Abstract
Although emotional engagement has received much attention in educational research, relatively little is known about its predictors among undergraduates. Accordingly, to predict emotional engagement, we used the learning and study strategies model, proposed and revised by Weinstein and Palmer, which has 10 strategies namely; information processing, selecting main ideas, test strategies, anxiety, attitude, motivation, concentration, self-testing, study aids, and time management. Thus, an adapted Arabic version of the learning and study strategies inventory-second edition and an emotional engagement scale were administered to 522 undergraduates. To ascertain the accuracy and generalizability of the final selected model in future samples, six candidate models were cross-validated using the leave-one-out approach. Results revealed that the final five-predictor model, which included information processing, anxiety, attitude, concentration, study aids, was the most parsimonious and had the lowest mean square error, explaining approximately 40% of the variation in undergraduates’ scores on emotional engagement. The study aids strategy was found to be the strongest predictor of emotional engagement. Educational implications, limitations, and directions for future research were also discussed.
Keywords
Student engagement has recently gained greater interest due to its association with students’ learning outcomes and academic performance in different learning contexts (e.g., in person, online) in higher education (see, e.g., Bond & Bedenlier, 2019; Lee, 2012; Martin & Bolliger, 2018; Stan et al., 2022). Specifically, researchers have found a positive association between student engagement and academic achievement in different majors including (a) accounting (e.g., Shernof et al., 2017), (b) English language (e.g., Karabıyık, 2019), and (c) stem education (e.g., Franco & Patel, 2017), to name a few.
According to the “Glossary of Education Reform” (2016), “student engagement refers to the degree of attention, curiosity, interest, optimism, and passion that students show when they are learning or being taught, which extends to the level of motivation they have to learn and progress in their education.” As a multidimensional construct, student engagement includes behavioral, emotional, and cognitive components (Fredricks et al., 2004; Kuchinski-Donnelly & Krouse, 2020). It is worth noting that few other researchers (e.g., Deng et al., 2020) adopt a fourth dimension of student engagement namely, social engagement, which mainly includes student-teacher and student-student relationships. However, in this study, we focus on the component of student engagement known as emotional engagement, which is sometimes labeled affective or psychological engagement (Appleton et al., 2006; Bowden et al., 2021; Jimerson et al., 2003), because it received less attention in the extant literature compared to other dimensions of student engagement. In addition, various scholars have noted the role of positive emotions in learning and academic outcomes (e.g., Rodrigo-Ruiz, 2016), and emotional engagement has been internationally studied by applied researchers and practitioners (Hewson, 2018; Özhan & Kocadere, 2020; Sagayadevan & Jeyaraj, 2012; Sakr et al., 2016; Taylor & Statler, 2014; Ulmanena et al., 2016; Wara et al., 2018). Despite this international interest in emotional engagement, little empirical research has been conducted on its predictors among undergraduates, especially using learning and study strategies (hereafter known as LASSs) model proposed and revised by Weinstein and Palmer (2002). Thus, examining the predictive power of the LASSs seems relevant to explain Egyptian undergraduates’ emotional engagement, which is the main objective of the current investigation (more information is provided below). In the sections that follow, we start by illustrating the LASSs model, followed by emotional engagement and its predictors, and then we discuss the study rationale with special emphasis on limitations of prior research, which underlie conducting the present study.
Learning and Study Strategies
Prior research has documented that LASSs utilized in different learning contexts have been important for achieving the intended learning outcomes in different study programs such as medical education (e.g., Nabizadeh et al., 2019), since higher performing students tend to choose more effective learning strategies (e.g., Geller et al., 2018). One of the most widely used models of LASSs was proposed and revised by Weinstein and Palmer (2002). Based on the revised model, LASSs refer to the behaviors and beliefs employed by students during learning and studying. This model includes 10 strategies: information processing, selecting main ideas, test strategies, anxiety, attitude, motivation, concentration, self-testing, study aids, and time management.
In the LASSI-II manual, Weinstein and Palmer (2002) operationally defined the 10 strategies. Information processing refers to students’ ability and skills to use organization strategies to support their learning and studying. Selecting main ideas refers to students’ skills at identifying main points from supporting details. Test strategies illustrate students’ effective use of both test preparation and test-taking strategies. Anxiety indicates the extent to which students worry about learning and their academic performance. Attitude refers to students’ tendency and interests toward learning and achieving success in college. Motivation is reflected on students’ persistence, self-discipline, and willingness to exert the effort necessary to complete academic tasks successfully. Concentration is expressed in students’ ability to pay attention and control distractors while learning and completing academic tasks. Self-testing refers to students’ use of reviewing and comprehension monitoring strategies to check their level of understanding of what they are learning and studying. Study aids refers to students’ use of the supportive techniques and available resources to help them learn and remember new information. Last, time management refers to students’ ability to manage their time wisely during learning and studying.
Learning strategies have been found to be significant predictors of students’ learning outcomes in different contexts (e.g., Dill et al., 2014; Seabi, 2011; West & Sadoski, 2011). It has been also argued if LASSs are predictors of academic performance, they are likely to be predictors of students’ emotional engagement, since the latter is an essential factor for academic performance as repeatedly discussed in literature (see, Wang & Sui, 2020). In more detail, types of activities utilized by instructors and students during learning and studying (e.g., self-monitoring) increase their engagement and learning outcomes (Bender, 2017). Put another way, when students adopt specific strategies (e.g., managing time wisely) to learn and study the course content, they are likely to become more productive and feel enjoyment of learning, which probably make them more emotionally engaged. Relatedly, when they positively collaborate to learn or complete a specific task, they are also likely to emotionally engage and achieve the intended outcomes. Thus, such learning behaviors and strategies have been thought of to be effective factors for students’ emotional engagement. With that said, students’ emotional engagement could be predicted by learning behaviors and strategies utilized by students in their learning and studying. As a result, LASSs are proposed to be explanatory variables (i.e., predictors), whereas emotional engagement is investigated as the outcome variable in the present study.
Emotional Engagement
Given the role of emotions in learning, emotional engagement has recently received special attention in both in-person and online learning environments (Dao & Sato, 2021; Esmail & Matthews-Roper, 2022). Various researchers have proposed definitions for emotional engagement characterized mainly by learners’ interactions with peers, staff, and faculty in ways that enhance their learning outcomes. For instance, Jimerson et al. (2003) define emotional engagement as students’ connections with institutions, instructors, and peers. Sheard et al. (2010) define it as students’ emotional feeling, perceptions about the learning environment, and their relationships with peers and faculty. In a higher education context, Lawson and Lawson (2013) describe it as the psychological and emotional association with college. Relatedly, emotional engagement is reflected in enjoyment, enthusiasm, desire, feeling of safety, and communication with peers and faculty (Uden et al., 2013). Other researchers define it in terms of belonging to the classroom and valuing the context of learning (Al-Amri, 2020). Synthesizing these definitions as well as others reviewed during the literature review phase of the research, emotional engagement is operationally defined in the context of the present study as undergraduate students’ positive interaction and collaboration with peers and faculty members to create a supportive learning environment, which enhances their learning and studying.
Predictors of Emotional Engagement
Given the utility of students’ emotional engagement in achieving the intended learning outcomes, researchers have studied its predictors from different perspectives. For instance, Ryan and Patrick (2001) stated “students’ perceptions of teacher support, and the instructor’s role in promoting interaction and mutual respect were related to positive changes in their motivation and engagement” (p. 437). Similarly, Klem and Connel (2004) found that teachers’ support was associated with student engagement. Instructors’ support can be viewed in terms of the opportunities created for students to effectively use LASSs in class activities. Zhao and Kuh (2004) indicated that participating in a learning community was positively related to student engagement, self-reported outcomes, and overall satisfaction in college. Relatedly, Heyward (2010) concluded that role-play enhanced emotional engagement, since it set the stage for mutual interaction among undergraduates and consequently increased their positive attitudes toward learning and studying. It has also been argued that active learning, peer relationships, and positive social skills are necessary for engaging students in learning and studying (e.g., Wentzel, 2012; Zepke & Leach, 2010). Sagayadevan and Jeyaraj (2012) found that students who had more effective interaction with instructors had higher levels of engagement.
Other researchers investigated the predictors of emotional engagement within different populations. Ainley and Ainley (2011) selected data for four countries (Columbia, Estonia, Sweden, US) from the Program for International Student Assessment (PISA) and argued that students’ positive emotions in the learning environment increased their enjoyment and attitudes, which in turn increased their engagement with learning science. Elffers et al. (2012) pointed out that positive experiences and attitudes increased emotional engagement among 909 students in postsecondary vocational education in the Netherlands. Mazer (2013) demonstrated the association between students’ emotional interest and their engagement based on the results of a study conducted among 183 undergraduates enrolled at a large university in the US-Midwest.
In a study among 19 distance-learning students at a New Zealand University, Kahu (2014) conducted a qualitative study to explore predictors of emotional engagement. Based on the analysis of interviews and other qualitative materials, participants felt less connected to the university and concluded that faculty and staff should consider students’ emotional engagement when designing and delivering courses by creating interactive learning environments. The design of learning activities that supports students’ use of LASSs has been a primary factor for their emotional engagement. In a study among 786 sixth graders at 10 public schools in the US-Midwest, Ben-Eliyahua et al. (2018) pointed out that mastery goals were positively correlated with emotional engagement. It has been argued that students who adopt mastery goals utilize LASSs effectively. Given the widespread implementation of online learning in graduate programs, Kuchinski-Donnelly and Krouse (2020) examined if autonomy, competence, and relatedness predict emotional engagement among 123 graduate nursing students. They concluded that competence was the only statistically significant predictor of emotional engagement. Palmgren et al. (2021) concluded that students who were less anxious (i.e., had low anxiety profiles) were more emotionally engaged. In an online course during COVID-19, El-Sayad et al. (2021) found that academic self-efficacy and perceived usefulness had significant direct effect on emotional engagement in a sample of 330 students.
It can be concluded from results of the above studies that students’ interactions in the learning environment and their attitudes, positive emotions, and mastery goals are related to their emotional engagement. Thus, such results raise a question about LASSs’ utility to predict undergraduates’ emotional engagement. Specifically, the impact of LASSs, utilized by undergraduates during learning and studying, on their emotional engagement is still largely unknown. Conducting such a research project adds to the educational and psychological literature of learning strategies, student engagement, and learning outcomes.
Study Rationale
The studies reviewed above were conducted in different countries (e.g., Columbia, Estonia, Netherlands, New Zealand, Sweden, and USA). Additionally, most of these studies were conducted to predict emotional engagement using other educational and psychological constructs, rather than LASSs, such as instructors’ promotion of interactions (Ryan & Patrick, 2001), instructors’ support (Klem & Connel, 2004), students’ participation in the learning community (Zhao & Kuh, 2004), role-play (Heyward, 2010), student-instructor relationships (Sagayadevan & Jeyaraj, 2012), positive experience (Elffers et al., 2012), emotional interest (Mazer, 2013), mastery goals (Ben-Eliyahua et al., 2018), autonomy, competence, and relatedness (Kuchinski-Donnelly & Krouse, 2020), and academic self-efficacy and perceived usefulness (El-Sayad et al., 2021). Stated differently, very few studies to date, if any, have examined the predictive power of the LASSs model on emotional engagement among Egyptian undergraduates, which is the overall objective of the present study. More specifically, we attempt to fill the gap identified through the review of literature and advance research on predictors of emotional engagement by attempting to answer three research questions. For the first, “using all subset regression, what is the best set of candidate LASSs models, which parsimoniously account for the variation in scores of students’ emotional engagement?” Second, “which candidate model can be generalized in future sample using the leave-one-out cross-validation approach (LOOCV)?” Third, “which strategy of the LASSs model is the strongest predictor of emotional engagement?”
The present study has the potential to advance research on student engagement in general and emotional engagement in particular. In more detail, using LASSs to predict emotional engagement in this new population is expected to contribute to our understanding of the association between learning strategies and students’ emotional engagement, which can have deeper insights in designing learning situations that better support students’ emotional engagement.
Method
Research Design
The present study follows a cross-sectional research design in which all participants responded to the study instruments as described in more detail below.
Participants
A total of 522 undergraduates (369 females, 153 males, Mage = 20.96 years, SDage = 0.61, age range: 20–27 years) were recruited to participate in the study at a large public university in the southern region of Egypt, which used to include three campuses in three governorates in Upper Egypt (Abulela & Bart, 2021). Participants were specialized in Arabic, Biology, Chemistry, English, History, Mathematics, and Mechanical Engineering majors. They also had diverse socioeconomic status (e.g., low, medium, and high), given the diversity of population surveyed. Though the sample was convenient, we proportionally represented gender subgroups to increase similarity of students’ demographics between the sample and population of interest as recommended in methodological research (e.g., Abulela & Harwell, 2019), which increases the likelihood of results generalizability. All participants completed an adapted Arabic version of the LASSI-II (Abdelsamea & Bart, 2019) and the emotional engagement subscale adapted from the student engagement scale (Abulela, 2015).
Instruments
The Learning and Study Strategies Inventory-Second Edition (LASSI-II)
The adapted Arabic version of the LASSI-II was utilized to assess participants’ LASSs. The LASSI-II consists of 10 subscales with eight items per subscale, totaling 80 items. The 10 LASSs assessed by the LASSI-II include information processing (e.g., I translate what I am studying into my own words), selecting main ideas (e.g., It is hard for me to decide what is important to underline in a text), test strategies (e.g., I review my answers during essay tests to make sure I have made and supported my main points), anxiety (e.g., Courses in certain subjects, such as math, science, or a foreign language, make me anxious), attitude (e.g., I have a positive attitude about attending my classes), motivation (e.g., I am up-to-date in my class assignments), concentration (e.g., I find it hard to pay attention during lectures), self-testing (e.g., I test myself to see if I understand what I am studying), study aids (e.g., When they are available, I attend review sessions for my classes), and time management (e.g., I put off studying more than I should).
The LASSI-II items were rated with 5-point ordinal response categories (i.e., ranging from 1 = totally inapplicable to me to 5 = totally applicable to me). There is no total score of the LASSI-II, since it is used for diagnostic purposes. Psychometric properties were estimated, where McDonald’s omega coefficient for score reliability ranged from a low of .65 (study aids) to a high of .86 (self-testing), with most coefficients falling between .70 and .80. To establish validity evidence based on internal structure, confirmatory factor analysis was utilized and yielded adequate goodness-of-fit indices, in which all indices fell within the recommended guidelines (e.g., Brown, 2015). For more information about the psychometric properties of the LASSI-II in Egyptian undergraduates, interested readers are referred to Abdelsamea and Bart (2019).
Emotional Engagement Scale (EES)
The EES assesses the interactions between undergraduates with their peers on the one hand and with faculty on the other, feeling of safety and support from instructors, enthusiasm, and joy (e.g., my instructors supports me when needed). It has 20 items where each item was a statement that participants rated on 5-point ordinal response categories (i.e., ranging from 1 = totally inapplicable for me to 5 = totally applicable to me). Scale psychometric properties were estimated. Score reliability was relatively high (McDonald’s omega coefficient = .84). Validity evidence based on the internal structure of the measure was also established by means of confirmatory factor analysis that provided satisfactory goodness-of-fit indices for the unidimensional model, χ2 (170) = 297.83, χ2/df = 1.7, Goodness of Fit Index (GFI) = 0.90, Root Mean Square Error of Approximation (RMSEA) = 0.050. All standardized estimates (loadings) were significant and above .30.
Procedure
Educational authorities were contacted to obtain permission for data collection. After approval, teaching assistants distributed the two instruments to undergraduate students in their classes, who were informed that their participation was voluntary. After their consent to participate, they were also informed briefly and clearly the purpose of the study and how to complete the two instruments based on the provided instructions. It took them 15 to 20 minutes to respond on the LASSI-II and 5 minutes to respond on the EES. Finally, we collected the responses and scored them.
Data Analysis
Data analysis was conducted in various sequential steps using R, version 4.0.3 (R Development Core Team, 2020). First, we used multiple regression analysis with backward elimination to select the final model by retaining predictors with p-value ≤.10. Second, we also used backward elimination to select the final model by retaining the final model with the lowest Akaike information criterion (AIC; Field et al., 2012). Third, we used all subset regression to fit all possible models. Since the number of all possible models equals 2 P − 1, where P is the number of predictors (P = 10 in the present study), the total number of possible fitted models equals 1023. We selected a candidate set of all plausible models based on the corrected AIC (AICc) such that all models within 2 units of the lowest AICc value were retained for further analysis (Burnham et al., 2011).
We conducted a cross-validation for the selected set of candidate models to increase the generalizability of the final selected model in different samples (Field et al., 2012). Relative to other methods, the LOOCV was shown to best correct against overfit in the cross-validation process (Ng, 1997). The algorithm for performing LOOCV includes: (a) removing the ith observation to be used as validation data and then using the remaining observations from the dataset as training data; (b) fitting the candidate set of models based on AICc to the training data; (c) using the estimated coefficients from the fitted models to compute the mean square error (MSE) for each model using the ith observation as validation data; and (d) repeating steps a to c for each observation. This process was repeated separately for all candidate models. Then, we computed the average cross-validated MSE across observations for each candidate model, since it is an index for the model accuracy such that the model with the lowest cross-validated MSE has the most accuracy and generalizability in different samples.
We also compared the candidate models using the nested ANOVA F test. If not statistically significant, it indicates the model with more predictors does not explain a statistically significant proportion of variation in the data beyond the variation that is already accounted for by the reduced model and consequently the latter should be selected for parsimony. Finally, we fitted the final selected model on the full data to obtain the regression coefficients and the amount of variance explained. Relatedly, we checked model assumptions to ensure the statistical conclusion validity and therefore the generalizability of study-based inferences (Abulela & Harwell, 2020).
Results
Table 1 shows the means, standard deviations, and intercorrelations between LASSs and emotional engagement. The correlation matrix was screened for evidence of multicollinearity (i.e., high correlation among predictors). The highest correlation was .61 between selecting main ideas and testing strategies, and consequently no evidence for multicollinearity was observed among the predictors (Field et al., 2012). We also noted that the highest correlation for the outcome variable “emotional engagement” was with study aids (r = .54).
Means, Standard Deviations, and Intercorrelations Between LASS and Emotional Engagement (N = 522).
Note. LASS = learning and study strategies; INP = information processing; SMI = selecting main idea; TST = test strategies; ANX = anxiety; ATT = attitude; MOT = motivation; CON = concentration; SFT = self-testing; STA = study aids; TMT = time management.
p < .05. **p < .01.
The following results are reported in accordance with the steps outlined in the Data Analysis section. Based on the results of backward elimination using both the p-value and the AIC criterion, we retained a model with five predictors; information processing, anxiety, attitude, concentration, and study aids. As shown in Figure 1, the model with five predictors has the lowest AIC value.
The backward elimination iteration process for capturing the final selected model based on the values of akaike information criterion (AIC).
Results of the 1,023 models fitted according to the all subset regression method yielded six models, where the value of AICc was within 2 of the lowest reported values as shown in Figure 2. We also noted that the predictors included in the five-predictor model retained based on the backward elimination using both the p-value and AIC criterions were retained in all six candidate models. This signifies the relative importance of the five predictors in predicting undergraduates’ emotional engagement. Furthermore, the five-predictor model had the lowest AICc value (1226.11).
The backward elimination iteration process for capturing the final selected model based on the values of the corrected akaike information criterion (AICc).
The six candidate models retained in Figure 2 were cross-validated to check their accuracy and generalizability in different samples. The values of the average cross-validated MSE were checked as an indicator for the model error accuracy in future observations. These values ranged from 0.5954 (Model 2: INP, SMI, ANX, ATT, CON, and STA) to 0.5963 (Model 1: INP, ANX, ATT, CON, and STA). As noted, there was not a substantial difference in the average cross-validated MSE between the model with five and six predictors. This called for conducting a statistical test to determine whether adding a sixth LASSs predictor explains a statistically significant proportion of variation in undergraduates’ emotional engagement scores.
The five-predictor model was compared against each of the six-predictor models in the candidate set, shown in Figure 2, using the nested ANOVA F-test. Results were in favor of the five-predictor model, since all p-values were >.05, indicating that adding a sixth predictor did not explain any significant proportion of variation in undergraduates’ emotional engagement scores (see Table 2).
Comparing the Six Candidate Models Using the ANOVA Nested F-test.
Note. INP = information processing; SMI = selecting main idea; TST = test strategies; ANX = anxiety; ATT = attitude; MOT = motivation; CON = concentration; SFT = self-testing; STA = study aids; TMT = time management; Δdf = Change in degrees of freedom.
Based on the above results, the five-predictor model was the most parsimonious and therefore was selected as the final model. After fitting the final selected model to the full data, results revealed that it was statistically significant, R2 = .401, F(5, 517) = 69.67, p < .001, explaining 40% of the total variation in undergraduates’ emotional engagement scores.
As shown in Table 3, the five LASSs predictors included in the final model were: information processing (β = .22, t(517) = 5.62, p < .01), anxiety (β = −.09, t(517) = −2.49, p = .01), attitude (β = .19, t(517) = 4.91, p < .01), concentration (β = .10, t(517) = 2.51, p = .01), and study aids (β = .36, t(517) = 8.31, p < .01). Study aids was the strongest predictor of emotional engagement and can be interpreted to mean that with every one standard deviation increase in study aids, there is 0.36 average standard deviation increase in undergraduates’ emotional engagement with other predictors held constant. It was also noted that anxiety was the only negative predictor of emotional engagement in the final model.
Multiple Regression Results for the Final Five-Predictor Model of Emotional Engagement.
Note. LASS = learning and study strategies; INP = information processing; ANX = anxiety; ATT = attitude; CON = concentration; STA = study aids; β = standardized regression coefficient; SE = standard error; CI = confidence intervals.
To ascertain statistical conclusion validity and consequently validity of the study-based inferences, we performed model checking for the final selected five-predictor model. Consistent with the initial screening of the correlation matrix, multicollinearity was not a concern, since all variance inflation factor values ranged from 1.14 for anxiety to 1.50 for concentration, and were under the recommended threshold of 10 (Bowerman & O’Connell, 1990; Myers, 1990, as cited in Field et al., 2012). The visual inspection of the residual scatterplot is commonly used to assess normality, linearity, homoscedasticity, and independence of residuals. A normality probability plot for the model residuals showed no evidence of non-normality. The values of kurtosis and skewness were also close to 0, indicating consistency with the plots. To assess linearity and homogeneity of variance, fitted values were plotted against standardized residuals and no discernible patterns were detected. Being a study design, independence of residuals was always checked via assuring that participants provided responses independently. In addition, results of Durbin-Watson statistic also provided evidence of the independence of residuals (test statistics = 1.95, p = .532). Specifically, we accepted the null hypothesis that the model-based residuals were not auto-correlated. Collectively, model assumptions hold, allowing for valid statistical inferences of the final model results.
Discussion
The overall objective of the study was to examine the utility of the LASSs model in predicting Egyptian undergraduates’ emotional engagement using an advanced cross-validation technique. The LOOCV approach has been recommended in model selection to decrease prediction error and consequently increase the generalizability of the selected model in different samples. Results revealed that the best-fitting model included information processing, anxiety, attitude, concentration, and study aids, and explained 40% of the variation in undergraduates’ emotional engagement scores. These results are consistent with those who reported that emotional engagement can be predicted by students’ attitudes, interactions with their faculty and peers, and their positive feelings (Ainley & Ainley, 2011; Ben-Eliyahua et al., 2018; Elffers et al., 2012; Kahu, 2014; Mazer, 2013; Sagayadevan & Jeyaraj, 2012). Among the five predictors in the final selected model, the strongest was study aids.
To explain the association between undergraduates’ LASSs and their emotional engagement, it has been argued that the effective use of study aids in learning and studying enhances students’ active participation, interactions with peers and instructors, and emotional commitment with their learning (Wentzel, 2012; Zepke & Leach, 2010), which potentially supports their emotional engagement. Put differently, since the study sample included students from lab-based majors such as chemistry and mechanical engineering, there was a great potential to collaborate in the lab setting to exchange tools and instruments as a form of collaboration and consequently their emotional engagement. Similarly, students who have positive attitudes toward college are likely to be more emotionally engaged (Nayir, 2015). In more detail, when students feel positive toward attending school or college, they become more attached or emotionally engaged, since such learning environment brings enjoyment and becomes valuable to them. On the other hand, students who have a low risk of anxiety are more emotionally engaged (Palmgren et al., 2021), which explains why anxiety has a negative coefficient in predicting undergraduates’ emotional engagement. In other words, less anxious students are generally more interactive with their teachers or emotionally engaged in the learning environment (Palmgren et al., 2021), since previous researchers reported a negative correlation between anxiety and student-teacher interaction (e.g., Taye, 2017), which in turn enhances emotional engagement. Despite being cognitive strategies, students who effectively utilize concentration and information processing are more likely to be emotionally engaged in the learning environment due to the association between positive emotions and cognitive processing (Tyng et al., 2017).
Educational Implications
Before illustrating educational implications of the study, it is important to emphasize that the present study has two important contributions to the existing educational and psychological literature of undergraduates’ emotional engagement. For the first, it is likely the first study in which the LASSs model was utilized to predict or explain emotional engagement, which contributes to the literature of emotional engagement predictors using an advanced and robust statistical methodology—the LOOCV approach. Second, conducting the study on this new population has the potential to set the stage for comparing the predictive power of the LASSs model across other populations.
A number of educational implications emerge from the present study for students, faculty, and stakeholders (e.g., counselors and advisors), since it highlights the significance of five LASSs in predicting undergraduates’ emotional engagement. Based on the empirical evidence obtained in the study, it is suggested that students should be supported to interact positively with their peers and faculty and use study aids effectively to increase their emotional engagement. Additionally, it is recommended that faculty should create learning opportunities and design learning activities, which support undergraduates’ emotional engagement and consequently their learning outcomes (Zepke & Leach, 2010).
Instructors and stakeholders should pay much attention toward the relative contribution of LASSs in students’ emotional engagement, which can be utilized to make policy decisions. For instance, due to limited resources, instructors and stakeholders sometimes need guidelines for the most important LASSs, which can support undergraduates’ emotional engagement. The overall message for instructors and stakeholders is to utilize the five LASSs included in the final selected model starting with study aids to emotionally engage undergraduates in the learning environment. Specifically, since study aids strategy was the strongest predictor of undergraduates’ emotional engagement, it is important for instructors and stakeholders to create the learning environment that supports undergraduates’ use of study aids to enhance their emotional engagement. Put another way, if the purpose of funding is to increase emotional engagement, the selection will be for study aids due to its highly predictive power compared to other LASSs. Similarly, if instructors are to prioritize only one LASS to improve emotional engagement due to time or any other constraints, they should select study aids.
Learning and teaching centers at educational institutions should periodically assess undergraduates’ LASSs to gain more information related to students’ emotional engagement, since five strategies can explain more than one-third in their emotional engagement scores. Following this assessment, it is advised that students, who have low profiles particularly on the five LASSs included in the final selected model, should participate in intervention programs that enhance those strategies. To conclude, undergraduates’ emotional engagement has been a key factor for their academic success and consequently studying its predictors remains a subject for future research in education and psychology.
Limitations and Suggestions for Future Research
Results of the present study should be interpreted and generalized with some limitations in mind. First, it utilized a convenience sample, so researchers should be cautious while generalizing the results to other populations. Second, this study used only the LASSs model to predict undergraduates’ emotional engagement. Thus, future research may replicate the study using other learning strategies models such as the deep versus surface learning approaches to predict undergraduates’ emotional engagement. In addition, subsequent research may also investigate the predictive power of the model depending on the context of the study (e.g., undergraduates vs. graduates). We argue that if the goal is different, the engagement level may be different too. In summary, Future research may also consider using a cross-validated elastic net model for estimating regression coefficients as introduced by Zou and Hastie (2005).
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this study was provided by Taif University Researchers Supporting Project number (TURSP-2020/177), Taif University, Taif, Saudi Arabia.
