Profiles of hope: How clusters of hope relate to school variables

Highlights

• Four theoretically consistent groups of hope were derived via cluster analysis.

• High hopers had the most adaptive school profile followed by high agency thinkers, high pathways thinkers, and low hopers.

• Different levels of hope may lead to different perceptions of school, which may result in differing levels of success.

Abstract

In a diverse sample of 297 adolescents, four clusters based on the agency and pathways subscales of the Children's Hope Scale were derived via cluster analysis: high hopers (n = 105), high agency thinkers (n = 73), high pathway thinkers (n = 57), and low hopers (n = 62). We examined differences among clusters on psychological (consideration of future consequences, perceived life chances, perceived stress, and self-esteem) and educational constructs (academic self-concept, academic investment, and self-reported academic achievement). We also examined differences among hope clusters by sex, grade, and socioeconomic status. Results indicated that (a) the hope clusters derived were theoretically consistent with hope theory, (b) there were differences in the demographic makeup of the hope clusters with effect sizes ranging from small to medium, (c) students with different hope profiles differed on the outcome variables with about 50% of the effect sizes ranging from medium to large, and (d) high hopers and high agency thinkers had the most adaptive outcomes. The findings suggest that hope may be a useful variable for determining academic and psychological risk as well as a potential avenue for intervention in adolescence.

Introduction

Academic success during the adolescent years has been linked to several positive outcomes later in life. Adolescents with higher grade point averages (GPA) earn more money as adults (French et al., 2015, Oehrlein, 2009), are more likely to be accepted into highly ranked colleges (Espenshade, Hale, & Chung, 2005), are more likely to be successful in college (Noble & Sawyer, 2004), and are more likely to be hired after graduating from college (Barr & Mcneilly, 2002) than those with lower GPAs. Further, students who graduate from college generally have higher status jobs, are happier overall, and live longer lives than those that do not graduate (Egerter et al., 2009, Pascarella and Terenzini, 2005). As can be expected, predicting academic success during adolescence is an area of research that receives substantial attention (see Hattie, 2009 for a review). Perception-based constructs have received substantial attention recently in the academic achievement literature (Paunesku et al., 2015, Yeager and Walton, 2011). This increased focus has come about for at least two reasons: (a) several perception-constructs have been implicated in academic functioning, including closing the achievement gap, and (b) interventions that target perception-based constructs can be quick, effective, and long lasting (Walton and Cohen, 2007, Walton and Cohen, 2011, Yeager and Walton, 2011).

One perception-based construct that appears to have the potential to substantially impact the academic success of adolescent students is trait hope (Snyder, 2002), “a relatively stable personality disposition” (Snyder, Lopez, Shorey, Rand, & Feldman, 2003, p. 123). Although hope can also be assessed as a state (Snyder et al., 1996), that is, “a temporary frame of mind” (Snyder et al., 2003, p. 123), in this paper, all discussions of hope are in terms of hope as trait. In addition to being strongly correlated with academic achievement (e.g., r = 0.69; Feldman & Kubota, 2015), intervention studies have indicated that hope can be changed in as little as 90 minutes (Feldman & Dreher, 2011), and the changes have been substantial (average d = 0.40; Weis & Speridakos, 2011) and have been maintained for as long as 18 months (Marques, Lopez, & Pais-Ribeiro, 2011).

In this paper, we examined the relationship between hope and several psychological and educational variables that are associated with adaptive functioning in adolescence. However, first, we review hope theory and the literature on hope in schools. Next, we discuss some influential school variables and how they relate to hope. Finally, we present a study examining how different profiles based on the two components of hope relate to these variables.

Hope, defined as one's perceived ability to execute envisioned paths to future goals, is a two-component cognitive-motivational construct (Snyder, 2002, Snyder et al., 1991). Hope encompasses how individuals choose goals, how they plan to accomplish chosen goals, their motivation for accomplishing chosen goals, and their belief in their capacity to accomplish chosen goals. Hope is primarily measured in child and adolescent populations using the Children Hope Scale (Snyder et al., 1997) and in adult populations using the Adult Hope Scale (Snyder et al., 1991).

One component of hope is pathways. Pathways is one's perceived ability to envision routes to one's goals (Snyder et al., 1991). Put another way, pathways is one's perceived capacity to envision and produce a roadmap to a better future, irrespective of one's current circumstances. For example, if a student in high school wants to obtain a job as a college professor, his ability to envision himself as a professor in vivid detail will make up part of his pathways thinking, whereas his ability to envision steps to accomplish that goal (e.g., going to college, taking the Graduate Record Examination, excelling in graduate school) will make up the other part. The theoretical importance of pathways is embodied in a quote by William Ward (n.d.): “If you can imagine it, you can achieve it”. Snyder (2002) argued that students who are high in pathways produce (a) more elaborate, creative, and specific plans of action to accomplish their goals, (b) more realistic goals, and (c) alternative paths to accomplish goals in the event their initial route proves to be untenable. In contrast, students who are low in pathways typically produce (a) unclear and vague courses of action to accomplish their goals and (b) idealistic and inappropriate goals for their current level of achievement; they also do not produce alternative paths to goals (Snyder, 2002).

The other component of hope is agency. Agency is one's belief, along with the corresponding motivation and confidence, that one can accomplish one's envisioned goals (Snyder, 2002, Snyder et al., 1991). Whereas pathways refers to individuals' perceived ability to see the goals they want to accomplish and the roadmap to achieve them, agency refers to their ability to believe in themselves to accomplish those goals, as well as the motivation to do the work that will propel them along the goal-achievement pathway. Agency also encompasses the determination to persist throughout the goal-achievement process when setbacks occur. Continuing the example above, the student's agency would be his belief, motivation, and confidence in himself that he could excel in college, on the GRE, and in graduate school, and finally attain his goal of being a professor. Students who are high in agency are typically more persistent (Snyder, 1994), more motivated to accomplish their goals (Snyder, 2002), and more likely to engage in strategies that help them to persist during stressful situations (e.g., positive self-talk, Snyder, Lapointe, Crowson, & Early, 1998).

As agency and pathways are both components of hope, they are intercorrelated (Adelabu, 2008, Arif and Yousuf, 2010). However, there is substantial structural validity evidence in support of hope's 2-factor structure. In the study introducing the Adult Hope Scale, Snyder et al. (1991) reported that the 2-factor structure was supported in eight samples – six college student samples and two outpatient samples – and reported correlations among the subscales ranging from 0.38 to 0.57. This finding was replicated in the study introducing the Children's Hope Scale (CHS; Snyder et al., 1997). Snyder et al. (1991), which concluded that agency and pathways were “related, but not synonymous” and speculated that “future research may unravel differential correlates of agency and pathways and may yield information pertaining to their separate construct validity and utility” (p. 582). Although Snyder (2002) continued to theorize about the two factors, a total hope score is typically used in research. In this study, we use the two factors of hope as initially theorized.

Hope has been distinguished from related constructs like self-efficacy and optimism both theoretically and empirically. Researchers have argued that hope is experienced under different conditions than self-efficacy and optimism, and is elicited in different circumstances (Bruininks and Malle, 2005, Snyder, 2002). For example, optimism is likely to be experienced when students believe that they will accomplish a desirable future goal, like getting an A in a math course, but do not know how the good grade will come about. This feeling of optimism changes to hope when those students envision a pathway to getting the A in the math class; that is, when they know how they will accomplish earning the A and in turn feel a sense of agency in the process of the A coming about (Snyder, 2002). Self-efficacy is different in that it is likely to be experienced before either hope or optimism. Self-efficacy is likely to be experienced in the can phase whereas hope and optimism are likely to be experienced in the will phase (Snyder, 2002). Continuing the example from above, students are likely to experience a sense of self-efficacy when they are deciding whether or not they can get an A in the math class.

Hope has also been shown to be empirically different than self-efficacy and optimism in several studies. Using confirmatory factor analysis, Bryant and Cvengros (2004) found that a joint examination of hope and optimism items resulted in the best fit when the items from the two constructs loaded on separate factors. Feldman and Kubota (2015) found that general hope shared 44% of the variance with general self-efficacy (r = 0.67) and that academic hope shared a similar amount of variance with academic self-efficacy (r = 0.66), indicating that about 56% of the variance in the constructs is unique. Mirroring these findings, Ben-Naim, Laslo-Roth, Einav, Biran, and Margalit (2017) reported that both the pathways and agency subscales of hope shared about 42% of variance with academic self-efficacy (r = 0.65 & 0.64 respectively), whereas Dixson, Worrell, Olszewski-Kubilius, and Subotnik (2016) found that hope shared about 20% of variance with academic self-efficacy (r = 0.45). All three of these studies indicate that more than half of hope's contribution is unique. Finally, correlations between hope and optimism range from 0.23 to 0.56 (Feldman and Kubota, 2015, Magaletta and Oliver, 1999, Vacek et al., 2010).

Hope is also meaningfully related to several educational and psychological constructs in the literature. Hope has been found to correlate with academic achievement at all levels of education, even after controlling for ability (Curry et al., 1997, Snyder et al., 2002) and academic engagement (Marques, Lopez, Fontaine, Coimbra, & Mitchell, 2015). Researchers have found that hope is positively associated with several additional outcomes, such as success in competition (Curry & Snyder, 2000), general wellbeing (Parker et al., 2015, Satici, 2016), problem solving ability (Snyder et al., 1991), resilience (Satici, 2016), and social competence (Sympson, 1999). Longitudinal studies have shown hope to be related to life satisfaction after a year (Marques, Lopez, & Mitchell, 2013) and to more favorable developmental trajectories over a three-year span (Schmid et al., 2011). Additionally, researchers have found hope is inversely related to several negative outcomes, such as anxiety (Arnau, Rosen, Finch, Rhudy, & Fortunato, 2007), depression (Snyder, 2004), and PTSD (Hassija, Luterek, Naragon-Gainey, Moore, & Simpson, 2012).

Several studies have examined how hope relates to gender, race, age, and SES. Results from the majority of studies examining gender have indicated that hope scores do not differ significantly based on gender (Adelabu, 2008, Snyder et al., 1997, Snyder et al., 2002, Snyder et al., 2003). However, Valle, Huebner, and Suldo (2004) found that women reported significantly higher hope scores than men, but with a small effect size (d = 0.16). Results from studies examining whether hope scores differ significantly across race are mixed. Although some studies indicate that hope scores do not differ significantly based on race (Adelabu, 2008, Snyder et al., 1991, Snyder et al., 1997), some researchers have found that African Americans report higher hope scores than European Americans, although with small effect sizes (e.g., McDermott et al., 1997, d = 0.21; Valle et al., 2004, d = 0.18). Finally, studies indicate that hope scores do not differ across SES or age (Snyder et al., 1991, Snyder et al., 1997, Valle et al., 2004).

As schooling typically involves long term goals for most students (e.g., the desire to get into honors classes, be valedictorian, to graduate, or get into college), hope, as theorized, has the potential to be a useful construct in schools. For example, students commonly have the goal of earning good grades across multiple subjects (Svanum & Bigatti, 2006). Snyder (2002) argued that successfully attaining the aforementioned academic goals is heavily influenced by a student's level of hope. However, very little research has been conducted on hope in school settings, and the studies that have been conducted have focused almost exclusively on the relationship between hope and academic achievement.

Snyder et al. (1997) found that hope scores correlated 0.50 with scores on the Iowa Test of Basic Skills in a sample of 372 students in Grades 4–6, an association similar in size to the one between IQ and school performance (Neisser et al., 1996). Subsequently, Snyder et al. (2002) conducted a 6-year longitudinal study tracking the hope levels, academic achievement, and the graduation status of 213 college freshmen (52% women, M_age = 18.17). They found, after controlling for ACT scores obtained before the study, that hope predicted higher cumulative grades (r = 0.21, p < 0.01) and that those with high hope were more likely to have graduated and less likely to have been dismissed from school over the 6-year period (Cramer's V = 0.19).

Adelabu (2008) examined how future time perspective, ethnic identity, and hope relate to academic achievement in a sample of 661 financially disadvantaged youth. She reported that (a) youths who were future oriented earned higher grades (r = 0.12, p < 01), (b) youths with higher agency earned better grades (r = 0.20, p < 01), (c) youths with higher pathways did not earn better grades (r = 0.08, p > 0.05), (d) youths with higher ethnic identity earned better grades (ethnic exploration r = 0.13, p < 0.01; ethnic affirmation r = 0.24, p ≤ 0.01), and (e) only agency predicted academic achievement among both rural and urban adolescents when the other variables were controlled for.

Recently, Feldman and Kubota (2015) conducted a study to determine which variable among academic self-efficacy, optimism, and academic hope was the strongest predictor of GPA after controlling for the other variables. Using path analysis with a sample of 89 college students, the researchers reported two major findings: (a) academic hope was the strongest predictor of GPA (β = 0.54), with a coefficient more than twice the coefficient for academic self-efficacy (β = 0.23), and (b) optimism did not predict GPA (β = 0.01). Similarly, Gallagher, Marques, and Lopez (2017) conducted a longitudinal study examining how hope and self-efficacy relate to academic achievement and academic persistence over the course of college in sample of 229 college students. They found that after controlling for high school GPA and ACT scores, academic hope was a significant predictor of semesters enrolled in college (b = 0.29, p < 0.05), graduating within four years (b = 0.14, p < 0.05), and academic performance (b = 0.21, p < 0.01, explaining 6.1% of GPA's variance), whereas academic self-efficacy did not contribute (p's > 0.05). Further, they found that academic hope was meaningfully associated with cumulative GPA across all 4 years of college (rs range from 0.30 to 0.33).

Outside of correlational studies which do not differentiate students at different levels of hope, most studies that examine hope and academic achievement use arbitrary cutoffs of overall hope scores (e.g., those that have above average hope are classified as high hopers and those that have below average hope are placed in a low hope group), and it is unclear whether the conclusions drawn from these studies are generalizable. Students who are high in both pathways and agency are theoretically different from those students who are high in either agency or pathways alone (Snyder, 2002). However, these latter two sets of students often end up in the same group when groups are formed based on cut scores (e.g., Snyder et al., 2003). Therefore, more research needs to be conducted that creates hope groups based on more robust analyses.

One such study is that of Gilman, Dooley, and Florell (2006). These researchers used cluster analysis to create different groups of hope in a sample of 341 adolescents. They found three clusters of hope based on students' pathways and agency scores: high hope, average hope, and low hope. They found that those with high hope earned significantly higher grades than those with average (d = 2.34) and low hope (d = 3.92). They also found that students with average hope earned significantly higher grades than those with low hope (d = 1.94). However, hope theory postulates that there are at least four hope groups (Snyder, 1994, Snyder, 2002, Snyder et al., 1991, Snyder, 1994): those that are high in both agency and pathways (high hopers), those that are high in one and average/low in the other (high agency thinkers and high pathways thinkers), and those that have both low agency and low pathways (low hopers). No study on hope and achievement has yet examined these four groups.

There are several variables that are theoretically linked to academic achievement that could help clarify the role of hope in the school context. In this study, we examined three groups of variables related to academic functioning and schooling. The first are academic variables, including academic achievement, academic investment (i.e., how important excelling in school is to a student; Sinclair, Huntsinger, Skorinko, & Hardin, 2005), academic self-concept (Bong & Skaalvik, 2003), school belonging (i.e., one's perceived connectedness to the school community; Booker, 2006), and educational expectations (i.e., how much schooling one expects to complete). The second group consisted of two psychological variables that have been shown to be related to academic functioning, that is, self-esteem (Rosenberg, 1965) and perceived stress (Cohen, 1988). The third and final group consisted of two future-oriented variables that have been linked to academic performance: perceived life chances (i.e., one's expectations that positive events will occur in one's future; Jessor, Donovan, & Costa, 1990) and consideration of future consequences (i.e., one's consideration of distant vs. immediate consequences of potential behavior; Strathman, Gleicher, Boninger, & Scott, 1994).

The choice of outcome variables was not arbitrary. All of these variables were included in the dataset used for this study for several reasons. They have been used in previous studies and have demonstrated associations with doing well in school or with adolescent wellbeing (Booker, 2006, Chavous et al., 2008, Rubin et al., 1977, Stevenson et al., 1998, Worrell and Hale, 2001, Worrell et al., 1999). Additionally, they span three different domains – academic, wellbeing, and future-orientation – that allowed for a broad-based analysis of the different psychosocial constructs studied in adolescent students. These variables can help us understand if students with different hope profiles perceive the school environment differently. Finally, these constructs all have theoretical associations to hope (Snyder, 2002). For example, one's sense of agency and goals in the school context are directly related to how much one values school, is thriving in the school context, and how much one feels that one is a part of the school community.

The present study answers the call for more research on hope in the school environment. There were several goals for the current study. The first goal was to examine whether cluster analysis based on pathways and agency scores would yield interpretable clusters. We hypothesized that we would find the four theorized hope groups. The second goal was to examine if hope clusters are associated gender, SES, and grade. Given the majority of previous literature indicates that hope does not differ across gender and age (e.g., Snyder et al., 2002, Snyder et al., 2003, Valle et al., 2004), we made no hypotheses about hope cluster membership in relation to gender or grade. With regard the SES, the literature is mixed. Empirical research has indicated that hope does not differ across SES levels (Snyder et al., 1991, Snyder et al., 1997, Valle et al., 2004), but hope theory suggests that those with more barriers are likely to have lower agency and pathways scores (Snyder, 2002) and ultimately lower levels of hope (Dixson, Keltner, Worrell, & Mello, 2017). Therefore, it was hypothesized that hope cluster membership would differ significantly across SES with higher SES students having a higher proportion of students in the high hoper cluster and lower SES students having a higher proportion of students in the low hoper cluster.

The third goal was to examine hope cluster differences across perceived stress, educational expectations, self-esteem, academic investment, consideration of future consequences, academic self-concept, perceived life chances, school belonging, and academic achievement. Based on how high hope is more adaptive than low hope (Schmid et al., 2011, Snyder, 2002), it was hypothesized that there would be significant differences across all school variables. More specifically, it was hypothesized that individuals in the high hoper cluster would report more adaptive scores across all variables than the three other clusters, and that individuals in both the high pathways and high agency clusters would report more adaptive scores across all variables than those in the low hoper cluster.