==== Front J Microbiol Biol EducJ Microbiol Biol EducJMBEJournal of Microbiology & Biology Education1935-78771935-7885American Society of Microbiology 10.1128/jmbe.v19i2.1589jmbe-19-80ResearchLife Science Majors’ Math-Biology Task Values Relate to Student Characteristics and Predict the Likelihood of Taking Quantitative Biology Courses † Andrews Sarah E. *Aikens Melissa L. Department of Biological Sciences, University of New Hampshire, Durham, NH 03824* Corresponding author. Mailing address: Department of Biological Sciences, 117 Spaulding Hall, 46 College Road, University of New Hampshire, Durham, NH 03824. Phone: 603-862-2379. E-mail: sarah.andrews@wildcats.unh.edu.2018 31 7 2018 19 2 jmbe-19-8007 2 2018 04 6 2018 ©2018 Author(s). Published by the American Society for Microbiology.2018This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial-NoDerivatives 4.0 International license (https://creativecommons.org/licenses/by-nc-nd/4.0/ and https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode), which grants the public the nonexclusive right to copy, distribute, or display the published work.Expectancy-value theory of achievement motivation predicts that students’ task values, which include their interest in and enjoyment of a task, their perceptions of the usefulness of a task (utility value), and their perceptions of the costs of engaging in the task (e.g., extra effort, anxiety), influence their achievement and academic-related choices. Further, these task values are theorized to be informed by students’ sociocultural background. Although biology students are often considered to be math-averse, there is little empirical evidence of students’ values of mathematics in the context of biology (math-biology task values). To fill this gap in knowledge, we sought to determine 1) life science majors’ math-biology task values, 2) how math-biology task values differ according to students’ sociocultural background, and 3) whether math-biology task values predict students’ likelihood of taking quantitative biology courses. We surveyed life science majors about their likelihood of choosing to take quantitative biology courses and their interest in using mathematics to understand biology, the utility value of mathematics for their life science career, and the cost of doing mathematics in biology courses. Students on average reported some cost associated with doing mathematics in biology; however, they also reported high utility value and were more interested in using mathematics to understand biology than previously believed. Women and first-generation students reported more negative math-biology task values than men and continuing-generation students. Finally, students’ math-biology task values predicted their likelihood of taking biomodeling and biostatistics courses. Instructional strategies promoting positive math-biology task values could be particularly beneficial for women and first-generation students, increasing the likelihood that students would choose to take advanced quantitative biology courses. ==== Body INTRODUCTION Efforts by undergraduate biology instructors and discipline-based education researchers to better integrate mathematics into undergraduate biology curricula have intensified in the last decade (1–8). Such efforts are motivated by the need to prepare life science majors for advanced coursework and careers in modern biology, which has become increasingly reliant on quantitative skills as technological advances and modern computing have led to the generation of large data sets and to more sophisticated modeling and predictive capabilities (9). Thus, to be successful researchers, students training in biology also need to have well-developed quantitative skills and become adept at using and applying quantitative reasoning to biological problems (9–11). However, it is important to recognize the role motivation may play in mediating students’ success at completing quantitative biology tasks. According to expectancy-value theory of achievement motivation (12, 13), students are motivated to engage in a task when they place value on it. These values for accomplishing a task, called task values, include interest (e.g., enjoyment of the task), utility value (e.g., belief that the task is useful for their future goals), and cost (e.g., spending extra time on the task) (12–14). Theory and empirical evidence suggest task values predict both achievement and academic-related choices such as course enrollment intentions (12, 13, 15–17). Students’ perceptions of the utility value of mathematics (18) and their mathematics anxiety (18–20) have been linked to mathematics performance. Students with higher interest in mathematics (21–23) or lower mathematics anxiety (24, 25) are also more likely to choose a mathematics/science major and/or take more mathematics in college. We currently lack sufficient evidence of students’ values of mathematics in a biology context (i.e., math-biology task values) to make inferences about the role such values play in student performance or academic-related choices. Despite a lack of empirical evidence, it is widely believed that biology students are math-averse. For example, Thompson and colleagues (5) found that instructors believed biology students had negative attitudes about mathematics in their biology courses; however, when students were polled, their attitudes were much more positive than predicted by instructors. Others found that students often believe mathematics is useful for their life science career (26–30). In contrast, recent empirical work found STEM majors reported mathematics to be slightly uncomfortable, frustrating, and unpleasant (31). Moreover, the data suggested life science majors may have more negative attitudes than physical science or mathematics majors (31). Thus, additional research is necessary before concluding biology students are math-averse. Examining math-biology attitudes across a diverse group of students is especially important, as expectancy-value theory (12, 13) predicts that sociocultural background will inform students’ task values. Students who have internalized gender- or cultural-based stereotypes related to mathematics ability are predicted to have more negative attitudes toward mathematics than students who are not stereotyped. Women in particular are often stereotyped as having lower mathematics ability than men (32–35), and research has demonstrated that women report lower interest in mathematics (36, 37) and greater mathematics anxiety than men (18–20, 24, 25). Students who are the first in their family to attend college (first-generation students) may also be negatively affected by stereotypes, as they are often of low socioeconomic status (38), and such students are stereotyped as having lower academic ability (39). Therefore, first-generation students may have more negative task-values than continuing-generation students. In contrast to expectations, first-generation students in developmental mathematics at a community college reported better attitudes toward mathematics than continuing-generation students (40). Findings related to race/ethnicity are also unclear, particularly as they are understudied compared with gender-related differences (41). While underrepresented minorities are also stereotyped as having lower mathematics or academic ability (42, 43), some studies have reported no differences in mathematics attitudes by race (25, 31), and even when race/ethnicity-related differences in mathematics attitudes have been shown (19, 25, 41, 44), findings are not always consistent. For example, one study found Black students had higher mathematics anxiety than White students (19), whereas another study found no difference (25). In summary, there are significant gaps in our understanding of the relationships among math-biology task values, academic-related achievement/choices, and student characteristics, and these gaps inhibit our ability to design effective quantitative instructional materials for biology courses that would meet the needs of diverse learners. Here we describe our research addressing the following questions: 1) What are undergraduate life science majors’ math-biology task values (i.e., are biology students really math-averse)? 2) To what extent do students’ math-biology task values differ according to their sociocultural background? 3) Do students’ math-biology task values predict their likelihood of taking advanced quantitative biology courses? In particular, we hypothesized the following: 1) on average, undergraduate life science majors would report high math-biology utility value and cost, but low math-biology interest, 2) women, first-generation, and underrepresented minority students would report lower math-biology interest and utility value, but higher math-biology cost, than men, continuing-generation, and White students, respectively, and 3) students who report higher math-biology interest and utility value, and lower math-biology cost, would be more likely to take advanced quantitative biology courses. METHODS Setting and participants Survey invitations were distributed to introductory and upper-level biology courses at 19 institutions between fall 2016 and fall 2017 (additional details in Appendix 1). In an effort to include students from traditionally underrepresented groups, we contacted colleagues at three minority-serving institutions and sent a request through the Society for the Advancement of Biology Education Research (SABER) listserv to invite instructors at institutions with high minority and/or first-generation student enrollment to participate in our study. While our study still includes relatively few African American/Black or Hispanic/Latinx students, one of the institutions was a Historically Black College/University (HBCU) and two were Hispanic Serving Institutions (HSI); two other institutions rank well above the national average in terms of racial/ethnic diversity. The anonymous survey was administered online through Qualtrics. Students self-identified as 18 years of age or older and as a life science major and consented to participate in order to enter the survey. Students who completed the survey were compensated with a $5 gift card. In total, 1,190 students responded, but 129 were dropped from analyses due to missing responses and an additional 35 were dropped for reasons described in Appendix 1, resulting in a final sample size of 1,026 (see Table 1 for student demographics). This study was approved by the IRB at the University of New Hampshire (#6562). TABLE 1 Demographic and academic background of life science majors included in the study (n = 1,026). Description n (%) Gender  Male 297 (29)  Female 729 (71) First-generation status  Continuing-generation 882 (86)  First-generation 144 (14) Race/Ethnicity  Asian 243 (24)  Black or African American 69 (7)  Hispanic or Latinx 136 (13)  White (not Hispanic or Latinx) 483 (47)  Other 25 (2)  Multiracial 70 (7) Highest high school mathematics course  Calculus 684 (67)  Pre-calculus 214 (21)  Algebra 58 (6)  Statistics 70 (7) Year in college  First year 451 (44)  Second year 198 (19)  Third year 213 (21)  ≥ fourth year 164 (16) Pre-professional program  Not in a pre-professional program 353 (34)  Pre-med 531 (52)  Other science pre-professional program 142 (14) Honors status  Not in honors 806 (79)  Honors 220 (21) Institution type  Research 893 (87)  Comprehensive 110 (11)  Primarily undergraduate 2 (< 1)  Community college 21 (2) Measures The online survey contained the Math-Biology Values Instrument (MBVI) (30), two items on students’ likelihood of choosing to take advanced quantitative biology courses, and a demographic/academic questionnaire. The MBVI was developed and validated as a measure of undergraduate life science majors’ math-biology task values (interest, utility value, and cost) and contained 11 Likert-type items on a 7-point response scale ranging from “strongly disagree” to “strongly agree” (30) (Appendix 1). Each subscale showed high internal consistency (interest: α = 0.96, n = 1,015; utility value: α = 0.90, n = 994; cost: α = 0.85, n = 991). For each student, we calculated a mean score for each math-biology task value. Following the MBVI, two items asked students to rate how likely they were to take an elective Mathematical Models in Biology course and an elective biology course involving statistics. Both were on a 7-point response scale ranging from “Extremely unlikely” to “Extremely likely” (Appendix 1). The final portion of the survey contained items on students’ demographic and academic characteristics. Although we were primarily interested in students’ gender, parents’ education (to create a first-generation variable), race, and ethnicity, we also asked students to report their highest high school mathematics course (math-preparedness), year in college, and any science pre-professional program in which they were participating. To control for achievement and math-biology related experiences, we asked students to report their SAT or ACT Math score, whether they were in an honors program, and the number of college math-biology courses taken (biology courses where mathematics was used regularly or mathematics courses where biology was used regularly). Finally, we asked students to report their institution, to be used as a random effect (see Appendix 1 for all survey item options and complete explanations for coding variables). Data analyses All analyses were conducted in R (R Foundation for Statistical Computing, Vienna, Austria [https://www.R-project.org/]). To test for differences in students’ math-biology task values due to student characteristics, we used separate linear mixed-effects models (lme4 (45) and lmerTest (46) packages) with students’ mean scores for each math-biology task value (interest, utility value, and cost) as the dependent variable. In each model, gender, first-generation status, race/ethnicity, highest high school mathematics course, year in college, and pre-professional status were included as fixed effects, honors status and number of college math-biology courses were included as controls, and institution was specified as a random effect (random intercept only). We were unable to include students’ self-reported SAT/ACT Math scores as a control for achievement because many students did not report these scores. Models originally included a gender*first-generation status interaction term, but it was not significant in any model and was removed (data not shown). Models also originally included a gender*race/ethnicity interaction term. However, due to small sample size within some levels of race/ethnicity (Table 1), we did not have enough power to examine an interaction with this effect, so the interaction term was removed. To determine whether students’ math-biology values predicted their likelihood of choosing to take advanced math-biology courses, we performed separate linear mixed-effects models as above, using students’ biomodeling and biostatistics likelihood scores, based on students’ responses to a 7-point Likert-type question, as dependent variables. Students’ mean scores for interest, utility value, and cost were treated as fixed effects, and all fixed effects and control variables described in the preceding models were included as control variables here. All models were fitted using restricted maximum likelihood (REML) and only complete cases (n = 1,015 for interest, n = 994 for utility value, n = 991 for cost, n = 954 for biomodeling, and n = 955 for biostatistics). We used ANOVA to test for the overall significance of categorical fixed effects (type III with Satterthwaite approximation for degrees of freedom). When a categorical fixed effect with more than two levels was found to be significant, we conducted all post-hoc pairwise comparisons using Tukey’s method for p value correction (lsmeans package, weights = proportional (47)). Marginal R2 (proportion explained only by fixed effects) and conditional R2 (proportion explained by fixed and random effects) were obtained using the MuMIn package (Version 1.15.6; Bartoń K, [https://cran.r-project.org/package=MuMIn]). We checked assumptions of linear regression (multicollinearity, normality, homoscedasticity, autocorrelation) and found no issues (variance inflation factors were calculated using code found here: https://github.com/aufrank/R-hacks/blob/master/mer-utils.R). RESULTS Mean scores for students’ math-biology interest, utility value, and cost were 4.5, 5.6, and 4.8, respectively (SD = +/− 1.7, 1.2, and 1.6, respectively; Fig. 1). In other words, life science majors on average were somewhat interested in using mathematics to understand biology and saw mathematics as useful for their life science career but perceived some cost associated with doing mathematics in their biology courses. Correlations among all variables used in this study can be found in Appendix 2. FIGURE 1 Overall means (± 1 SD) for life science majors’ mathbiology task values (interest n = 1,015, utility value n = 994, cost n = 991). Effects of student characteristics on students’ math-biology task values Women reported almost a half-point lower mathbiology interest score (β = −0.41, p < 0.001) and higher cost (β = 0.27, p = 0.01) than men (Fig. 2A, Appendix 3). First-generation students reported lower interest and utility value (β = −0.40 and −0.28, p = 0.01 for both) than continuing-generation students (Fig. 2B, Appendix 3). Race/ethnicity was a significant predictor of cost (p = 0.04), with Asian (β = 0.28, p = 0.03), Hispanic/Latinx (β = 0.33, p = 0.04), and multiracial students (β = 0.50, p = 0.01) reporting higher cost than White students (Fig. 3, Appendix 3). However, when all pairwise comparisons were conducted using Tukey’s method for p value correction, no comparisons within race/ethnicity were significant (Appendix 3). FIGURE 2 Effect of (A) gender and (B) first-generation status on life science majors’ math-biology interest (n = 1,015), utility value (n = 994), and cost (n = 991). Bars are marginal means (averaged over all other effects in each model) with 95% confidence intervals (*** p < 0.001, ** p < 0.01, * p < 0.05, ns = not significant; p values based on modeled regression coefficients). FIGURE 3 Effect of race/ethnicity on life science majors’ mathbiology cost (n = 991). Bars are marginal means (averaged over all other effects in each model) with 95% confidence intervals. Indicators of significance are in comparison with the reference level, White (*** p < 0.001, ** p < 0.01, * p < 0.05, ns = not significant). Here we show uncorrected p values based on modeled regression coefficients; however, post-hoc comparisons using Tukey’s correction showed that none of the pairwise comparisons were significant (Appendix 3). Highest high school mathematics course, a measure of students’ math-preparedness, was a significant predictor in all math-biology task value models (interest and cost p < 0.001, utility value p = 0.008; Appendix 3). Pairwise comparisons demonstrated that students who completed Calculus in high school reported higher interest than students who had completed only pre-Calculus (β = −0.61, p < 0.001), Algebra (β = −0.93, p < 0.001), or Statistics (β = −0.97, p < 0.001) (Fig. 4, Appendix 3). Students who completed Calculus in high school also reported higher utility value than students who completed only Algebra (β = −0.48, p = 0.03) and reported lower cost than those who completed only pre-Calculus (β = 0.68, p < 0.001), Algebra (β = 1.47, p < 0.001), or Statistics (β = 0.80, p < 0.001) (Fig. 4, Appendix 3). Students who completed pre-Calculus also reported lower cost than those who had completed only Algebra (β = 0.79, p = 0.003; Fig. 4, Appendix 3). FIGURE 4 Effect of highest high school mathematics course taken on life science majors’ math-biology interest (n = 1,015, p < 0.001), utility value (n = 994, p = 0.008), and cost (n = 991, p < 0.001). Bars are marginal means (averaged over all other effects in each model) with 95% confidence intervals. Within each task value, bars with different letters are significantly different (post-hoc pairwise comparisons were conducted using Tukey’s method for p value correction). Year in college was also a significant predictor of students’ cost (p = 0.04), though the only significant difference was that students who were ≥ fourth year reported lower cost than students who were in their first year (β = −0.42, p = 0.03; Appendix 3). Students’ pre-professional status was not a significant predictor of any math-biology task value (Appendix 3). The proportion of variance explained by fixed effects (marginal R2) was 9% in the interest and cost models and 4% in the utility value model; adding in the random effect increased the proportion of variance explained (conditional R2) to 10% for the interest model and 9% for the utility value model but did not affect the proportion of variance explained for the cost model (Appendix 3). Predicting future course intentions from math-biology values Students’ likelihood of taking a biomodeling course was positively related to their reported math-biology interest (β = 0.63, p < 0.001) and negatively related to their reported cost (β = −0.18, p < 0.001) (Fig. 5, panels A and C; Appendix 4). Students’ likelihood of taking a biostatistics course was positively related to their reported interest and utility value (β = 0.45 and 0.14, p < 0.001 and p = 0.005, respectively; Fig. 5, panels D and E; Appendix 4). The proportion of variance explained by fixed effects (marginal R2) was 43% in the biomodeling model, and 26% in the biostatistics model; adding in the random effect did not affect the proportion of variance explained for either model (Appendix 4). FIGURE 5 Predicting students’ likelihood of taking a biomodeling course (A–C; n = 954) or biostatistics course (D–F; n = 955) from their math-biology interest (left), utility value (center), or cost (right) scores. Note that “likelihood” is based on students’ responses to a 7-point Likert-type question (responses ranged from “Not at all likely” to “Very likely”). Regression lines are model predicted with 95% confidence intervals (shaded area). DISCUSSION Consistent with our first hypothesis and recent findings (31), we found most life science majors in our study reported at least some cost associated with doing mathematics in their biology courses, and approximately one-third reported high cost (≥ 6 on the 7-point scale). However, we also found that, on average, students reported high utility value, which also supports our first hypothesis and is in line with multiple studies showing students generally find mathematics useful for their life science careers (26–30). In contrast to our first hypothesis, we found most students were at least somewhat interested in using mathematics to understand biology, and nearly one-third reported high interest (≥ 6 on the 7-point scale). Thus, while our findings lend some support to the belief that biology students are math-averse, we found that students are much more interested in using mathematics to understand biology than previously believed. While students’ math-biology task values were on average near-neutral at worst, there was high variation among students, with many reporting low interest and high cost in particular. Some of this variability can be explained by students’ sociocultural and academic background (discussed more below), but the conditional R2 for these models is low, suggesting that there are factors unaccounted for in our models. For example, perceived social support from parents (48, 49), peers (49), and instructors (50, 51) may all influence students’ mathematics attitudes and achievement. Future research on these and additional factors is needed to understand the unexplained variation in students’ mathbiology task values. Influence of student characteristics on math-biology task values We found significant gender-related differences in math-biology task values, with men reporting higher interest and lower cost than women, partially supporting our second hypothesis. Mathematics is frequently stereotyped as a male domain (32–35) and thus gender is predicted by expectancy-value theory of achievement motivation (12, 13) to influence mathematics task values. Despite decreasing or disappearing gender-related gaps in mathematics achievement (20, 52, 53), it is still common for women to report more negative attitudes toward mathematics than men (31, 35), in particular, lower interest (37) and higher anxiety (20). First-generation students reported lower interest and utility value than continuing-generation students, also partially supporting our second hypothesis. Lower interest may be due to cultural stereotypes related to socioeconomic status, though we would expect that negative cultural stereotypes related to academic ability would also influence cost (as discussed below for race/ethnicity). Alternatively, lower interest and utility value in first-generation students may be a result of a lack of cultural capital. As their parents never attended college, first-generation students are less likely than continuing-generation students to receive guidance from their parents on connections between curriculum and career expectations (54). This may explain their comparatively lower perception of the usefulness of mathematics for their life science career, which in turn may explain their lower interest in using mathematics to understand biology. Given our relatively small sample size (< 150) of first-generation students and the paucity of other research on their mathematics attitudes, additional research on larger populations of these students is necessary. Similar to others who have found differences in mathematics attitudes by race (19, 41, 44) or ethnicity (25), we found that in comparison with White students, Asian, Hispanic/Latinx, and multiracial students reported higher math-biology cost, also partially supporting our second hypothesis. While post-hoc comparisons using a p value correction showed no significant differences, we do not want to dismiss the possibility that this is a real finding that our study was underpowered to detect. Negative cultural stereotypes related to academic ability (43) could explain higher cost among Hispanic/Latinx students. However, Asian students are perceived as the model minority, with high academic aptitude in mathematics and science (55), and thus would not be predicted to report higher cost based on negative cultural stereotypes. One possible explanation for this unexpected result may lie in our broad race categories that likely mask important differences between ethnic groups. Our category of Asian encompassed all people originating from anywhere on the Asian continent, yet Southeast Asian students are stereotyped as academically inferior to White students, while East Asian students are stereotyped as academically superior to White students (55). Thus, disaggregating Asian data by ethnic groups may reveal nuances that could explain the observed results. Similarly, students selecting more than one race or ethnicity were binned into a single “multiracial” category; thus, the many different cultural backgrounds represented by this category make it difficult to speculate on why we see higher perceived cost in these students compared with White students. Finally, although we sought a diverse sample of students, we had a small sample size for Black or African American students (< 70). Further research utilizing much larger sample sizes of students from diverse cultural backgrounds is necessary before drawing conclusions about how race or ethnicity influences life science majors’ math-biology task values. Although we found significant differences in mathbiology task value scores related to gender, first-generation status, and race/ethnicity, these differences were small (Figs. 2 and 3; Appendix 3). However, subsequent regression analyses suggest the small differences observed between women and men and between first-generation and continuing-generation students may impact students’ academic-related choices. We re-ran the biomodeling and biostatistics regression models without task values and found that gender significantly predicted students’ likelihood of taking a biomodeling course (β = −0.51, p < 0.001) and first-generation status significantly predicted students’ likelihood of taking a biostatistics course (β = −0.51, p = 0.003). When task values were originally included in these regression models, neither gender nor first-generation status were significant (Appendix 4; discussed below). Thus, gender and generation differences in the likelihood of taking advanced quantitative biology courses appear to be mediated by gender and generation differences in task values, though additional research is needed to explore this possibility (e.g., path analysis). Further research should also determine whether the small differences in task values between groups differentially affects student performance on quantitative biology tasks. In addition to relationships with students’ sociocultural background, we found that math-biology task values were significantly related to students’ math-preparedness, measured as their highest high school mathematics course completed. Students who completed Calculus in high school reported higher interest and lower cost, which is consistent with expectancy-value theory (12, 13) and empirical research showing that students’ math-preparedness influences their mathematics task values (18, 19, 24). Collectively, as predicted by expectancy-value theory (12, 13), our findings demonstrate that undergraduate life science majors’ mathbiology task values are informed by their sociocultural and academic backgrounds. However, it should be noted that most students in our study came from research institutions. In particular, it is likely that community college students may have very different math-biology task values than students from research institutions. Further research on populations of students from community colleges and primarily undergraduate institutions is necessary to understand the math-biology task values of students attending these institution types. Math-biology task values predict future course intentions Interest was the strongest predictor of students’ likelihood of taking advanced biomodeling and biostatistics courses, which supports our third hypothesis and is consistent with expectancy-value theory (12, 13) and the findings of others (21–23). However, students’ perceptions of the usefulness of mathematics for their life science career were only positively related to their likelihood of taking a biostatistics course. This may be a result of students’ perceptions of the types of mathematics useful to life science careers. The use and interpretation of statistics is often stressed throughout the undergraduate life science curriculum. Therefore, when life science majors report high math-biology utility value, they may be primarily thinking about the importance of statistics. If so, utility value would be expected to predict the likelihood of taking a biostatistics course but would not necessarily predict the likelihood of taking a biomodeling course. Moreover, when evaluating the likelihood of taking an elective biostatistics course, students may have weighted its perceived usefulness more than the costs of taking the course, which would explain the lack of significance of cost in the model. In contrast, students may have put significant weight on cost when evaluating their likelihood of taking the biomodeling course due to the less familiar, and potentially more intimidating, nature of the course. Future research could utilize an explanatory sequential mixed-methods study, in which selected students are interviewed based on their survey responses, to better understand students’ motivational reasons underlying their likelihood of taking particular quantitative biology courses. Additionally, as we only measured students’ likelihood of taking hypothetical quantitative biology courses, future research could longitudinally track students to determine actual advanced quantitative biology courses they enroll in during their undergraduate life science career. While further research linking math-biology task values to students’ mathematics achievement in biology contexts is still needed, we believe that engendering positive mathbiology attitudes in students is in itself a worthy goal of quantitative biology instruction. Thus, instructors incorporating quantitative skills into their courses can design more effective quantitative instruction by implementing strategies that may lead to more positive math-biology task values. Although students generally perceived mathematics to be useful for their life science career, first-generation students may benefit from a utility value intervention, in which students make connections between course topics and their lives or other interests (56, 57). This type of intervention has also been successful in increasing interest in a topic (16, 56), and thus could also be beneficial for women and mathematically underprepared students. Student-centered learning methods focused on solving authentic problems (57–59) can also promote student interest in quantitative biology. Similar strategies have been shown to be successful at reducing mathematics anxiety, along with providing encouragement and positive reinforcement and making testing environments less intimidating (60). Based on our findings, anxiety-reducing strategies could be beneficial for women, first-generation, minority, and mathematically underprepared students. CONCLUSION Our findings suggest that, on average, undergraduate life science majors perceive some cost associated with doing mathematics in their biology courses and are therefore somewhat math-averse. However, students also generally see mathematics as useful for their life science career and are much more interested in using mathematics to understand biology than previously believed. As predicted by expectancy-value theory (12, 13), aspects of students’ sociocultural background and their math-preparedness informed their math-biology task values, which predicted their likelihood of choosing to take advanced quantitative biology courses. SUPPLEMENTAL MATERIALS Appendix 1: Additional methodology and details on the survey (instructions, items, and response options), Appendix 2: Correlations among all variables used in the study, Appendix 3: Unstandardized regression coefficients, standard error, and p values for fixed effects used in interest, utility value, and cost linear mixed-effects models, Appendix 4: Unstandardized regression coefficients, standard error, and p values for fixed effects used in likelihood of taking a biomodeling and biostatistics course linear mixed-effects models Click here for additional data file. ACKNOWLEDGMENTS We thank the faculty that distributed our survey invitation and reminders to their students. Comments from Carrie Hall, E. Erin Morris, and four anonymous reviewers greatly improved this paper. Support for this work was provided by a grant from the National Science Foundation (NSF DUE-1640347). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF. The authors declare that there are no conflicts of interest. † Supplemental materials available at http://asmscience.org/jmbe ==== Refs REFERENCES 1 Chapman BS Christmann JL Thatcher EF 2006 Bioinformatics for undergraduates: steps toward a quantitative bioscience curriculum Biochem Mol Biol Educ 34 180 186 10.1002/bmb.2006.49403403180 21638667 2 de Pillis L Adolph SC 2010 Mathematical biology at an undergraduate liberal arts college CBE Life Sci Edu 9 417 421 10.1187/cbe.10-08-0099 3 Goldstein J Flynn DFB 2011 Integrating active learning & quantitative skills into undergraduate introductory biology curricula Am Biol Teach 73 454 461 10.1525/abt.2011.73.8.6 4 Wightman B Hark AT 2012 Integration of bioinformatics into an undergraduate biology curriculum and the impact on development of mathematical skills Biochem Mol Biol Educ 40 310 319 10.1002/bmb.20637 22987552 5 Thompson KV Cooke TJ Fagan WF Gulick D Levy D Nelson KC Redish EF Smith RF Presson J 2013 Infusing quantitative approaches throughout the biological sciences curriculum Int J Math Educ Sci Technol 44 817 833 10.1080/0020739X.2013.812754 6 Hester S Buxner S Elfring L Nagy L 2014 Integrating quantitative thinking into an introductory biology course improves students’ mathematical reasoning in biological contexts CBE Life Sci Edu 13 54 64 10.1187/cbe.13-07-0129 7 Baumgartner E Biga L Bledsoe K Dawson J Grammer J Howard A Snyder J 2015 Exploring phytoplankton population investigation growth to enhance quantitative literacy Am Biol Teach 77 265 272 10.1525/abt.2015.77.4.6 8 Klug JL Carey CC Richardson DC Darner Gougis R 2017 Analysis of high-frequency and long-term data in undergraduate ecology classes improves quantitative literacy Ecosphere 8 e01733 10.1002/ecs2.1733 9 American Association for the Advancement of Science 2011 Vision and change in undergraduate biology education: a call to action: a summary of recommendations made at a national conference organized by the American Association for the Advancement of Science July 15–17, 2009 Washington, DC 10 National Research Council 2003 BIO2010: Transforming undergraduate education for future research biologists The National Academies Press Washington, DC 11 Association of American Medical College-Howard Hughes Medical Institute (AAMC-HHMI) 2009 Scientific foundations for future physicians Washington, DC 12 Eccles J Adler TF Futterman R Goff SB Kaczala CM Meece JL Midgley C 1983 Expectancies, values, and academic behaviors 74 146 Spence JT Achievement and achievement approaches W.H. Freeman and Company San Francisco, CA 13 Wigfield A Eccles JS 2000 Expectancy-value theory of achievement motivation Contemp Educ Psychol 25 68 81 10.1006/ceps.1999.1015 10620382 14 Eccles JS Wigfield A 2002 Motivational beliefs, values, and goals Annu Rev Psychol 53 109 132 10.1146/annurev.psych.53.100901.135153 11752481 15 Bong M 2001 Role of self-efficacy and task-value in predicting college students’ course performance and future enrollment intentions Contemp Educ Psychol 26 553 570 10.1006/ceps.2000.1048 11681832 16 Hulleman CS Godes O Hendricks BL Harackiewicz JM 2010 Enhancing interest and performance with a utility value intervention J Educ Psychol 102 880 895 10.1037/a0019506 17 Wigfield A Cambria J 2010 Students’ achievement values, goal orientations, and interest: definitions, development, and relations to achievement outcomes Dev Rev 30 1 35 10.1016/j.dr.2009.12.001 18 Pajares F Miller MD 1994 Role of self-efficacy and self-concept beliefs in mathematical problem solving: a path analysis J Educ Psychol 86 193 203 10.1037/0022-0663.86.2.193 19 Pajares F Kranzler J 1995 Self-efficacy beliefs and general mental ability in mathematical problem-solving Contemp Educ Psychol 20 426 443 10.1006/ceps.1995.1029 20 Miller H Bichsel J 2004 Anxiety, working memory, gender, and math performance Pers Individ Dif 37 591 606 10.1016/j.paid.2003.09.029 21 Gainor KA Lent RW 1998 Social cognitive expectations and racial identity attitudes in predicting the math choice intentions of Black college students J Couns Psychol 45 403 413 10.1037/0022-0167.45.4.403 22 Lapan RT Shaughnessy P Boggs K 1996 Efficacy expectations and vocational interests as mediators between sex and choice of math/science college majors: a longitudinal study J Vocat Behav 49 277 291 10.1006/jvbe.1996.0044 8980085 23 Maltese AV Tai RH 2011 Pipeline persistence: examining the association of educational experiences with earned degrees in STEM among US students Sci Educ 95 877 907 10.1002/sce.20441 24 Hackett G 1985 Role of mathematics self-efficacy in the choice of math-related majors of college women and men: a path analysis J Couns Psychol 32 47 56 10.1037/0022-0167.32.1.47 25 Hembree R 1990 The nature, effects, and relief of mathematics anxiety J Res Math Educ 21 33 10.2307/749455 26 Elliott B Oty K McArthur J Clark B 2001 The effect of an interdisciplinary algebra/science course on students’ problem solving skills, critical thinking skills and attitudes towards mathematics Int J Math Educ Sci Technol 32 811 816 10.1080/00207390110053784 27 Matthews KE Adams P Goos M 2009 Putting it into perspective: mathematics in the undergraduate science curriculum Int J Math Educ Sci Technol 40 891 902 10.1080/00207390903199244 28 Matthews KE Adams P Goos M 2010 Using the principles of BIO2010 to develop an introductory, interdisciplinary course for biology students CBE Life Sci Educ 9 290 297 10.1187/cbe.10-03-0034 20810961 29 Thompson KV Nelson KC Marbach-Ad G Keller M Fagan WF 2010 Online interactive teaching modules enhance quantitative proficiency of introductory biology students CBE Life Sci Educ 9 277 283 10.1187/cbe.10-03-0028 20810959 30 Andrews SE Runyon C Aikens ML 2017 The math-biology values instrument: development of a tool to measure life science majors’ task values of using math in the context of biology CBE Life Sci Educ 16 ar45 10.1187/cbe.17-03-0043 28747355 31 Wachsmuth LP Runyon CR Drake JM Dolan EL 2017 Do biology students really hate math? Empirical insights into undergraduate life science majors’ emotions about mathematics CBE Life Sci Educ 16 ar49 10.1187/cbe.16-08-0248 28798211 32 Hyde JS Fennema E Ryan M Frost LA Hopp C 1990 Gender comparisons of mathematics attitudes and affect: a meta-analysis Psychol Women Q 14 299 324 10.1111/j.1471-6402.1990.tb00022.x 33 Steele CM 1997 A threat in the air: how stereotypes shape intellectual identity and performance Am Psychol 52 613 629 10.1037/0003-066X.52.6.613 9174398 34 Spencer SJ Steele CM Quinn DM 1999 Stereotype threat and women’s math performance J Exp Soc Psychol 35 4 28 10.1006/jesp.1998.1373 35 Nosek BA Smyth FL 2011 Implicit social cognitions predict sex differences in math engagement and achievement Am Educ Res J 48 1125 1156 10.3102/0002831211410683 36 Campbell NK Hackett G 1986 The effects of mathematics task performance on math self-efficacy and task interest J Vocat Behav 28 149 162 10.1016/0001-8791(86)90048-5 37 Matarazzo KL Durik AM Delaney ML 2010 The effect of humorous instructional materials on interest in a math task Motiv Emot 34 293 305 10.1007/s11031-010-9178-5 38 Terenzini PT Springer L Yaeger PM Pascarella ET Nora A 1996 First-generation college students: characteristics, experiences, and cognitive development Res High Educ 37 1 22 10.1007/BF01680039 39 Spencer B Castano E 2007 Social class is dead. Long live social class! Stereotype threat among low socioeconomic status individuals Soc Justice Res 20 418 432 10.1007/s11211-007-0047-7 40 Otts CD 2011 Self-regulation and math attitudes: effects on academic performance in developmental math courses at a community college PhD Thesis University of Kansas Lawrence, KS 41 Else-Quest NM Mineo CC Higgins A 2013 Math and science attitudes and achievement at the intersection of gender and ethnicity Psychol Women Q 37 293 309 10.1177/0361684313480694 42 Blascovich J Spencer SJ Quinn D Steele C 2001 African Americans and high blood pressure: the role of stereotype threat Psychol Sci 12 225 229 10.1111/1467-9280.00340 11437305 43 Hudley C Graham S 2001 Stereotypes of achievement striving among early adolescents Soc Psychol Educ 5 201 224 10.1023/A:1014438702266 44 Chen C Stevenson HW 1995 Motivation and mathematics achievement: a comparative study of Asian-American, Caucasian-American, and East Asian high school students Child Dev 66 1215 10.2307/1131808 45 Bates D Mächler M Bolker B Walker S 2015 Fitting linear mixed-effects models using lme4 J Stat Softw 67 1 48 10.18637/jss.v067.i01 46 Kuznetsova A Brockhoff PB Christensen RHB 2017 lmerTest package: tests in linear mixed effects models J Stat Softw 82 1 26 10.18637/jss.v082.i13 47 Lenth RV 2016 Least-squares means: The R package lsmeans J Stat Softw 69 1 33 10.18637/jss.v069.i01 48 Gunderson EA Ramirez G Levine SC Beilock SL 2012 The role of parents and teachers in the development of gender-related math attitudes Sex Roles 66 153 166 10.1007/s11199-011-9996-2 49 Leaper C Farkas T Brown CS 2012 Adolescent girls’ experiences and gender-related beliefs in relation to their motivation in math/science and English J Youth Adolesc 41 268 282 10.1007/s10964-011-9693-z 21769612 50 Jackson CD Leffingwell RJ 1999 The role of instructors in creating math anxiety in students from kindergarten through college Math Teach 92 583 586 51 Beilock SL Gunderson EA Ramirez G Levine SC 2010 Female teachers’ math anxiety affects girls’ math achievement Proc Natl Acad Sci USA 107 1860 1863 10.1073/pnas.0910967107 20133834 52 Hyde JS Lindberg SM Linn MC Ellis AB Williams CC 2008 Gender similarities characterize math performance Science 321 494 495 10.1126/science.1160364 18653867 53 Lindberg SM Hyde JS Petersen JL Linn MC 2010 New trends in gender and mathematics performance: a meta-analysis Psychol Bull 136 1123 1135 10.1037/a0021276 21038941 54 Vargas JH 2004 College knowledge: addressing information barriers to college The Education Resources Institute (TERI) Boston, MA Accessed May 31, 2018 https://www.teri.org/files/pdf/research-studies/CollegeKnowledge.pdf 55 Museus SD Park JJ 2015 The continuing significance of racism in the lives of Asian American college students J Coll Stud Dev 56 551 569 10.1353/csd.2015.0059 56 Hulleman CS Harackiewicz JM 2009 Promoting interest and performance in high school science classes Science 326 1410 1412 10.1126/science.1177067 19965759 57 Harackiewicz JM Canning EA Tibbetts Y Priniski SJ Hyde JS 2016 Closing achievement gaps with a utility-value intervention: disentangling race and social class J Pers Soc Psychol 111 745 765 10.1037/pspp0000075 26524001 58 Yuretich RF Khan SA Leckie RM Clement JJ 2001 Active-learning methods to improve student performance and scientific interest in a large introductory oceanography course J Geosci Educ 49 111 119 10.5408/1089-9995-49.2.111 59 Renninger KA Hidi S 2011 Revisiting the conceptualization, measurement, and generation of interest Educ Psychol 46 168 184 10.1080/00461520.2011.587723 60 Haynes AF Mullins AG Stein BS 2004 Differential models for math anxiety in male and female college students Sociol Spectr 24 295 318 10.1080/02732170490431304