are manifested in negative attitudes towards members who belong to undesirable and devalued social categories (Allport, 1954; Campo-Arias & Herazo, 2015; Ramírez et al., 2016). A stereotype is a set of qualities, attributes, or specific traits attributed to the social group by which the group is homogenized and qualified (Campo-Arias & Herazo, 2014; Lippman, 1991). Stigma, prejudices, and stereotypes materialize in discrimination, which encompasses all behaviors, subtle or overt, of exclusion and systematic violence against groups that disturb norms, values, and social well-being according to the hegemonic group or that holds the social, political, or economic power (Awad & Rackley, 2017; López et al., 2008).
The stigma-discrimination complex can be presented against any social group. However, the most studied fields correspond to groups discriminated by race, ethnic origin, gender, religion, and sexual orientation (Carter et al., 2017; Ghumman & Ryan, 2018; Heilman & Caleo, 2018; Jones et al., 2017; Sugarman et al., 2018; Zschirnt & Ruedin, 2016). The experiences of discrimination, and more if they are repetitive, constitute a risk factor for the physical and mental health of the discriminated groups (Carter et al., 2019; De Freitas et al., 2018; Levefor et al., 2020; Semlyen et al., 2016). The stigma-discrimination complex is considered a growing public health problem due to the negative consequences in the short and long term on people’s well-being (Campo-Arias et al., 2014; Gulliford, 2019).
Consequently, Krieger et al. (2005) carried out the process of construction and validation of the Experiences of Discrimination (EOD) in order to quantify the experiences of discrimination on ethnic-racial grounds in a sample of 228 African Americans and Latinos between 25 and 64 years, residents of the United States. The EOD is made up of nine items that are scored from zero (never) to three (four times or more) and measure the frequency with which an individual has perceived stigma-discrimination in nine contexts: at school, when applying for employment, in work, at home, when renting a house, receiving medical attention, applying for a loan, on the street, and by judicial authorities. The EOD showed a one-dimensional internal structure with acceptable indicators of goodness-of-fit (chi-square = 200.5, degree of freedom = 80, p = .05, RMSEA = .05, and CFI = .94) and an acceptable internal consistency (Cronbach’s alpha of .74) (Krieger et al., 2005). In Colombia, Campo-Arias et al. (2014) explored the EOD structure and internal consistency in a group of 361 medical students from Bucaramanga. They reported a one-dimensional internal structure that explained 52.9% of the variance, using exploratory factor analysis (EFA) and high indicators reliability (Cronbach’s alpha of .88 and McDonald’s omega of .89).
Although the EOD was constructed to measure experiences of discrimination by ethnicity or race, the wording of the items allows modification for experiences of discrimination by other characteristics (Krieger et al., 2005). Therefore, Lee et al. (2016) adapted six of the nine items for experiences of discrimination based on sexual orientation during the last year: when obtaining insurance coverage or receiving medical attention, in public places, obtaining a job, work, being admitted to an educational institution and in courts or by the police. However, they did not report any psychometric performance indicators due to the study’s objectives. Subsequently, in Macedonia, Stojanovski et al. (2017) reported that EOD showed high internal consistency (Cronbach’s alpha of .82) in 208 adults of sexual minorities and 188 adults of hegemonic sexual identity.
Therefore, in the present study, a specific adaptation was carried out to know the EOD’s performance by sexual orientation (SOEOD-9) in people with diverse sexual identities, lesbians, gays, bisexuals, and travesties, transgender, transsexual, intersexual, and queer (LGBTTTIQ). LGBTTTIQ people tell a long history of stigma-discrimination based on non-hegemonic sexual identity (Cassey et al., 2019; Wang et al., 2019).
Besides, nomological validity, also known as a hypothesis or theoretical validity, is explored, which seeks to observe the presence or absence of correlation between the construct evaluated by a scale and another independent construct, which theoretically has a statistically significant association (Adcock & Collier, 2001). The present study used the variable “coming out” with the parents to explore the nomological validity. When the individual assumes sexual orientation before the family, it is more frequent for them to behave publicly as a sexually diverse person, and consequently, the risk of discrimination in different socialization contexts increases (Doan Van et al., 2019; Gattamorta & Quidley -Rodríguez, 2018; Wax et al., 2018).
Evidence indicates that experiences of discrimination based on sexual orientation increase the risk of mental health problems and disorders in LGBTTTIQ, including depressive symptoms of anxiety and the spectrum of post-traumatic stress disorder with a significant deterioration in the quality of life (Bialer & McIntosh, 2017; Mason et al., 2018; Petruzzella et al., 2020; Richardson & King, 2017). However, little is known about discrimination experienced in the population with diverse Colombian sexual orientations (Nieves-Lugo et al., 2020). Knowing the implications of discrimination in LGBTTTIQ people requires measurement instruments, such as the SOEOD-9, that can be used in population studies and allow the valid and reliable measurement of experiences of discrimination (González-Rivera & Pabellón-Lebrón, 2018). This knowledge will allow affirmative actions to reduce the impact of discrimination based on sexual orientation and promote the health and well-being of LGBTTTIQ (Lee & Kanji, 2017; Puckett et al., 2016).
This study aimed to know the structure and internal consistency of SOEOD-9 in a group of LGBTTTIQ from Santa Marta, Colombia.
Method
Study design and ethical issues
A validation study of a scale was carried out. According to Resolution 8430 of 1993 of the Ministry of Health and Social Protection, the research did not represent any risk for the participants; techniques and methods were applied without intervening or modifying psychological variables anonymously confidentially. Participants must sign an informed consent.
Participants
Three hundred three individuals with diverse sexual orientations participated, selected by non-probabilistic, snowball-type sampling. The participants’ ages ranged between 18 and 70 years (M = 25.7, SD = 7.3). More information is available in Table 1.
Characteristic of the sample.
Instrument
Participants completed SOEOD-9. The SOEOD-9 comprises nine items that explore discrimination events in nine contexts, from school to interaction with police and authorities.
Each item offers four response options that are scored from zero to three, where zero is “never,” one is “once,” two is “two or three times,” and three is “four times or more.” The version used appears in Table 2.
SOEOD-9.
Procedure
The information was collected in person. The printed version consisted of the informed consent, a sociodemographic questionnaire, and the SOEOD-9. The participants filled out the questionnaire anonymously, and each one was assigned a code to preserve confidentiality. The information collected was digitized into a database.
Analysis of data
The statistical programs IBM-SPSS version 25 and STATA were used. The Bartlett (1950) test of sphericity was performed, in which a high square-chi and probability values (p) less than 5% were expected. Likewise, the Kaiser-Meyer-Olkin, KMO (Kaiser, 1974) sample adequacy test was applied, in which a value greater than 0.70 was expected. Adequate indicators in both measures reveal one or more latent factors in a set of items. With the maximum likelihood method and Promax rotation, EFA was used to explore the structure of SOEOD-9, and the communalities and loadings were observed.
In the confirmatory factor analysis (CFA), the goodness-of-fit coefficients were found with the chi-square test, with degrees of freedom (df) and p-value, Root Mean Square Error of Approximation (RMSEA) and confidence intervals of 90% (90%CI), Comparative Fit Index (CFI), Tucker-Lewis index (TLI), and Standardized Mean Square Residual (SRMR). A probability value greater than 5% is expected in the chi-square, with values less than 0.06 for RMSEA and SRMR, and greater than .89 for CFI and TLI. At least three of these indicators were expected to be within desirable parameters (Hu & Bentler, 1999). Other versions would be tested if this were not achieved by observing the modification indices’ results.
The scale’s internal consistency was calculated using the coefficients Cronbach’s alpha of Cronbach and McDonald’s omega of McDonald. Cronbach’s alpha is the internal consistency measure most used in the research; however, McDonald’s omega is a more accurate measure when the tau equivalence principle is not fulfilled, necessary for the precise calculation of internal consistency (Campo-Arias & Oviedo, 2008). It is recommended to report both coefficients to have at least two reliability indicators in the validation studies. Cronbach’s alpha and McDonald’s omega values higher than .70 are expected (Campo-Arias & Oviedo, 2008).
The authors compared the scores on the SOEOD-5 between the participants who reported coming out to their parents and those who did not, using the Student’s t-test (1908) after verifying the homogeneity of the variance with the Levene test (O’Neil & Mathews, 2002). Probability values less than 1% (p < .01) were accepted as significant.
Results
Dimensionality
The SOEOD-9 showed scores between 0 and 18 (M = 3.64, SD = 3.87). Barlett’s sphericity test indicated a chi-square of 654.9; df = 36, p < .001, and the KMO test was .80. These values allowed us to advance in the EFA, in which two latent factors were evidenced: the first with an eigenvalue of 3.4 that explained 37.9% of the variance, which was preliminarily called “proximal discrimination” (items 1, 6, 8, and 9), and the second one called “distal discrimination” (items 2, 3, 4, 5 and 7), with an eigenvalue of 1.2 responsible for 13.7% of the total variance. The subsections showed communalities between .272 and .728 and coefficients between .498 and .875. See details in Table 3.
Communalities and loadings of the soeod-9.
In the CFA, the goodness-of-fit indicators were poor for the two-dimensional solution (chi-square = 104.54, df = 26, p < .001, RMSEA = .10, 90%CI .08 - .12, CFI = .88, TLI = .83, and SRMR = .06). Given these findings, a one-dimensional model was tested with the nine items that make up the scale. Likewise, this solution showed modest goodness-of-fit coefficients (chi-square = 141.18, df = 27, p = .001, RMSEA = .12, 90%CI .10 - .14, CFI = .81, TLI = .74, and SRMR = .07).
However, other versions of five items were explored due to the internal consistency coefficients presented below, and the best performance was shown by the version that included items 3, 5, 6, 8, and 9 (SOEOD-5). The SOEOD-5 presented scores between 0 and 12 (M = 2.17; SD = 2.68). The Bartlett test of sphericity showed a chi-square of 288.03; df = 10, p < .001, and the KMO test of .795. A single factor with an eigenvalue of 2.5 was retained, which explained 49.4% of the total variance. The goodness-of-fit coefficients were chi-square = 8.11, df = 5, p = .15, RMSEA = .05, 90%CI .01 - .10, CFI = .99, TLI = .98, and SRMR = .02. See Table 4 for communalities and coefficients of the SOEOD-5.
Communalities and loading of the SOEOD-5.
Internal consistency
In the one-dimensional model, the SOEOD-9 showed Cronbach’s alpha of 0.75 and McDonald’s omega and .79; and in the two-dimensional model, Cronbach’s alpha was .66 for Factor 1, and Cronbach’s alpha of .70 for Factor 2; and McDonald’s omega of .71 for Factor 1 and McDonald’s omega of .73 for Factor 2. The SOEOD-5 presented Cronbach’s alpha of .71 and McDonald’s omega of .74.
Nomological validity
The scores on the SOEOD-5 in those who said they had informed their parents of their sexual orientation showed higher scores than those who did not (M = 2.70, SD = 2.85 vs M = 1.48, SD = 2.26, t = 4.15, df = 300.99, p < .01). The t value was taken for non-homogeneous variances (F = 11.97, p < .001).
Discussion
In the present study, among LGBTTTIQ people of Santa Marta, Colombia, the SOEOD-9 shows an internal structure of two factors with unacceptable goodness-of-fit coefficients; however, it presents high internal consistency. The SOEOD-5 shows better overall performance with good structure and internal consistency coefficients.
The present investigation observed that the SOEOD-9 presents high internal consistency for one- and two-dimensional structures. However, the goodness-of-fit indicators for these structures are below what is accepted. In contrast, the SOEOD-5 showed a one-dimensional structure with excellent goodness-of-fit indicators and high internal consistency. The SOEOD-5 is the result of eliminating four items with poor performance: item 1 (school), item 2 (applying for a job), item 4 (renting a house), and item 7 (bank loan).
The performance of SOEOD versions is not known with certainty; to date, there is no exploration similar to that of the present study. In Macedonia, Stojanovski et al. (2017) found Cronbach’s alpha of .82, apparently for the SOEOD-9. In ideal conditions, one-dimensional scales are preferred, and the retained factor explains at least 50% of the variance (Campo-Arias et al., 2017; Gorsuch, 1997). Besides, it is expected to observe the goodness-of-fit coefficients within the range preset as acceptable, previously noted (Hu & Bentler, 1999). Finally, it is necessary to have adequate reliability, and for this, it is expected to find internal consistency values between .70 and .95 for Cronbach’s alpha and McDonald’s omega (Campo-Arias & Oviedo, 2008).
Regarding the nomological validity, a higher perception of discrimination was found in the participants who had come out of the closet with their parents. These findings are consistent with the available evidence. Keeping a secret sexual orientation is often used to avoid discrimination among adolescents and young adults (Dueñas et al., 2022; Zheng et al., 2020). Public expression of sexual diversity significantly improves psychological well-being (Perales et al., 2020); however, visibility increases the risk of discrimination, especially in hostile sociocultural contexts toward sexual diversity (Suppes et al., 2021; Wei et al., 2017).
The SOEOD-5 is a short, valid, and reliable measure that quantifies discrimination experiences based on sexual orientation. The best overall performance was observed for the SOEOD-5 in the present investigation. This observation is promising, given that it allows us to have an instrument for measuring the experience of discrimination based on sexual orientation in a diverse Colombian sexual population (González-Rivera & Pabellón-Lebrón, 2018). These findings constitute a significant empirical contribution to the medical and social sciences and research that explores the experiences of discrimination based on sexual orientation and mental health in the LGBTTTIQ population.
The present study makes a notable contribution to scientific knowledge; however, it is necessary to recognize some research limitations. Homosexuality and bisexuality are still taboo subjects in Colombian society (Zambrano et al., 2019), making it difficult to approach the study population. At the time of application, only people who publicly recognized, to a greater or lesser extent, diverse sexual orientation was included in the sample. The psychometric performance of the different versions of SOEOD can show different indicators if the individuals who remain “in the closet” are evaluated. It is possible that in this group of people, a very distant performance is observed, and, consequently, the results should be interpreted with caution and can only be reproduced in people who are publicly self-recognized as LGBTTTIQ (Baams et al., 2013).
Exploring the specific psychometric performance of the SOEOD-5 in each sexual orientation was not recommended due to the small groups formed with the LGBTTTIQ group’s segmentation. Caution should be exercised because sociocultural factors moderate the manifestation of discriminatory behaviors, which vary according to sexual orientation, identity, and gender expression (Balsam et al., 2013; Cerezo, 2020).
It is concluded that the SOEOD-9 is an instrument with adequate indicators of internal consistency and two factors with imperfect goodness-of-fit (proximal discrimination and distal discrimination). The SOEOD-5 shows better overall performance, both in dimensionality and internal consistency. It is necessary to carry out other studies to know the psychometric performance of different versions of SOEOD.
Data Availability
The data supporting this study's findings are available from the corresponding author upon reasonable request.
Footnotes
Email: yeisondgb{at}gmail.com
Author notes This project was funded by Colciencias in the 2018 Young Researchers and Innovators program (Call 812 of 2018) through the Special Cooperation Agreement No. 80740-251-2019 with the Universidad del Magdalena, Santa Marta, Colombia. Authors declare that they have no interest in the research. Correspondence concerning this article should be addressed to Adalberto Campo-Arias, Universidad del Magdalena, Carrera 32 No. 22-08, Santa Marta, Colombia (Código postal 470004) Teléfono: 57 5 4381000 Ext. 1338. Email: acampo{at}unimagdalena.edu.co
Data availability statement: The data supporting this study’s findings are available from the corresponding author upon reasonable request.
The authors included a nomological validity to have an external validation variable.
References
