Optimism and university context

The scientific literature has identified variables relevant to their relations with optimism within the university context. As to the students’ demographic segments, the differences of optimism according to sex have been explored, without the result pointing consistently in one direction or another. Thus, the results of ^{Puskar et al.’s (2010)} study evidence less levels of optimism in women than in men, while in other works significant differences have not been found (^{Huan, Yeo, Ang, & Chong, 2006}; ^{Patton, Bartrum, & Creed, 2004}). Regarding age, ^{Londoño’s (2009)} study on university students showed a positive correlation of this variable with optimism. In the same line, the results of ^{Zou et al.’s (2016)} study observed a positive linear relation with age. That is to say, the older the subject, the greater the degree of optimism. These results empirically support the proposals of authors such as ^{Seligman (2003)} and ^{Martin (2002)}, who show in their theories that optimism is a characteristic learnt in the adult age.

The relation between optimism and other psychosocial variables measured in students has also been analyzed. It has been found that optimism, in connection with high levels of social support, increased the subjective well-being of university students (^{Marrero & Carballeira, 2010}) and enabled them to successfully face those demands which are characterized as being stressful (^{Fernández-González, González-Hernández, & Trianes-Torres, 2015}). ^{Chapman & Chi’s (2017)} proposal shows how perceived social support acts as a mechanism for optimistic people to more actively face the challenges which emerge on a daily basis.

On the other hand, optimistic future expectations in university students are associated with high self-esteem (^{Montgomery et al., 2003}). These relations have been moderately evidenced in different empirical works within the university context (^{Kapikiran & Acun-Kapikiran, 2016}; ^{Liu et al., 2017}). A positive correlation between self-esteem and optimism has also been found in samples with secondary education students (^{Monzani, Steca, & Greco, 2014}).

Another variable which has centered the attention of researchers is perceived self-efficacy. Analyzing the adaption to the university context in first-year students, ^{Morton, Mergler & Boman (2014)} note the direct relation between self-efficacy and optimism, as well as both variables being associated with university adaptation. In the same sense, ^{Jovanović & Gavrilov-Jerković (2013)} found positive correlations between perceived self-esteem and dispositional optimism in a sample made up of university and secondary school students. ^{Pu, Hou & Ma (2017)} support the relation between self-efficacy and dispositional optimism, although they argue that it is self-efficacy which fosters an optimistic view. Inversely, low self-efficacy would explain a person’s pessimistic perspective.

The connection between optimism and academic achievements is another study area which has caught the attention of researchers. ^{Monteiro et al. (2008)} found a positive and significant relation between both variables, concluding that those students who have high levels of optimism showed a higher academic performance when they finished the academic term. In ^{Chemers et al.’s (2001)} study, optimism and self-efficacy were factors taken into account to explain the academic performance of first-year students. The results clearly indicated that both variables influenced academic performance and the students’adjustment to the new context.

The literature has consistently pointed out the power of optimism as a variable which favors first-year students’ university adaptation. For this reason, different authors have carried out empirical works on optimism in university students whose results recommend the design and implementation of programs that foster optimism (^{Fernández-González et al., 2015}; ^{Monteiro et al., 2008}; ^{Perera & McIlveen, 2014}). With a view to specifying some of the aspects which give substance to intervention programs or strategies in this area, it is interesting to identify the factors that are associated to a greater extent with optimism. This is why, this work has proposed the aim of analyzing the relevance of different psychosocial and academic variables in explaining the optimism with which students face the start of their university studies.

Design

A quantitative ex-post facto study was carried out, adopting a descriptive correlational design according to the typology indicated in ^{Hernández, Fernández & Baptista (2003)}. As to the time dimension, the design is transversal, given that the data collection has been carried out at a single moment.

Participants

The population considered in this study was located in the Spanish context. Specifically, it was made up of students who accessed a university degree program for the first time in the University of Seville. Through a non-probabilistic sampling procedure and addressing the criteria of accessibility, the sample consisted of 750 first-year university students of five university teaching areas (see Table 1). Neither repeaters of the first course-year, nor those who had finished another university degree or had begun but not completed other degrees were included in the sample. 57.5% were women and 42.5% men. The average age was 18.94 years old (S.D.=2.20).

Table 1 Sample of first-year university students 

In accordance with the analysis procedures which are described in a later section, in the initial sample two subsamples of students were identified, characterized by high optimism and by low optimism groups of 247 and 248 students, respectively.

Instruments

Dispositional optimism was considered as the dependent variable, measured through the Life Orientation Test (LOT-R), in its revised version (^{Scheier, Carver, & Bridges, 1994}). This test is made up of 10 items with a Likert-type answer scale, of which three refer to optimism, three to pessimism and four items are included that serve to make the content of the test less evident and are not taken into account for scoring purposes. The answers range between totally disagree (1) and totally agree (5). For this instrument, Scheier, Carver & Bridges (1994) obtained a reliability estimated by a Cronbach alpha of .78. In this work, the instrument showed an alpha value of .75 for the total sample. This value can be considered acceptable, taking into account that the measure obtained with the LOT-R test is supported only in six items. Scheier, Carver & Bridges (1994) proposed a mono-factorial structure, considering the optimism-pessimism construct to be a sole feature with two poles. The adaptation to Spanish used in this study is that employed by ^{Ferrando, Chico & Tous (2002)}. Contrary to the unidimensional structure attributed originally to the LOT-R, the authors cited bring together different studies which propose the instrument’s bidimensionality and note the existence of two factors which correspond to optimism and pessimism. With a view to clarifying this issue, this study carried out an exploratory factor analysis (EFA) to identify the number of dimensions underlying the scale, followed by a confirmatory factor analysis (CFA). 50% of the sample was randomly selected to carry out the EFA and the CFA was done with the other half. The suitability of the matrix of correlations to perform this analysis has been confirmed by Bartlett’s sphericity test (χ2=502.49; p=.000) and by Kaiser-Meyer-Olkin’s measure of sampling adequacy (KMO=.80). Following the recommendations of ^{Lloret-Segura, Ferreres-Traver, Hernández-Baeza & TomásMarco (2014)}, for the extraction of factors the maximum likelihood method was used, previously checking that the values of asymmetry (included in absolute values between .09 and .90) enable trusting the normality of the variables. In the factorial solution a single dimension was retained, taking into account that the distance between the selfvalue of the first factor (2.757) and of the second (.957) is considerably greater than that registered between this and the self-values for the successive factors (.770 in the third, .566 in the fourth). The retaining of a sole factor is also a result compatible with the initial theoretical criterion and, moreover, the six items present high saturations (see Table 2).

Table 2 Factor loadings of items of the Life Orientation Test (LOT-R) 

The CFA corroborated that the single factor model satisfactorily fit the data, confirming the unidimensional structure of the scale. The goodness of fit indexes showed acceptable values (χ² = 53.263, p < .000, χ²/gl = 9, GFI= .954, CFI= .916) or very close (RMR= .063) to the values which would permit stating the goodness of the model (^{Byrne, 2001}). Consequently, the dispositional optimism feature was used in this analysis as a single score.

As independent variables, self-esteem, self-efficacy, perceived social support, previous academic support and the demographic features of sex and age were included. Self-esteem was measured through ^{Rosenberg’s Self-Esteem Scale (Rosenberg, 1965}), adapted to Spanish by MartínAlbo, Núñez, Navarro & Grijalvo (2007). These authors used the instrument twice with a sample of university students, confirming its unidimensional structure and obtaining Cronbach alpha reliability measures of .85 and .88 in them. For this study, the instrument gave an alpha value of .87. This instrument has 10 items, where five are enunciated positively and five negatively. The answers of the subjects are expressed through a 4-point Likert scale ranging from totally disagree (1) to totally agree (4).

To measure the self-efficacy variable, the General SelfEfficacy Scale (^{Baessler & Schwarcer, 1996}) was used. This instrument measures a single construct through 10 items which adopt the format of a Likert-type scale with four degrees of agreement ranging from totally disagree (1) to totally agree (4). The Spanish adaptation of this instrument, used in this study, is that proposed by ^{Sanjuán, Pérez & Bermúdez (2000)}, which obtained a reliability of α=.87, close to the value shown in this study: α=.84.

The measurement of the variable perceived social support was carried out with the Scale of Family and Friends Support (AFA-R), by ^{González-Ramírez & Landero-Hernández (2014)}. This is made up of 14 items referring to the support which is received from the family or friends. The items are answered using a frequency scale with five levels, where 1 corresponds to never and 5 to always. Applying it to university students, its authors found a reliability of .92 for the instrument. Like the results of internal consistence, this study showed an alpha value that was also high: α=.89. The variables age, sex and previous academic performance were obtained from questions the respondents were asked when filling out the measurement instruments. In the case of academic performance, they were asked to record the grade obtained in the University entrance test with which they accessed the degree program that they were pursuing.

Procedure

The data collection took place in the second four-month term of the 2016/2017 academic year, specifically during April 2017. The teachers who were teaching subjects of the first year of university in different degree programs were contacted via e-mail. The questionnaires were applied in the classroom and lasted approximately 15 minutes, including the presentation of the study, the explanation about the data collection instruments and the filling out of the questionnaires by the students. In accordance with the Code of Good Practices in Research of the University of Seville (https://investigacion.us.es/investigacion/apoyo), the anonymity of the participants was guaranteed. They answered the questions voluntarily after having received information about the study’s proposal and nature. The students did not receive any academic benefits from participating in the study. It was also guaranteed that those who declined to take part in the study would not suffer adverse consequences due to this. The data collection was carried out by the researchers themselves.

Data Analysis

For the analysis, the answers to the items used in the scales of self-esteem and self-efficacy were codified from 1 to 4, while the five modalities of answers for the items related to perceived social support and optimism were codified from 1 to 5. In the case of items in a negative sense, these values were inverted, so that a high score always corresponds to a high level in the feature measured. A very low incidence of lost values was registered (less than 1% in all the variables), so as a general procedure it was opted for working with the cases for which complete information was obtained for the variables involved in each analysis (casewise procedure). The analysis began with the calculation of the descriptive statistics and internal consistency indices, both for the measure of dispositional optimism and for the rest of the variables considered in the study. The sample was divided into three groups using the factorial scores in optimism, derived from an EFA similar to that applied to examine the instrument’s unidimensionality, although this time applied to all the participants. To do so, the terciles distribution was used, enabling the differentiation of optimistic students, students who did not opt for optimistic or pessimistic positions and clearly pessimistic students. Retaining those who were in the group of greater and lesser optimism, the binary logistic regression analysis allowed an exploration of the extent to which the psychosocial and academic variables studied contribute to the differentiation between the optimists and the pessimists. The SPSS v.23 and AMOS v.24 statistical softwares were used.

Descriptive study of the variables

Just like the demographic variables of age and sex, already described when presenting the participants in the study, information was gathered about the mark obtained in the university entrance tests. The average was 10.9 (SD=1.69) on a score scale with a maximum of 14. In the case of the variables optimism, self-esteem, self-efficacy and social support, the total scores in each of the instruments used for their measurement were divided by the number of items, so that values between 1 and 5 were expressed for the variables optimism and social support, and values from 1 to 4 for self-esteem and self-efficacy. Table 3 shows that the university students perceive strong support from their family and friends, reflected in an average score of 4.22. They tend to show themselves as being optimistic, with a score of 3.39, slightly above the value 3 which is the central point of the scale, and have high levels of selfefficacy and self-esteem, with values above 3 on a scale which goes from 1 to 4.

Table 3 Descriptive statistics 

Explanation of optimism from psychosocial and academic variables

Binary logistic regression analysis was used to value to what extent students’ optimism is associated with the psychosocial and academic variables considered in this study. The binary dependent variable is optimism, differentiating between optimistic students (value 1) and pessimistic students (value 0). As independent variables, sex, age, prior academic performance, self-esteem, self-efficacy and perceived social support were included. According to Hosmer & Lemeshow’s test, the model constructed with these variables adequately fits the values observed, as the chi-squared contrast statistic =6.837 is associated with a degree of significance of p=.554. This permits the maintenance of the null hypothesis of equality between the values observed and the values expected according to the model. The model’s goodness has also been estimated from the R squared of Cox & Sell (R²=.384) and that of Nagelkerke (R²=.512), whose values indicate a good explanation of the dependent variable by the independent variables. Furthermore, its explanatory power is supported by the high percentage of students correctly classified as optimistic and pessimistic. This correspond to 81.5% and 73.3%, respectively, with a global percentage of successful prognostics for 77.5% of the students.

Table 4 indicates the results of the logistic regression analysis, showing that the variables with a greater degree of association with dispositional optimism are self-esteem and self-efficacy (p<.001). The effect of the variables perceived social support (p=.020) and, to a lesser extent, previous academic performance (p=.047) is also significant. As to the remaining variables, the demographic features of sex and age have not turned out to be relevant in the explanation of optimism (p>.05).

Table 4 Coefficients, odds ratios and Wald for the effects on optimism 

¹ Odds ratio associated with the increase of one unit in the independent variable.

² Reference category: man.

Addressing the odds ratios, the importance of each of the variables in the differentiation of optimistic and pessimistic university students can be quantified. Its value indicates how much more likely it is to have an optimistic disposition for each unit in which the score registered in the independent variable considered increases. Thus, the unit increase of the score in self-esteem means approximately multiplying by thirteen the probability of being an optimist (odds ratio 13.093), while augmenting a point in self-efficacy multiplies almost fivefold the likelihood of the student being optimistic (odds ratio 4.714). Less important are the effects of prior performance or social support (odds ratios of 1.175 and 1.672, respectively).