Introduction

Sexually transmitted infections (STIs), principally human immunodeficiency virus and acquired immunodeficiency syndrome (HIV/AIDS), are considered chronic worldwide health problems. They affect thousands of men, women, and children through serious issues such as acute disability, infertility, and even death (^{World Health Organization, 2016}). In addition, these pathologies, impact people’s quality of life, increasing the probability of developing depression and anxiety (^{Salazar Campos & Valencia Ortiz, 2018}) as well other psychological issues associated with perceived threats about the future and life projects (^{Moodley, Staunton, de Roubaix, & Cotton, 2016}).

These consequences are worrying, especially considering that 37.9 million people have been diagnosed with HIV/AIDS worldwide, and that 100,000 new cases were registered in Latin America in 2018, a 7% increase compared to 2010. The countries with the largest increases in incidence between 2010 and 2018 are Chile (34%), Bolivia (22%), Brazil (21%), and Costa Rica (21%) (^{Joint United Nations Program on HIV/ AIDS, 2019}). In Chile, the number of confirmed infections progressively increased to a total of 71,000 cases between 1984 and 2018. There was an incidence of 0.27 per 1,000 in 2018 for inhabitants who were mainly concentrated in the first two regions of the extreme north of Chile (^{Cáceres-Burton, 2019}; ^{Joint United Nations Program on HIV/AIDS, 2019}). Most people, infected with HIV, are young (i.e., between 18 and 35 years old), who have a higher frequency of sexual risk behaviours (SRBs) (^{Barrera-Herrera & Vinet 2017}; ^{Cáceres & Pino, 2018}). This was the focus group for the epidemiological research (^{Joint United Nations Program on HIV/AIDS, 2019}).

STIs, especially HIV/AIDS, are mainly transmitted through SRBs (e.g., sexual activity with multiple partners, inappropriate use of protective barriers, and sexual activity under the influence of alcohol or drugs), which increase both the probability of acquiring STIs and unwanted pregnancies (^{FerrerUrbina et al., 2018}). Therefore, the prevention of STIs and/ or HIV/AIDS requires these behaviours to be identified and explained (^{Orcasita, Uribe, Castellanos, & Gutiérrez, 2012}).

As part of this context, research has focused on identifying variables to increase the understanding of the SRBs, and several personality traits stand out (^{Allen & Walter, 2018}; ^{Fernández et al., 2013}; ^{Teva, Bermúdez, Ramiro, & BuelaCasal, 2012}; ^{Wang et al., 2020}). Personality traits have shown consistent relationships with SRBs, principally factors such as sensation seeking and sexual sensation seeking (^{Charnigo et al., 2013}; ^{Danko, Buzwell, & Earle, 2016}; ^{Heidinger, Gorgens, & Morgenstern, 2015}; ^{Xu, Zheng, Liu, & Zheng, 2016}), which could be a better predictor, compared to general personality factors, because they are closer to the behaviour of interest (^{Anglim & Grant, 2014}; ^{Anglim & O’Connor, 2019}; ^{Ashton, Paunonen, & Lee, 2014}; ^{Paunonen, Haddock, Forsterling, & Keinonen, 2003}).

^{Zuckerman (1971)} defines sensation seeking as the tendency to seek new, varied, and complex experiences and sensations, involving physical and social risks, to simply enjoy such experiences. This is measured by the Sensation Seeking Scale, which is designed to assess four dimensions (thrill and adventure seeking, experience seeking, disinhibition, and boredom susceptibility) (^{Zuckerman, 1971}). Since this approach has a broad scope, it is a weak predictor when constrained to more specific domains such as sexuality (^{Anglim & Grant, 2014}; ^{Anglim & O’Connor, 2019}; ^{Ashton, Paunonen, & Lee, 2014}; ^{Paunonen et al., 2003}). For that reason, ^{Kalichman et al. (1994)} designed the Sexual Sensations Seeking Scale.

The Sexual Sensations Seeking Scale is a onedimensional scale designed to assess sensation seeking in sexual domains; it can be defined as seeking sexually novel experiences to obtain the optimal sexual arousal. This scale has been used in a wide variety of research, evidencing consistent relationships with SRBs (^{Burri, 2017}; ^{Gullette, Booth, Wright, Montgomery, & Stewart, 2014}). The original version comprises nine items (e.g., “The physical sensations are the most important thing about having sex” or “I am interested in trying out new sexual experiences”) and shows evidence of validity based on the relationship with other variables (convergent and discriminant) and reliability reports (α > .75). In a later revision, ^{Kalichman and Rompa (1995)} decided to modify the scale in order to increase the internal coherence by discarding two items (e.g., “I have made promises I did not mean to keep to get a person to have sex with me” and “I have felt curious about having anal intercourse without a condom”) and adding four new items that included physical aspects not previously considered (e.g., “The physical sensations are the most important thing about having sex”, “I enjoy the sensation of intercourse without a condom”, “My sexual partners probably think I am a “risk taker”, and “When it comes to sex, physical attraction is more important to me than how well I know the person”).

The 11-item original version has been applied to samples of men who have sex with men (^{Kalichman & Rompa 1995}) and heterosexual undergraduate students (^{Gaither & Sellbom, 2003}). This scale has also been adapted and validated for use in Spain, specifically for adolescents (^{Ballester-Arnal, Ruiz-Palomino, Espada-Sánchez, MorellMengual, & Gil-Llario, 2018}; ^{Teva & Bermúdez, 2008}), young undergraduate students (^{Santos-Iglesias, Moyano, Castro, Granados, & Sierra, 2018}), and gay, lesbian, and bisexual people (^{Gil-Llario, Morell-Mengual, GiménezGarcía, Salmerón-Sánchez, & Ballester-Arnal, 2018}), as well as Mexican married or cohabiting adults (^{de la Rubia, 2018}) and Portuguese adults (^{Pechorro et al., 2015}). Although this scale, with its respective validation’s studies in other contexts, has reported proper psychometric properties, the original one-factor dimensionality of the scale has been discussed in most studies. We have noted the presence of a second factor, characterized by the physical sensation of attraction, and formed by different sets of items that depend on the validation study. These different proposals for the factorial structure of the instrument show a difficulty in interpreting the scores; furthermore, they do not completely represent ^{Zuckerman’s (1971)} theoretical background since ^{Kalichman and Rompa’s (1995)} revised instrument, in both its oneand two-dimensional versions, did not consider important dimensions such as susceptibility to boredom.

In this context, the available measuring instruments have presented an unstable structure and restricted the multidimensionality offered by the original general proposal (^{Zuckerman, 1971}). Therefore, given the limitations in the available studies and considering that the development of measurement instruments in a specific culture may present greater advantages over linguistic adjustments (^{Cohen, Gafni, & Hanani, 2007}), as well as the need to provide evidence of reliability and validity specific to each culture (^{Vázquez, 2014}), the purpose of this work is to develop a new scale for the assessment of sexual sensations seeking, with evidence of validity based on internal structure and relationship to other measures, for use in young people and adults in a Latin American context, which allows an adequate understanding of the observed scores and, consequently, provide ethical guarantees to support the conclusions and decisions, derivates from the measurement process (American Educational Research Association, ^{American Psychological Association, & National Council on Measurement in Education, 2014}). In order to promote the development of research and intervention related to HIV/ AIDS prevention in at-risk populations.

Method

Results

Table 1 shows the fit indicators of three ESEM measurement models, an (M1) original version (three dimensions with 26 items), a (M2) middle debugged version (two dimensions with 15 items), and a (M3) final debugged version (two dimensions with 9 items). Additionally, M3 was tested in a second order CFA version (M4).

Table 1 Global fit of measurements models 

Note: M1 = ESEM with 27 items and 3 dimensions; M2 = ESEM with 15 items and 2 dimensions; M3 = ESEM with 9 items and 2 dimensions; M4 = second order confirmatory factor analysis with 9 items. N° Par = number of parameters; χ2 = chi-square; df = degrees of freedom; CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation. CI = confidence interval; Low = lower; Upp = upper.

According to the most common fit criteria in the specialized literature (CFI > .95; TLI >.95; RMSEA <.06: ^{Schreiber, 2017}), the original model (M1) is not an explanation enough for the observed covariations matrix. For this, we proceeded to review the initial factorial structure since the factors risky sexual emotions seeking and novel sexual emotion seeking presented a considerable number of items with strong cross-loadings. Therefore, some items were eliminated, and both factors were combined in a new dimension (M2) called sexual emotions seeking. This assumed an improvement in the explanation of the observed covariations matrix, albeit below the expected standards (e.g., CFI > .95; TLI > .95; RMSEA > .050: Schreiber, 2017). Subsequently, the M2 was debugged by considering three criteria: (1) selection of strong factorial loads (λ > .50); (2) removing redundant items; and (3) removing items with strong cross-loadings (> .30).

The final debugged scale has nine items divided into two dimensions: Sexual Emotion Seeking (four items) and Tendency for Sexual Boredom (five items). The fit statistics (Table 1), both comparative (e.g., CFI; TLI) and absolute (e.g., x2/gl; RMSEA), suggest that the debugged ESEM (M3) and Second Order CFA (M4) models are good population representations of the observed relationships. Factor loadings, factor covariances, and reliability estimates for each dimension are presented in Table 2.

Table 2 Standardized factorial loadings, factorial covariations and reliability coefficients (Cronbach´s Alfa and McDonald´s omega) for each dimension 

Note: Factorial Loading >.4 are bolded; **p < .01; *p < .05. SES = Sexual emotion seeking; TSB = Tendency to sexual boredom.

The factorial loadings show adequate representations for each factor (λ > .50), and they have low levels of cross-loadings (λ ≤ .30). Structural relationships between sexual emotion seeking and tendency to sexual boredom are high (r > .50; ^{Cohen, 1988}). Reliability estimates are adequate, or at least sufficient (α > .70; ^{Cho & Kim, 2015}), depending on whether McDonald’s hierarchical Omega or Cronbach’s Alpha is used.

Table 3 shows the results of the invariance tests for the final version of the scale (M4) for men and women. The CFI and RMSEA deltas show considerable fit changes in the metric or scalar model compared to the configuration model, indicating that the equivalence between factor loadings and factor intercepts does not have the same meaning between groups. It can, therefore, be assumed that factor loadings and factor intercepts do not behave equivalently between groups.

Table 3 Measurement invariance testing 

Note: ꭓ2: chi-square; gl: degrees of freedom; p: significance; Δꭓ2: change in chi-square; Δgl: change in degrees of freedom; Δp: change in significance.

Finally, Table 4 shows the relationships between the latent dimensions of the Spanish-speaking scale of sexual sensation seeking and scale of risk behaviours. The proposed model has adequate fit (CFI =.973; TLI =.968; RMSEA =.038). According to the observed relationships, Sexual Emotion Seeking has medium (>.30; ^{Cohen, 1988}) and large (>.50; Cohen, 1988) direct effects on risky sexual behaviours. This proves to be statistically different from 0 in the population, while Tendency for Sexual Boredom also has large direct effects (>.50; Cohen, 1988) on risky sexual behaviours, with exception of the relationship between TSB and IUPB, which is small and non-statistically significant.

Table 4 Standardized relationships between latent dimensions of the two-dimensional scale of sexual sensation seeking and scale of risk behaviors 

Note: **p < .01; *p < .05. SAMP = sexual activity with multiple partners; IUPB = inappropriate use of protective barriers; SAIAD = sexual activity under the influence of alcohol or drugs.

Discussion

The purpose of this study was to develop a multidimensional scale to assess sexual sensation seeking to be used in young people and adults. The statistical fit of the final model (M3), the size of the factorial loadings, and the absence of meaningful cross-loadings allow the model’s structure to be supported. The evidence is valid, based on the internal structure for the proper interpretation of the dimensional scores. Additionally, reliability coefficient estimations allow us to support the fact that each dimension has an adequate internal consistency level, which minimizes measurement errors.

In terms of the validity evidence based on the relation with other variables, the relationships between the dimensions of the Two-Dimensional Sexual Sensation Seeking Scale and the Scale of Risk Behaviours are as expected, according to the specialized literature (^{Burri, 2017}, ^{Gaither & Sellbom, 2003}; ^{Gullette et al., 2014}). The only exception is the relationship between the dimensions of Tendency for Sexual Boredom and inadequate use of protective barriers, which is small and not statistically significant.

The proper statistical fit of the second-order model (M4) and observed factorial covariations in ESEM model (M3) allow us to use the Two Dimensional Sexual Sensation-Seeking Scale in both its two-dimensional and one-dimensional form: this measuring instrument is able to reflect the sexual sensation seeking construct through specific dimensions (Sexual Emotion Seeking and Tendency for Sexual Boredom) that represent most of ^{Zuckerman’s theoretical background (1971)} in a unidimensional way that represents ^{Kalichman et al.’s (1994)} initial proposal. Although, the two-dimensional model reported in this study and the models presented in the validation studies carried out by ^{Santos Iglesias et al. (2018)}, ^{Ballester-Arnal et al. (2018)} and ^{Gil-Llario et al. (2018)} address seeking novel experiences and have adequate psychometric properties, they differ through the substantial content of the second dimension. The validation studies identify this as the sensation of physical aspects. This dimension contains items (e.g., “I enjoy the sensation produced by sexual relations without a condom”) that could correspond to another construct such as risky sexual behaviour, which may not be expressing the true values of the associations between these two different constructs.

According to invariance standards suggested by ^{Chen (2007)}, it is not possible to support metric and scalar invariance of the measurements between genders. Caution should therefore be exercised in interpreting levels of sexual sensation seeking when making comparisons between men and women. Although this result is contrary to the findings in the sexual sensations seeking scale, validation studies have shown complete or partial invariance by gender in the adult Mexican population and the Spanish population of adolescents and gay, lesbian or bisexual persons (^{Ballester-Arnal et al., 2018}; ^{Gil-Llario et al., 2018}; ^{de la Rubia, 2018}). This is consistent with the previous validation study by ^{Santos-Iglesias et al. (2018)} on Spanish adults. This discrepancy could be related to cultural and social aspects, which influence men’s and women’s behaviour. Multiple studies show that heterosexual men manifest higher levels of sexual sensation seeking than heterosexual women (^{Burri, 2017}; ^{Flanders, Arakawa, & Cardozo, 2013}; ^{Gaither & Sellbom, 2003}; ^{Martín, 2017}). This result holds when heterosexual men are compared with homosexual men (^{Gil-Llario, Morell-Mengual, Ballester-Arnal, Giménez-García, & Castro-Calvo, 2015}). Based on the theory of social roles (^{Chrisler & McCreary, 2010}), certain sexual behaviours are accepted for one gender and rejected for the other. This is because each gender faces a different socialization: women are characterized by greater affection and expressiveness (^{Pascoal, Lopes, & Rosa, 2019}), while men are characterized by the achievements and instrumentalization of the sexual act (^{Vanwesenbeeck, 2013}).

The main restrictions of this study correspond to the size and representativeness of the sample; therefore, it is important to make new psychometric studies using this scale to increase the capacity of its generation, mainly for higher risk populations (adolescents, people with substance use disorders, men who have sex with men, sex workers, etc.). In order to obtain additional evidence of validity and representativeness, we recommend applying this instrument in health, medical, and educational contexts. Including this scale in an assessment protocol in health services or educational establishments will make it possible to evaluate the need for the application of specific psychological techniques, such as self-control or training to identify risky behaviour. Furthermore, the information provided by this measurement instrument could also be useful to identify groups of subjects that require targeted preventive interventions. Consequently, designing current interventions and strategies that promote sexual health among young people and young adults will be complemented and improved.

Finally, despite the limitations, the present scale not only constitutes a new brief instrument developed with contemporary psychometric techniques, but it also establishes an updated proposal to evaluate sexual sensation seeking as a specific aspect of the personality, which shows evidence about its effects on risky sexual behaviours SRBs (^{Allen & Walter, 2018}; ^{Burri, 2017}; ^{Gullette et al., 2014}; ^{Fernández et al., 2013}; ^{Teva et al., 2012}; ^{Wang et al., 2020}).

Conclusions

The final version (nine items) of the Two-Dimensional Sexual Sensation Seeking Scale (2-SSS) shows evidence of reliability and validity based on the internal structure of the test and on the convergence with other measures. Initial evidence suggests that the current scale constitutes an alternative to assess sexual sensation seeking; in addition, it can be used to develop research on the psychological factors involved in sexual behaviours¹.