Application of the Future Expectation Scale for Adolescents ( FESA ) in Brazil

This exploratory study analyzed the factor structure of the Future Expectation Scale for Adolescents (FESA) in a Brazilian sample of emerging and young adults to determine its utility for research and clinical use in Brazil. The sample included 547 young adults, ranging from 18-29 years old (M = 22; SD = 3.9), from different socioeconomic status and 351 (64.2%) females. Results of Confi rmatory Factor Analysis revealed that the original model found for the FESA did not fi t the Brazilian sample well. Therefore, a subsequent Exploratory Factor Analysis was performed with the 24 self-reported FESA items. Results demonstrated a cohesive factor structure for the FESA, but the factor structure was different from the original. For instance, items belonging to the original factor ‘Marriage and Family’ loaded better with the original factor ‘Children’s Future,’ which was renamed ‘Children and Family.’ The original factor ‘Church and Community’ was slightly different and was renamed ‘Church’ to be consistent with the Brazilian context. The factor structure of the FESA was consistent and the measure was found to be useful for further studies in this area.

and future expectations (McWhirter & McWhirter, 2008;Wyman, Cowen, Work, & Kerley, 1993).In the current study, the term 'expectation' will be used, and is defi ned as 'believing that something will happen' (Kernerman, 2007).Ramos, Seidl-de-Moura, and Pessôa (2013) identifi ed three central theoretical orientations relevant to the future goals of young people: (a) Future Time Perspective Theory, (b) Self-determination Theory, and (c) cognitive and neural processes related to the development of goals for the future.Future Time Perspective Theory investigates the current infl uence of future goals (Lens, 1993;Nuttin, 1983;Schmitt, 2010) and describes future expectations as (a) short-term, when individuals establish more immediate future goals (for example, "I will apply for a job tomorrow"); (b) extended, when individuals establish more distant future goals (for example, "I will fi nd a job in the service industry as soon as I graduate from high school"); and (c) long-term, when individuals establish long term future goals, which will be pursued for years (for example, "One day I will own my own business"; Lens, 1993).Self-determination Theory investigates intrinsic and extrinsic motivations and the factors that promote these motivations (Deci & Ryan, 2000).Studies in the fi eld of education highlight the role of the teacher in promoting intrinsic motivation, and the infl uence of intrinsic motivation in students' learning engagement, persistence, and use of educational strategies (Guimarães & Boruchovitch, 2004;Vansteenkiste, Soenens, Verstuyf, & Lens, 2009).Future Time Perspective Theory and Self-determination Theory overlap.For instance, one investigation concluded that individuals whose future goals present an intrinsic motivation are more autonomous and more persistent in taking actions towards their future goals (Vansteenkiste et al., 2009).The third theoretical orientation focuses on the similarities between cognitive and neural processes involved when stimulating thoughts about past and future goals (Ramos et al., 2013).Within this perspective, a study brought evidences that decision making under uncertainty and trial-by-trial feedback highly stimulates activity of the parietal lobule, what was dissociable from general attention in the brain (Vickery & Jiang, 2009).Using data of pupillary response, another study indicated that future-oriented people invested more cognitive resources and response accuracy than presentoriented people in reaction time, when exposed to items with a temporal orientation towards the future (Nowak, Milfont, & Van Deer Meer, 2013).Other research found that age-related changes in affective brain regions and circuitry can infl uence how people experience affect during decision-making and how people recruit affect to make choices (Weierich et al., 2010).Future orientation affects an adolescent's defi nition and exploration of planning for and commitment to goals that guide their developmental course in three domains: (a) higher education; (b) work and career; and (c) marriage and family (Seginer, 2008).Positive beliefs in relation to the future are associated with long term plans for higher education, positive thoughts in relation to work, better social and emotional adjustment at school, and self-perceptions of competency (Catalano et al., 2004).Individuals having extended and long-term future perspectives manifest a stable and healthy model of planning, are able to delay gratifi cation, and are more motivated and committed to pursue their goals, especially in comparison to individuals with short-term future goal perspectives (Bilde, Vansteenkiste, & Lens, 2011;King & Gaerlan, 2014;Lens, 1993).This suggests the potential protective role of positive, long-term future orientations in youth outcomes (McWhirter & McWhirter, 2008).In addition, adolescents will invest in the future to the extent that they perceive reward and return on their investment (Seginer, 2008).
Associations between future expectations and positive outcomes in development have been identifi ed in studies with children.Positive expectations of children in relation to the future were predictors of adjusted emotional regulation and an internal locus of control in later years, acting as a protective factor in stressful situations (Wyman et al., 1993).An investigation in Finland evaluated adolescent future perspectives in relation to three basic processes: motivation; planning; and evaluation (Nurmi, 1991).The results showed that the goals and interests of adolescents were affected by their family context, predominantly in activities associated with the end of adolescence and beginning of adulthood, and that worries about the future tended to decrease with age.Another study revealed that highly challenging situations infl uenced the future orientation of adolescents, and these situations were mediated by hope and moderated by four factors: cultural orientation, developmental stage, interpersonal relations, and intrapersonal characteristics (Seginer, 2008).
When considering the topic of future expectations during the transition to adulthood, the infl uence of culture, developmental stage, and personality should be taken into account (Seginer, 2008).Asian and Latin American young people differ from European young people with respect to a cultural tradition of family assistance, support, and respect.The former include the family as part of their future perspectives, which does not happen with Europeans.At the same time, life style changes (e.g., the opportunity to attend college) for members of diverse ethnic groups can alter the role of family and its infl uence in their transition to adulthood (Fuligni, 2007).
Considering the infl uence of future expectations in the current life of individuals (Catalano et al., 2004;Lens, 1993;Nurmi, 1991;Seginer, 2000Seginer, , 2008;;Wyman et al., 1993;Vansteenkiste et al., 2009), this study selected the Future Expectation Scale for Adolescents (FESA, McWhirter & McWhirter, 2008) to be analyzed in a Brazilian sample.Other measures about future expectation exists, such as the Future Time Perspective (Carstensen, 1996) and the Future Orientation Questionnaire (Seginer, 2000(Seginer, , 2008)), but they only see future expectation as a function of anticipated achievements or desired behavior changes.Therefore, the FESA was chosen in the present study, because it adds an orientation of future expectations towards specifi c domains, namely, "work and education", "children's future", "marriage and family", "health", and "church and community".These specifi c domains were defi ned based on the literature in the area of future expectations (McWhirter & McWhirter, 2008).
The FESA was built in the context of a larger study about risk and protective factors among Chilean adolescents.The items originate from diverse sources of information, such as a review of risk and protective factors among Chile and U.S. adolescents; and readings on the Chilean historical, political, and cultural context.Participants in this study were 389 7th-to 12th-grade students in a poor and working-class neighborhood, in urban central Santiago, Chile (McWhirter & McWhirter, 2008).
The FESA (McWhirter & McWhirter, 2008) items assess the extent to which participants agree with each statement about their future lives.The referential statement 'when I'm an adult...' is followed by future possibilities, such as 'I will achieve the level of education that I desire' and 'I will have a healthy diet'.Answers are based on a 5-point Likert scale, ranging from 'I do not believe this at all' to 'I certainly believe this'.
First, authors examined the FESA through an exploratory factor analysis.Second, they examined its construct validity, by exploring its full-scale and subscale correlations with other theoretically related measures.They found signifi cantly and positively correlation with each of the protective factors included in the analysis (e.g.Hope scale and Children's School-Related Self-Effi cacy); and negative correlations with risk factors were small (e.g., Barriers Diffi culty, r =−.12; p < .05;McWhirter & Mc-Whirter, 2008).
Based on their initial fi ndings, the original measure was revised, specifi cally, four items with smaller pattern coeffi cients were deleted, three items were added, and one item was revised.This resulted in a 24-item measure with possible total scores ranging from 24 to 120.For all subscales, higher scores represent the more positive future expectations associated with the respective domains of work and education, marriage and family, children's future, health, and church and community (McWhirter & McWhirter, 2008).Internal consistency reliabilities for the revised version and subscales are presented in Table 1 (McWhirter & McWhirter, 2009).Despite the fact that the measure was developed for assessing adolescent future expectations, the FESA addresses life domains also very salient to emerging adults.There is a growing international literature about an extension of the process of transitioning to adulthood observed in a number of industrialized societies (Arnett, 2007;Barry, Nelson, Davarya, & Urry, 2010;Douglass & Col-lege, 2007;Facio, Resett, Micocci, & Mistrorigo, 2007;Fuligni, 2007;Galambos & Martínez, 2007;Leonard & Scott-Jones, 2010).
The factors measured by the FESA (Work and Education, Children's Future, Marriage and Family, Church and Community, and Health) are relevant to young people outside of the Chilean context.The factor structure of the measure may vary in different countries as a function of cultural or contextual factors.Understanding such differences is an important fi rst step in using this tool to investigate future expectations in Brazil, and will also serve as a reference to compare the measure's factor structure in different countries.Thus, the present study aimed to analyze the factor structure of the Future Expectation Scale for Adolescents in a Brazilian sample.We expected that the measure may provide important information also about future expectations in young individuals from different countries.The FESA is a new scale and no other studies to date have used the measure.Therefore, the present study is one of the fi rst ones to analyze FESA factorial structure in a sample from another country.

Sample and Procedures
The sample included 547 Southern Brazilians, between 18 and 29 years old (Md = 22 years; IQR 1 = 19-26), 196 (35.8%) males, and 351 (64.2%) females, of low (n = 194, 35.5%) and high (n = 353, 64.5%) socioecomic status (SES).The SES criterion was based on the socioeconomic classifi cation from the Brazilian Association of Institutions of Market Research (Abipeme, 2008).This classifi cation attributes scores for "domestic comfort items" (e.g.washing machine, freezer, and television) and the household's level of education.In the present research, the criterion used to determine the head of household was the parent with the higher level of education.High SES in this study corresponds to scores above 58.The sample of individuals from differing SESs was accessed through different educational institutions: (a) two technology courses (n = 99); (b) three universities (n = 223); (c) two courses focused on preparing students from low SES to the university entrance exam (n = 50); and (d) two schools that work with young individuals and adults with limited or no previous education (e.g., people with writing and reading diffi culties and with solving basic math problems; n = 33).There were 142 participants, which accessed the questionnaire exclusively online, and they were not attached to a specifi c institution.Considering the total sample, 43.1% of participants were working full time (n=236); 9.1% part time (n=50); .4% occasionally (n=4); and 47.3% were not currently working (n=259).Regarding their educational status, 22.5% had less than High school (n=123); 60.5% had High school degree (n=331); 43.1% had College degree (n=236); and 17% had some Graduate school (n=93).
The study questionnaire and instruments were digitalized and applied by two methods: online (i.e., the participant was invited by e-mail to access an external website where the instruments were available) and in-person (i.e., the participant completed the instruments online in a computer lab with one researcher present).The in-person method was used to provide support to participants with limited education who could present diffi culties to understand the questionnaire.Research group members were available to answer their doubts during the application.In both methods, the instruments were self-administered.The average time for answering the instrument was one hour.
Translation of the measure took place in several steps.The fi rst step was an independent translation of the English version of the FESA into Portuguese by the fi rst author.The version was reviewed by two bilingual judges, who had no knowledge about the instrument.They compared the version in English to the version translated into Portuguese.Several changes were made based on suggestions by the judges.The revised version in Brazilian Portuguese was then backtranslated into English by a Professor living in U.S. who is fl uent in both English and Brazilian Portuguese.Next, the same version was reviewed and backtranslated into English by two bilingual judges, also fl uent in both English and Brazilian Portuguese.The backtranslated version was deemed similar to the original.

Ethical Considerations
The research project was approved by the Research Ethics Committee of the Federal University of Rio Grande do Sul (Protocol Number 2011014), institution where the investigation was developed.

Results
Confi rmatory Factor Analysis (CFA) was used to evaluate whether the original structure of the FESA's fi ve factors fi t to the Brazilian data.To take into account the categorical nature of the scale items, the Weighted Least Square Mean Variance (WLSMV) estimation method was employed in the Mplus software.For fi t indices, the Chi-Square (χ²), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation with 90% Confi dence Interval (RMSEA -90% CI), and Weighted Root Mean Square Residual (WRMR) were calculated.Goodness of fit criterion included a nonsignifi cant Chi-Square (p > .05),values of the CFI and TLI above .90or close to .95,values of the RMSEA close to or below .05(with below .08 an acceptable fi t) and values of the WRMR below .60 (Yu, 2002).
The number of factors extracted was based on the following criteria: (a) instrument theoretical background; (b) Kaiser (1960) criteria (the number of factors extracted is equal to the number of eigenvalues higher than one); (c) Scree plot; and (d) parallel analysis, which compares the eigenvalues found empirically with random eigenvalues (Hayton, Allen, & Scarpello, 2004;Horn, 1965).
Using PAF (Barlett's test of sphericity = 5952,566, df = 276; p < .0001;KMO = .88),six factors were extracted (i.e., eigenvalue > 1).These six factors accounted for 50% of the total variance.Communalities of the 24 items presented a range of .128(item 9) to .836(item 20).One of the six factors consisted of only two items (16 and 18), and item 18 cross loaded on two factors.The scree-plot test and the parallel analysis (see Figure 1) suggested a four factor solution as the most representative of the data.
Next we tested a 5 factor solution and compared this with the results of the parallel analysis.A fi ve factor solution is consistent with the original FESA fi ndings, and the 6 th factor was not viable.In addition, our review of the parallel analysis results showed that the four factor solution combined Factor three (Marriage and Family) and Factor fi ve (Children's Future).In the Brazilian sample, items about having children that originally loaded on Factor three in the Chilean sample loaded on Factor fi ve, thus, we determined that the most descriptive name for this factor was 'Children and Family'.This left, Factor three was composed of items related to the expectation of having a partner and getting married, but no items focused on having children and building a family.Considering the current diversity of family structures (Borges & Magalhães, 2009;Vieira & Rava, 2010;Zordan, Falcke, & Wagner, 2009), and some couples' option of not having children (Costa, 2007), the fi ve factor structure seems to provide more specifi c information about respondents, and retains the same number of factors as the original FESA.
Seven items loaded signifi cantly (i.e., factor loadings > .30) on the Factor 1 (Work and education; eigenvalue = 7.35) and accounted for 30% of the variance; fi ve items loaded signifi cantly Factor 2 (Church and community; eigenvalue = 2.95) and accounted for 12% of the variance; four items loaded signifi cantly on the Factor 3 (Marriage; eigenvalue = 1.52) and accounted for 6% of the variance; three items loaded signifi cantly on the Factor 4 (Health; eigenvalue = 1.30) and accounted for 5% of the variance; fi ve items loaded signifi cantly on the Factor 5 (Children and Family; eigenvalue = 1.08) and accounted for 4% of the variance.Considering the parallel analysis suggestion, it is possible that in other samples, a factorial structure of four factors will result; in such case we hypothesize that factors three and fi ve will merge into a single factor.In Brazil, the FESA presented a factor structure similar to the original instrument.However, there were some differences in the factors 'Children's Future', 'Marriage & Family', and 'Church and community'.The items 'I will dedicate time to spend with my family' (item 19) and 'My children will live in peace' (item 23), originally in the factor 'Marriage and family', were associated with the factor 'Children's Future'.The factor 'Church and community' also differed in that the items 'I will be a leader in my community' (items 9) and 'I will do volunteer work in my community' (item 15) had factor loadings lower than .30and were removed from the subscale.
The new structure of FESA in the Brazilian sample presented the following characteristics: 22 items divided into the following subscales (a) Work and Education (items 1, 7, 12, 13, 16, 17, and 22) 2).Table 2 Exploratory Factor Analysis of the 24-Item FESA

Items
Original Items When I'm an adult...

Brazilian Portuguese Items
No futuro eu...

Items
Original Items When I'm an adult...

Brazilian Portuguese Items
No futuro eu...

Factor loadings WE Ch Ma
He CF

Discussion
The present study aimed to contribute to research about the power of future expectations in individuals' current lives, by investigating the factor structure of a measure of young people's future expectations in a number of domains.The original measure, the FESA, was developed in Chile and administered to young people in high school.The current study presents data from a sample of 547 young individuals from differing SESs status in Brazil.Results reinforce the importance of considering the infl uence of contextual specifi cities in the measures' factorial structure.
The FESA factorial structure in a Brazilian sample was analyzed, and a factor structure with differences and similarities in relation to the original instrument was observed.For instance, minor differences emerged in the Factors Children's Future, Marriage and Family, and Church and Community.Two items about family and children loaded onto the factor Children's Future instead of the factor Marriage and Family.Both factors had stronger internal consistency reliabilities in the Brazilian sample.Specifi cally, from α=.73 in the Chilean sample (McWhirter & McWhirter, 2009), the subscale of Children's Future was α=.86 in the Brazilian sample.Similarly, the internal reliability of the Factor Marriage & Family was stronger in the Brazilian sample without these items (α=.80 vs. .82).
The items 'I will dedicate time to spend with my family' (item 19) and 'My children will live in peace' (item 23) were more strongly associated with the Factor Children's Future, considering that family includes not just the couple, but also other members (e.g., kids).The Factor Marriage and Family referred to 'couple's life'.It was composed of the items 'I will get married' (item 3), 'I will get married before I am 30 years old' (item 8), and 'My marriage will last forever' (item 14).Therefore, the Factor three retained only items related to the expectation of having a partner and getting married, not necessarily those focused on having children and building a family.
Based on these differences, the term 'Family' was added to the Factor Children's Future, which was named 'Children and Family'.The Factor Marriage and Family was re-named 'Marriage'.The difference in these factors may suggest that in this Brazilian sample, marriage is not necessarily associated with having children.People get married for diverse reasons, such as sexual satisfaction, and feelings that they are admired and valued.Couples look for company and security; they avoid loneliness and try to solve previous family problems.Social prestige, material values, legal, and fi nancial aspects are also motivations for marriage.Individuals do not necessarily or invariably get married for the obligation of having children.They look for a relation of affection (Branden, 2000;Carter & McGoldrick, 2001;Costa, 2007).
In addition to this, currently, marriage refl ects the extension of adulthood and even of adolescence, and planning children is not the main goal of a marriage.Parenthood became a project possible to happen after years of union.Establishing relations with different partners has become frequent, and marriage became a less critical transition in family cycle than before (Branden, 2000;Carter & Mc-Goldrick, 2001).Thus, differences in the factor structure may be due to differences in the role of marriage between Brazil and Chile.It is also possible that the difference in the age of the samples contributed to the difference in the factor structure.It may be that for adolescents in high school, marriage and having children are viewed as more similar life roles, while for emerging adults who may be closer to making marriage decisions, having children is more clearly a distinct role.
The Factor 'Church and community' also presented a new structure.The items 'I will be a leader in my community' (items 9) and 'I will do volunteer work in my community' (item 15) presented factor loadings lower than .30and were removed from the subscale.This new structure of the instrument may be related to the plural meanings of the term 'Community' (Schmidt, 2010;Silva & Simon, 2005).In Brazil, the conception of community is frequently associated with poorer environments.Intervention placed in low SES contexts use the term 'community' to name these fragile environments (Elvas & Moniz, 2010).Poorer settings lack of governmental investment and members of the neighborhood needs to fi nd different strategies to deal with their diffi culties, in a communitarian direction (Schmidt, 2010).The context of insecurity faced by these families leads them to fi nd strategies to support each other, count on their broad family and neighbors as reference, and build a solidarity network (Bem & Wagner, 2006;Llanos, Orozco, & Garcia, 1999).Considering the sample was composed of LSES and HSES participants, it might be the case these two items did not correspond to all participants' reality.
The items related to religiosity, namely 'I will regularly go to mass or other religious services' (item 4), 'I will participate in many church activities' (item 20), and 'I will instill faith in my children or nieces and nephews' (item 24) were maintained in the Factor 'Church and Community'.It is probably related to the fact that Brazil is a predominantly Catholic country (Cerqueira-Santos & Koller, 2009) and religiosity is present in low and high SES contexts, constituting part of people's activities.
Continued research on the utility of the FESA is warranted.Future research recommendations include investigating the properties of the FESA with Brazilian adolescents.The FESA may be useful in research with young adults in Brazil on how their future expectations infl uence important behaviors such as decisions about postsecondary schooling or training, civic engagement, timing of marriage or committed partnerships and childbearing, and health behaviors.Such research might be carried out from the framework of Future Time Perspective Theory (Lens, 1993;Nuttin, 1983;Schmitt, 2010).

Conclusion
This study analyzed the factor structure of FESA in the Brazilian context and aimed to provide a research tool for subsequent studies about the infl uence of young individuals' future expectations in their current life.Young individuals' future prospects are signifi cant because they are related to long term plans about the future including (e.g. higher education and self-perception of competence).
The investigation is an important contribution for studies about the influence of future perspectives in young individuals' present-day.The psychometric analysis of FESA properties in the Brazilian context may contribute to the improvement of the original measure and for future comparisons of the instrument's factorial properties in different contexts.Coming studies may consider including other regions of Brazil in order to examine the inventory factorial structure within the country.Future investigations may also compare the mean score of FESA's dimensions among different groups of comparisons (sex, SES, age etc.).

Figure 1 .
Figure 1.Scree-plot of the empirical eigenvalues and random eigenvalues of the FESA.

Table 1
Subscales Reliability of the FESA Revised Measure in Chilean Adolescent Sample Note.Total scale alpha with an adolescent Chilean sample (α= .88).