Introduction

The concept of quality of life combines situations and conditions of the persons' living environment and the interpretations and evaluations that the individuals make on the objectives aspects of their surroundings (^{Meda et al., 2008}). Subjective well-being is a part of quality of life, expressing people's satisfaction and their degree of pleasure with specific or global aspects of their existence, in which positive moods prevail. Therefore, it is considered part of health in its most general sense, being relevant to every sphere of human activity (^{García-Viniegras & González, 2000}). ^{Martínez (2004)} mentioned that research into subjective well-being is one of the three ways to assess people's quality of life, together with economic and social indicators, thus suggesting that how people feel and what they think about their life is essential to understanding their well-being. From this perspective, the concept of subjective well-being has two dimensions: cognitive and affective. The cognitive component of subjective well-being is life satisfaction, either globally or by specific domains, whereas the affective component refers to the presence of positive feelings, which some authors call happiness (^{Arita, 2005}).

Life satisfaction can be defined as the positive evaluation that a person makes of his life in general terms, or of particular aspects (family, studies, work, health, friends, free time) (^{Diener, 1994}; Diener, Suh, Lucas, & Smith, 1999; ^{García-Viniegras & González, 2000}; ^{Liberalesso, 2002}). This satisfaction arises through a transaction between the individual and his micro-and macrosocial environment, in addition to the current and historical elements of this environment (including material objective and social conditions) that offer certain people opportunities for personal development. It is an experience linked to the present with projection into the future, as well as the balance between these two periods (García-Viniegras & González, 2000). ^{Gilman, Easterbrooks and Frey (2004)} reported that the relevance of studying of life satisfaction is based on this concept transcending the immediate effects of life events (Heady & Wearing, 1992) and moods (^{Seidlitz, Wyer, & Diener, 1997}); it is less ephemeral than the affective state (Diener & Diener, 1995) and more stable over time (Diener, 2000; ^{Eid, & Diener, 2004}). Ultimately, life satisfaction is not simply a result of other individuals' psychological states, but it is also influenced by important psychosocial relationships (^{Galindez & Casas, 2011}).

In Central and Latin America, including Chile, life satisfaction has tended to be assessed using the Life Satisfaction Scale by ^{Diener, Emmons, Larsen and Griffin (1985)} and adaptations of it. This scale is guided by the overall assessment of the concept, as reports of studies on the evaluation by domains are found to be insufficient in these cultures. In light of previous findings, approaching the construct of life satisfaction including its various domains contributes to a fuller understanding of the variables that affect it, which requires instruments that include a multidimensional measurement. At the same time, its measurement in university students is relevant given that they are at a complex stage of life, associated with the resolution of tasks not comparable with other times in life (^{Gomberoff, 2003}).

According to previous studies with university students, the level of life satisfaction could affect a variety of life domains, such as eating behavior patterns (^{Schnettler, Miranda, Sepúlveda, & Denegri, 2011}), physical and social well-being (^{Zubieta & Delfino, 2010}), self-concept (^{Zegers, Rojas-Barahona, & Forster, 2009}), future time perspective (^{Martínez, 2004}), among other variables.

Negotiating their way through university is a complicated experience for many young people, where various competencies and abilities are put into play to achieve the goals of the chosen professional training, facing a myriad of stressors. These include responding to greater academic demands, changes in social support networks, demands for greater autonomy and prominence, vocational or academic failure, living alone for the first time if they come from another city, among others (^{Micin & Bagladi, 2011}). Accordingly, their levels of life satisfaction must be identified early, since this stage is a precise space to change negative behaviors (^{Florenzano, 2005}; ^{Gomberoff, 2003}; ^{Martínez, 2007}; Micin & Bagladi, 2011).

According to previous studies with university students, the level of life satisfaction could affect a variety of life domains, such as eating behavior patterns (Schnett-ler, Miranda, Sepúlveda, & Denegri, 2011), physical and social well-being (^{Zubieta & Delfino, 2010}), self-concept (^{Zegers, Rojas-Barahona, & Forster, 2009}), future time perspective (^{Martínez, 2004}), among other variables. Differences according to gender and socioeconomic level have also been observed. In this latter case, both women (Zubieta & Delfino, 2010; Zubieta, Muratori, & Fernández, 2012) and those at the highest socioeconomic levels (Martínez, 2004; Zubieta & Delfino, 2010) have a greater overall life satisfaction than their comparison groups.

^{Huebner (1994)} contributes to the study of this construct with the Multidimensional Students Life Satisfaction Scale (MSLSS) by revealing the need to know the different levels of satisfaction related to specific dimensions in greater depth. This instrument is designed to provide a multidimensional profile of the youth population, originally schoolchildren, with respect to five domains (family, studies, friends, neighborhood and self), and it also assesses overall life satisfaction. The domain Family refers to satisfaction with the student's relationships with family members and the relationships the family members have with each other; the domain School refers to satisfaction with different aspects of the school environment, such as interest in school, learning and school activities; the domain Friends seeks to measure satisfaction with the relationships that students maintain with their peers; the domain Neighborhood evaluates satisfaction with the surrounding spaces and with the people in the neighborhood; and finally, Self refers to the satisfaction felt with oneself and the opinion of others.

The MSLSS was reviewed (^{Huebner, Laughlin, Ash, & Gilman, 1998}) and generally showed acceptable internal consistency coefficients for all the domains and globally. Additionally, in subsequent studies conducted to evaluate its psychometric properties in its English version and in the original target population, stability was established in terms of reliability and support for the multidimensionality (^{Greenspoon & Saklofske, 1 997; Greenspoon & Saklofske, 1998}; Gilman, Huebner, & Laughlin, 2000).

In Chile, only recently has the MSLSS come into use, applying the instrument to children and adolescents (^{Chavarría & Barra, 2014}; ^{Oyanedel, Alfaro, & Mella, 2015}; ^{San Martin & Barra, 2013}). Although ^{Schnettler et al. (2014)} evaluated the scale preliminarily in university students and there was little evidence of its psychometric behavior in the country, this study was designed to describe multidimensional life satisfaction in a broader sample of university students from northern, central and southern Chile. Specifically a) to examine the psychometric properties of ^{Huebner's Multidimensional Life Satisfaction Scale (1998)} in its application to university students, b) to describe the levels of life satisfaction in students, c) to determine whether there are differences among students about life satisfaction according to gender, and d) to determine if there are differences among students about life satisfaction according to socioeconomic level.

In this sense, in relation to the first objective, it was expected that psychometric properties would be found adequate, that is why the following hypotheses are proposed: H1: Internal consistency greater than 0.70 will be observed between elements of global scale and domains. H2: The proposed five factors structure which compose the scale will have well-adjusted with the data. , Regarding the second objective, previous studies have reported that university students assess their lives positively (^{Barraza & Ortiz, 2012}; ^{Domínguez-Espinoza, Salas, Contreras, & Procidano, 2011}; ^{Gómez, Villegas, Barrera, & Cruz, 2007}; ^{Martínez, 2004}; ^{Schnettler et al., 2011}; ^{Moyano, Cornejo, Carreño, & Muñoz, 2013}; Schnettler et al., 2013; ^{Zubieta & Delfino, 2010}; Zubieta et al, 2012), which provides the descriptives hypothesis: (H₃) A level above the central point of the scale of life satisfaction by domains will be observed in university students. In relation to the third and fourth objectives, the current state-of-the-art reveals differences according to gender and socioeconomic level where women (Zubieta & Delfino, 2010; Zubieta et al., 2012) and the highest socioeconomic levels (Martínez, 2004; Zubieta & Delfino, 2010) present more positive assessments than their comparison groups, and it is therefore expected that: (H₄) In female university students, the level of life satisfaction by domains will be greater than in male university students. (H₅): The higher the socioeconomic level of the university students, the greater will be the level of life satisfaction globally and by domains.

Method

Results

For this section, the results of the psychometric properties of the scale corresponding to the first objective will appear first; then the findings will be discussed according to the rest, and to contrast the hypotheses.

Psychometric properties

In terms of the reliability of the scale and its domains, the total test reveals a Cronbach's alpha coefficient of 0.90, and by domain it showed the following values: 0.89 for Family, 0.86 for Friends, 0.84 for Neighborhood, 0.79 for University, and finally 0.86 for Self. With these results, H1 was confirmed, indicating that there would be internal consistency greater than 0.70 between elements of the global scale and by domains, being consistent with the previous studies, and reaching an adequate parameter of reliability.

In terms of the validity of the scale according to the evidence-based on the internal structure and considering that all the exploratory factor analyses (EFA) run in previous studies had shown a stable structure of 5 domains, a CFA was conducted on the measurement model. This revealed adequate goodness-of-fit indices with the following results: X²(730) = 1 961.421, X²/df ratio = 2.69, RMSEA = 0.074, CFI = 0.918, TLI = 0.913. According to the considerated cut-off points, four of the indicators are in acceptable ranges to indicate a proper adjustment, being X2 the one which doesn't have an expected behave, however this indicator is susceptible to the number of participants, since if the sample is big enough, any model will be unsuitable, because increasing the sample size increases the power to detect statistically significant discrepancies which are irrelevant from a practical point of view (^{Abad et al., 2011}). From the data obtained from this analysis on the corresponding matrix, the model of five correlated factors was accepted (see Figure 1), with 40 items in total (see Table 1 ), confirming H2, which indicated than the proposed structured will be well-adjusted with the data, being consistent with previous studies where was evaluate the instrument's internal validity.

Note: *p <.001

Figure 1 Model obtained from the CFA and correlations between factors. 

Table 1 Non-standardized factorial loads of the items by factor. 

Note: ־p <.001University students' levels of life satisfaction.

In terms of the reliability of the scale and its domains, the total test reveals a Cronbach's alpha coefficient of 0.90, and by domain it showed the following values: 0.89 for Family, 0.86 for Friends, 0.84 for Neighborhood, 0.79 for University, and finally 0.86 for Self. With these results, H1 was confirmed, indicating that there would be internal consistency greater than 0.70 between elements of the global scale and by domains, being consistent with the previous studies, and reaching an adequate parameter of reliability.

Additionally, with regards to obtaining evidence of the relation with others variable, the CFA of the previous model and the satisfaction factor with the global life according to the data of the scale of Dierner et al. (1 985) revealed adequate goodness indices, with the following results: X² (930) = 2204,779, p = 0.001, X² / gl = 2.37 ratio, RMSEA = 0.067, CFI = 0.921, TLI = 0.916, likewise, four of the indicators are acceptable to indicate a good adjustment ranges, with X² escaping as expected, previously justified. It can be mentioned that in the six factors, mean and significant correlations are observed as shown in Table 2. The scale of ^{Diener et al. (1985)} correlates with all domains of the ^{Huebner et al. (1998)} giving evidence of the convergent validity of the latter, and therefore confirming the hypothesis 3 that indicated that there is an association between the values of the scales that measure the same construct.

Table 2 Standardized correlations between domains of MSLSS y SWLS. 

Note: *p <. 001

The results related to the second objective show that the comparison of the individuals' averages and the central point of the total scale and its dimensions reveals statistically significant differences with a large effect size in Global Satisfaction and the domains Family, Friends, University and Self, whereas the dimension Neighborhood shows a medium effect size, as illustrated in Table 3. That allows to accept the descriptive hypothesis 3

Table 3 Comparison of means with the central point of the scale (globally and by domains). 

Note 1: *p <.001

Note 2: gl = degrees of freedom

Differences in life satisfaction according to demographic variables

For the third objective, oriented toward detecting differences according to the demographic variable gender, the student's t-test indicated significant differences in the domains Overall Satisfaction, Friends and University. In Overall Satisfaction (t(303)= -2.593, p = 0.001, d = 0.30) women (W = 4.69, SD = 0.54) reported higher levels of satisfaction than men (W = 4.51, SD = 0.59) (with a small effect size). Likewise, in the domain Friends (t(303)= -3.078, p = 0.002, d = 0.36) women (W = 5.23, SD = 0.69) presented higher means men (W = 4.96, SD = 0.78) (with a small effect size), and in University (t(303)= -3.501, p = 0.001, d = 0.41), again women (W = 4.64, SD = 0.73) had higher levels of satisfaction than men (W = 4.33, DS = 0.78) (with a small effect size). This result means that hypothesis 4 is accepted.

For the fourth objective and the variable socioeconomic level, The Kolmogorov-Smirnov test was performed, giving a test statistic of 0.163, which being greater than 0.05 shows a normal distribution. The linear ANOVA test results showed significant differences only in the domain Neighborhood (F(1) = 10.617, p = 0.001, η² = 0.033) (with a moderate effect size, 33%), presenting a fit to the straight line according to SEL. That doesn't allow to accept the hypothesis 5.

Discussion

In recent years, research into quality of life has widened its areas of focus, including the measurement of subjective well-being as an important part of the assessment that people make of their current situation. In this context, evaluation of the psychometric properties and specifically the analysis of reliability and a CFA of the MSLSS, an instrument with ample theoretical support and good behavior through the years of use in other contexts and countries, is a relevant methodological contribution. In this study, the instrument showed adequate psychometric behavior, which contributes with a new valid and reliable tool at the level of construct validity for the multidimensional measurement of life satisfaction in young Chilean people at university.

University students reported a high level of life satisfaction in general and by domains, which is consistent with results from other studies conducted on this population with other scales (^{Barraza & Ortiz, 2012}; ^{Domínguez-Espinoza et al., 2011}; ^{Gómez et al., 2007}; ^{Martínez, 2004}; ^{Moyano et al., 2013}; ^{Schnettler, et al., 2011}; Schnettler et al., 2013; ^{Zubieta & Delfino, 2010}; Zubieta et al., 2012). This implies that despite the difficulties inherent in youth development, the stress associated with academic life and the changes which students face seemed not to affect the assessments the students made of their life satisfaction (^{Páez, Bilbao, & Javaloy, 2008}). In terms of the domains that seem to influence the assessment of students' life satisfaction, it was observed that although all of these had positive values, the highest was in the domain Friends. This could be related to a characteristic of youth centered on the need to maintain positive relations with other people, preserving stable social bonds and having friends one can trust (^{Díaz et al., 2006}). By contrast, the lowest mean was in Neighborhood, which may be associated in part with the concrete comparison between the neighborhood of origin and the one of current residence, since a high number of students must leave their family surroundings and migrate to cities far from their homes to attend university. Also, it is in part to the very fact of having to leave their familiar surroundings where they raised and have bonds and go somewhere that seems more distant to them (^{Micin & Bagladi, 2011}; ^{Denegri et al., 2014}).

In terms of the relation with sociodemographic variables, the analyses showed that women presented higher overall satisfaction and in the domains Friends and University than the men. This may be associated with certain characteristics of gender identity and the differential socialization between the sexes. In this process, women are socialized early on to be more social, friendly and oriented toward others (^{Aguilar, Valdez, González-Arratia, & González, 2013}).

In terms of the relation with sociodemographic variables, the analyses showed that women presented higher overall satisfaction and in the domains Friends and University than the men. This may be associated with certain characteristics of gender identity and the differential socialization between the sexes. In this process, women are socialized early on to be more social, friendly and oriented toward others (^{Aguilar, Valdez, González-Arratia, & González, 2013}). This also implies a high value placed on the relationship with the social surroundings as a facilitator of goals and stability and confidence on a relational level (^{Zubieta & Delfino, 2010}; Zubieta et al., 2012). On the other hand, increased access to education and the female presence in the labor market are achievements of recent decades (^{Vega-Robles, 2007}), and this could generate a more positive assessment by women as they become part of this change.

By contrast, the results do not support the notion that life satisfaction increases when SEL rises. This information may be explained in part by studies suggesting that the relationship between income and subjective well-being appears instead to be curvilinear, and that once basic needs have been satisfied, there is no necessary relationship between the wealth of nations and the sense of well-being, nor between this and the degree of individual wealth (^{Lane, 2000}; ^{Moyano et al., 2013}; ^{Veenhoven, 1991}). It may be hypothesized that when basic needs are satisfied, individuals move towards a post-materialistic stage in which the core aspiration is self-realization (^{Inglehart, 1990}).

To sum up, on the one hand the findings here confirm the results of previous studies and therefore contribute to the consolidation of the framework of variables that affect life satisfaction during youth (^{Barraza & Ortiz, 2012}; ^{Denegri et al., 2014}; ^{Moyano et al., 2013}; ^{Schnettler et al., 2013}; ^{Zubieta & Delfino, 2010}; Zubieta et al., 2012). On the other hand they provide more specific background information in relation to the student population which can serve as a reference for decision-making in public policies and intervention in mental health. This is established by considering that students with a greater subjective well-being will be better able to orient their life plans, make them concrete, and thus achieve a greater personal potential, contributing to their societies' creative capacity, where progress and well-being depend on the health and development of those who comprise it (^{Díaz, Carús, Macossay, & Barbosa, 2010}).

As a limitation, the study characteristics and the nature of the sample prevent generalization from university students to the youth population at large although, the results allows to accept the descriptive hypotheses; therefore, these results are only applicable to individuals with characteristics similar to those of the respondents, i.e. university students from the mentioned universities, it is necessary to consider future studies with representative samples that will allow the generalization of the results to the entire population of university students.

A future line of enquiry may be to delve more deeply into the relations and meanings that underlie the level of satisfaction in the domain Neighborhood and its relation to the migration of students for study purposes. Based on the obtained results and considering future researches with representative samples, it is expected to contribute with data to the residence policies for university students.