Life Science Undergraduate Mentors in NE STEM 4U Significantly Outperform Their Peers in Critical Thinking Skills

INTRODUCTION

Critical thinking is a skill routinely cited as preferred by employers over basic content understanding (National Science Board, 2010; Association of American Colleges and Universities, 2013; National Association of Colleges and Employers [NACE], 2014; New York Academy of Sciences, 2014; National Academies of Sciences, Engineering, and Medicine, 2016) and is a core learning objective of science education (Phillips and Bond, 2004; Carneval et al., 2010; Langdon et al., 2011; Dowd et al., 2018). Moreover, as the employment landscape becomes more competitive, it is imperative that students have the opportunity for a dynamic, well-rounded professional development experience at the college level. The acquisition of so called “soft skills” such as critical thinking translates across areas of content expertise, not excluding the sciences. Undergraduates (UGs) are increasingly being encouraged to become involved in activities such as mentoring and service learning to develop these soft skills. While there is demonstrated impact of mentoring and service learning on the mentees, there is little empirical evidence of the impacts of these activities on the UG mentors themselves (Carpenter, 2015; Nelson and Cutucache, 2017). Given this, we studied whether participating in the intervention of Nebraska Science, Technology, Engineering, and Math 4U (NE STEM 4U) as a UG mentor impacted the critical thinking skills of UG life science majors.

NE STEM 4U

The NE STEM 4U program is a professional training program for UG and graduate students in science, technology, engineering, and mathematics (STEM) majors who provide outreach using inquiry-based learning to students in grades K–8 (Cutucache et al., 2016). NE STEM 4U, as a program, utilizes a threefold training platform of teaching, mentoring, and research. Specifically, all participants must serve as teachers in the program and provide mentorship to youth, and all participants receive mentorship from faculty advisors and peers throughout the program.

The UG mentors in the program are students in STEM majors (predominantly from the life sciences) who have little or no background in working with youth, after-school programming, research, teaching, or any formal mentorship. Prospective mentors apply to the program and go through an interview process and a background check to work with youth. After a student is formally admitted to the program, he or she shadows a veteran group of mentors two to three times before being partnered with a veteran mentor(s) to implement his or her own teaching, using inquiry-based practices (i.e., not traditional, transmittal lecture). Using a team-teaching model, we aim to have two to three NE STEM 4U mentors per school. In addition to these practices, there are once-monthly “experiment nights” and “STEMinars.” At experiment nights, the mentors troubleshoot, in teams, the curriculum for that month. During STEMinars, we host speakers to present on a range of topics applicable to the out-of-school time teaching realm, such as classroom management, engaging special needs students, multicultural awareness, youth voice, relationship building, empathy, and understanding current budgetary challenges schools face. UG mentors also have the opportunity to participate in STEM education research.

Once the mentors are in the classroom, they engage the youth twice weekly at each school for 1 hour each time. Approximately half of our mentors mentor at more than one school per week. We estimate that the workload of 1 hour of teaching per week equates to a total of 4 hours invested for the mentor for teaching preparation and transportation to and from the site. In the classroom, the mentors engage no more than 15 young people at a time. To aid in relationship building, most schools ensure that the same K–8 students attend each time. The demographics of participating mentors/teachers is included in the Methods section and Supplemental Table 1. Of the students who have graduated from the program thus far (n = 117), 95.9% of them have both completed an academic degree in STEM and entered into the STEM workforce or graduate school upon graduation.

Critical Thinking

Critical thinking is delineated by a wide variety of definitions, although most agree with the philosophies of Socrates, Plato, and Aristotle (Gutek, 2009; Bailey and Mentz, 2015). Holyoak and Morrison (2005), describing critical thinking as an array of cognitive processes surrounding everyday life (Holyoak and Morrison, 2005; Dowd et al., 2018). One of the most cited definitions of critical thinking comes from what has come to be called “the Delphi Report,” in which 46 critical thinking experts across many disciplines came together to define critical thinking as “purposeful, self-regulatory judgment, which results in interpretation, analysis, evaluation, and inference, as well as explanation of the evidential, conceptual, methodological, criteriological, or contextual considerations upon which judgment is based” (Facione, 1990, p. 2). It is the Delphi Report that provides the foundation for the design of the California Critical Thinking Skills Test (CCTST), which has been used worldwide to measure critical thinking for more than 25 years (Insight Assessment, 2017) and was used in the current study.

Research on fostering critical thinking skills has primarily focused on how to teach critical thinking skills (Rowe et al., 2015; Paris, 2016; Dowd et al., 2018; Watanabe-Crockett, 2018). The majority of these studies investigated the influence of inquiry-based and problem-based teaching methods on thinking skills and showed that these instruction methods enhance the thinking skills of the learner (Greenwald and Quitadamo, 2014; Magrabi et al., 2018). Research has also been conducted in the context of peer- and near-peer–led team learning. For example, Gellin (2003), Quitadamo et al. (2009), and Smith (1977) found that peer interactions in the classroom had a positive impact on the critical thinking of UGs. Conversely, Snyder and Wiles (2015) found that serving as a peer mentor had no significant impact on critical thinking when mentors were compared with nonmentors. Beyond these types of studies, there exists a void in the literature regarding whether UGs who teach/mentor younger audiences (i.e., not peer or near-peer learning) enhance their own critical thinking skills. The few studies that are available on UG mentors typically focus on UGs who are pre-service teachers and typically do not measure changes in critical thinking (Malone et al., 2002).

Study Rationale

While some studies examine the effect of serving as a mentor from the UGs’ perspectives, the gap in the literature becomes especially pronounced upon review of the methods used in published studies, which consist primarily of qualitative, self-reported data being used to determine the impact on the UGs in various mentoring or teaching/teaching assistant roles (Holmes et al., 2013; Tenenbaum et al., 2014; Ward et al., 2014; Everhard, 2015; Walsh et al., 2015). While self-reported data are valuable as a good starting point for research or in-depth qualitative understanding of a phenomenon, they can be considered unreliable or biased and are listed as a limitation in many studies (Linn et al., 2015; Owen, 2017). Furthermore, qualitative data may not permit researchers to fully gauge how mentoring impacts specific skills such as critical thinking, which can be difficult to measure empirically (Gellin, 2003). Moreover, the length of time that an individual participates in an intervention (i.e., one semester vs. 1 year or more) may also play a role, as Snyder and Wiles (2015) did not find a significant gain in the critical thinking of peer mentors after they served as mentors for one semester, but there remains a gap in the literature involving interventions (inclusive of teaching by UGs) lasting longer time periods.

In the rare case that quantitative data are present in a published mentoring study, they typically are not the result of use of comprehensively tested instruments (Hannafin et al., 1997). This suggests that there is abundant opportunity for quantitative data collection and analysis in the mentoring literature, particularly studies that employ valid and reliable instruments. A recent study by Dowd et al. (2018) serves as a model example of such a quantitative study, as it employs the CCTST to examine the impacts of scientific writing on the critical thinking of UGs; however, the focus of this study is based on a classroom intervention and not on mentoring youth.

Taken together, the studies to date investigate how we should teach students (at all levels) to think critically, and how peer interactions may impact the critical thinking of UGs, but there exists a void in the literature regarding whether UGs who teach/mentor younger audiences enhance their critical thinking skills. Studies that use valid and reliable instruments for this measurement are noticeably absent. Therefore, we suggest that this is the first report using a well-validated assessment (the CCTST) of a link between teaching experiences by UGs who mentor younger audiences and critical thinking skills.

RESEARCH QUESTION AND STUDY DESIGN

In this study, we focused on the UG mentoring and teaching component of NE STEM 4U. This study aims to understand whether UG mentors demonstrated gains in critical thinking after at least two semesters of mentoring to K–8 students when compared with nonmentor UGs, using the CCTST. All of the individuals within both groups (mentors and nonmentors) were life science majors at the University of Nebraska at Omaha (UNO) who took similar courses during their matriculation and were comparable in distribution of gender, ethnicity, and class standing (Supplemental Table 1). The nonmentor life science UGs served as a control group and took the CCTST at the same time periods as the mentors. Using these two groups, this study was informed by the following research questions:

1. Does serving as a mentor impact the overall critical thinking of UG mentors compared with nonmentor life science UGs, as indicated by pre/post CCTST scores?

2. Are there specific subscales of the CCTST that indicate significant differences between mentor and nonmentor life science UGs?

METHODS

The NE STEM 4U program began in 2013, and since then, UGs have learned about NE STEM 4U at new student orientation, through flyers in hallways, or on university-sponsored student-group pages. The mentees come from participating K–8 schools that have a 1-hour after-school “enrichment” time, during which the UGs deliver an inquiry-based lesson in the form of experiments. UGs pick up participating students at the school cafeteria and take them to a classroom to carry out experiments and lessons. The UGs are solely in charge of the youth during the 1-hour enrichment window and serve as a replacement of public school staff during this time (as such cannot exceed a ratio of 15 students to 1 UG). Additional programmatic information, lesson plans, and other materials to replicate the program can be found at https://nebraskaomaha
.orgsync.com/org/nestem4u.

The demographics of the mentor participants in NE STEM 4U are ∼85% UGs and 15% graduate students; however, UGs are the focus of this study. UG mentors have an incoming GPA range of 1.5 to 4.0 (on a 4.0 scale) and most have declared a major in a STEM discipline. More detailed information about the general mentor characteristics is found in Nelson et al. (2017). Detailed demographic information regarding the parti­cipants in the current study is available in Supplemental Table 1. Additionally, the demographics of the youth (i.e., K–8 participants) are included in Cutucache et al. (2016) and Leas et al. (2017), but are summarized here: >50% of participants are African American, Latino, or declare both African American and Latino as their race, and just under 50% are Caucasian, Asian, or Pacific Islander students. All schools served in the Omaha Public School District have free or reduced lunch rates >47%, with the majority of the schools served being at 97% free and reduced lunch rates or higher (Leas et al., 2017).

Experimental Design

This quasi-experimental pre/posttest study used quantitative data from the CCTST to test the hypothesis that mentoring positively influenced the critical thinking of mentors in the NE STEM 4U program at the UNO when compared with nonmentor life science UGs. Groups were normalized to ensure matches for year in school, prior course work, and completion of science course work (n = 37). Informed consent was collected from all voluntary participants in accord with institutional review board regulations (IRB# 548-12-EX). This study took place over two academic years, with the same groups (NE STEM 4U mentor life science majors and nonmentor life science majors, respectively) and phases (quantitative pre/post) but different UGs each year.

Both mentor and nonmentor life science UGs took the CCTST at the beginning and end of the academic year (i.e., after two semesters of mentoring and course work or two semesters of course work only, respectively). The CCTST is a roughly 50-minute electronic assessment that provides an overall critical thinking score in addition to eight subscale scores: analysis, interpretation, inference, evaluation, explanation, induction, deduction, and (an optional test) numeracy. See Table 1 for a detailed definition of each measure provided by Insight Assessment (2017). The test is consistently updated based upon input from experts in fields such as assessment, psychometrics, measurement, statistics, and decision sciences, among others, and is based on the recommendations of the Delphi Report (Insight Assessment, 2017). According to the test designers, the subscores are not intended to represent completely independent factors; however, because many of the subscores are not inherently discrete units, they work together to represent the overall critical thinking ability of the student (Insight Assessment, 2017).

The questions used in the CCTST to measure reasoning skills come from a question pool that has been tested for over two decades by international measurement experts (Insight Assessment, 2017). This test is unique, because it is the only instrument that measures both cognitive and metacognitive skills, as recommended in the Delphi Report (Facione, 1990), and has been extensively evaluated for validity and reliability. A commonly cited definition of validity was provided by Eisenhart and Howe (1992, p. 1) as “the trustworthiness of inferences drawn from data.” In other words, how well does an instrument measure what it is thought to measure? Reliability is generally defined as “the degree to which an assessment tool produces stable and consistent results” (American Educational Research Association et al., 1985, p. 11).

Notably, many sources report on the robust validity of the CCTST (Williams et al., 2003; Sorensen and Yankech, 2008; O’Hare and McGuinness, 2015). Reliability tests for the eight subscales resulted in Cronbach’s alpha values ranging from 0.71 to 0.80 and a Cronbach’s alpha of >0.9 for the overall instrument (Facione and Facione, 1997), scores that indicate a strong instrument (Miller and Salkind, 2002). Additionally, the test has been used internationally across a wide variety of fields, including education research, science, nursing, psychology, and engineering, among others (Insight Assessment, 2017).

Analytical Procedures

All statistical tests were completed using SAS v. 9.4 (SAS Institute, Cary, NC). Before data collection, we estimated the sample size required to detect an effect using a power level of 80% and statistical significance cutoff of p ≤ 0.05 for this study (n = 11 mentors; n = 26 nonmentors; total n value = 37). After data were collected, we tested them for normality using the Anderson-Darling test. Scores of the subscale analysis were transformed using a reciprocal transformation to achieve normality. To assess whether mentors and nonmentors differed in the pretest, we conducted one-way analyses of variance (ANOVAs) with the effect group (i.e., mentor vs. nonmentor). To investigate whether mentors improved more than nonmentors, we conducted repeated-measures ANOVAs, which included the effects test (pre vs. post; the repeated measure), group (mentor vs. nonmentor), and interaction between test and group. Specifically, we used the interaction between test and group to test for differences in improvement between mentors and nonmentors.

RESULTS

At the beginning of the academic year (Supplemental Table 2), the overall CCTST score on the pretest did not significantly differ between life science students who were NE STEM 4U mentors or nonmentors, F(1, 35) = 3.32, p = 0.0771. However, mentors scored higher in the subscales inference, F(1, 35) = 4.92, p = 0.0332; interpretation, F(1, 35) = 5.18, p = 0.0291; and numeracy, F(1, 35) = 4.51, p = 0.0409 (Supplemental Table 2).

The repeated-measures ANOVAs showed that mentors increased their scores substantially in the subscales analysis, inference, and numeracy, while nonmentors showed no change in their scores (Table 2 and Figure 1; raw data in Supplemental Table 2). Although not significant, the overall score in the CCTST test, as well as in all the other subscales, showed a visually similar pattern on average (Figure 1, Supplemental Table 2, and Supplemental Figure 1), with mentors showing an increase in all scales from pre- to posttest and nonmentors demonstrating little to no change, or negative change, from pre- to posttest (Supplemental Table 2 and Supplemental Figure 1). Test (i.e., pre- vs. posttest) and group (i.e., mentor vs. nonmentor) were significant within the repeated-measures ANOVA for the overall score and some subscores (Table 2 and Figure 1). However, this is largely because of the pull of the mean; therefore, we present the ANOVAs to ensure a focus on the specific, significant gains.

DISCUSSION

The overarching objective of this study was to determine whether participation in the NE STEM 4U intervention (i.e., the professional development program for UG STEM majors) led to significantly improved gains in critical thinking skills. Specifically, we had two research questions: 1) Does serving as a mentor impact the overall critical thinking skills of mentors (compared with nonmentors)? 2) Are there specific subscales of the CCTST that indicate significant differences between mentors and nonmentors? For the first question, serving as a mentor does not statistically significantly impact student gains in critical thinking. However, mentoring does lead to statistically significant gains for student participants in analysis, inference, and numeracy—three subscales of critical thinking. We found that, in terms of overall critical thinking score, serving as a mentor did not have a significant impact, although a marginal increase was observed. Moreover, mentoring did statistically significantly impact the critical thinking subscale scores of analysis, inference, and numeracy.

Interestingly, previous studies (Madison, 2002; Golbeck et al., 2005) and the summaries of the skills (listed in Table 1) indicate a degree of relatedness between these subscale measures. Specifically, the three subscales of analysis, inference, and numeracy all relate to mathematical skills or quantitative literacy (Madison, 2002). Abilities in analysis and inference are also considered to indicate a higher level of quantitative literacy than basic numeracy or basic computational ability (Golbeck et al., 2005).

While it is not completely clear why UGs who mentor K–8 youth would show significant gains in measures related to math, the fact that mentors did display these gains post-mentoring is important, as studies indicate math skills are a strong predictor of future success (Trapmann et al., 2007). Trapmann et al. (2007) found that math grades were good predictors of future success for math, engineering, and natural science majors. Notably, Trapmann et al. (2007) found that, for engineering students, math grades were better predictors of academic success than an aptitude test specific to engineering. While the current study involved life science majors and not engineering students, it is interesting to note that mentoring significantly improved critical thinking abilities related to math skills.

In the current study, the observed increases in mathematical skills and quantitative literacy are likely due to the structure of the lesson plans/curricula for the NE STEM 4U program, as these are approximately equally balanced throughout the academic year to include math lessons in addition to science lessons. Moreover, many science lessons also include heavy use of mathematics, so this may be a by-product of frequency of exposure to and practice with these principles. This should be further explored in future studies. Additionally, more work should be done to understand why mentors did not demonstrate significant gains in the subscale scores of interpretation, evaluation, explanation, induction, and deduction. Specifically, the subscale of explanation, at least intuitively, seems to be an area that would be heavily used by a mentor/teacher of younger audiences, yet mentors did not show significant gains in this area. These questions remain to be discerned in future studies with larger sample sizes.

Another significant question on what is driving the improvement in critical thinking scores is whether it is the process of teaching itself (“learning better by teaching”) or whether it is the fact that the participants in the NE STEM 4U program are in a group dynamic (“camaraderie encouraging improvement”). The latter has been demonstrated by Springer et al. (1999) in a meta-analysis of decades of data on STEM UGs. We recognize the challenges in trying to tease apart the contribution of the process of teaching on the UGs as compared with the group dynamic. We have several studies ongoing with different cohorts of students (some in cohesive groups, others not) who serve as teachers for youth, and we expect to be able to address this limitation in future work, but it remains a significant barrier for understanding the precise contribution (if such a phenomenon can be determined) herein.

Overall, the findings in this study provide evidence that mentoring in NE STEM 4U improved critical thinking of the mentors when compared with nonmentor life science UGs, but more work needs to be done to further understand and corroborate these findings. For example, the findings of this study would be more robust if we had: a larger sample size, additional mentoring programs outside of NE STEM 4U, and a broader variety of STEM majors from different universities included. Additionally, the length of time that UGs participate in similar interventions (i.e., one semester vs. 1 year or more) should be further investigated to determine whether mentoring duration plays a role in critical thinking development, as Snyder and Wiles (2015) did not find a significant gain in the critical thinking of peer mentors after they served as mentors for one semester.

However, these preliminary findings do strongly suggest that serving as a UG mentor can improve critical thinking. Therefore, encouraging UGs to serve as mentors may be a way to fulfill the 21st-century skill development that many researchers say courses and other experiences are not meeting (Singer et al., 2012; NACE, 2014). In addition, serving as a UG mentor significantly improved quantitative skills such as analysis, inference, and numeracy, which are known to be strong indicators of future success for UGs in academics and their future careers (Trapmann et al., 2007). Overall, this quantitative study supports the findings of a previous qualitative study, wherein former UG mentors self-reported that they felt their experience improved their critical thinking (Nelson and Cutucache, 2017). More studies such as these should be conducted to provide strong empirical evidence of the impact serving as a mentor has on UG mentors.

We suggest that the incorporation of an innovative model that provides transferable skills to UGs for future employment, coupled with gains in critical thinking skills to apply to their course work and then ultimately to meet community stakeholder needs is a win-win-win. Finally, the levels of retention to academic degree completion as well as placement in the STEM workforce for NE STEM 4U mentors were significantly higher (i.e., 95.9%, as reported in the Introduction) than the national average, thus suggesting the importance of this program for recruitment and retention of STEM majors. Overall, this study suggests that serving as a teacher/mentor to younger audiences may lead to gains in specific subscore or components of critical thinking for UGs.

REFERENCES