These psychodynamic factors illustrate societalattitudes reflecting a cultural divide between scien-tists and non-scientists identified more than 50years ago (Snow 1961). The pervasive negativeattitudes toward math can be attributed to cultur-ally embedded conceptions that math is difficultand accessible to only a few extraordinary individ-uals (Belbase 2013). Restrictive images of math,science, and research are perpetuated through sto-ries told by peers and parents, school experiences,and media representations (Hannover and Kessels2004; Murtonen et al. 2008). These attitudes arereinforced or even exacerbated through a steadystream of negative messages about research meth-ods courses, with sometimes exaggerated claimsabout difficulty, amount of work, and the caveatthat success depends on proficiency in math.Researchers have attributed students’ scienceaversion to the disconnect between the culture ofscience and students’ images of themselves(Hannover and Kessels 2004; Taconis and Kessels2009). Through a process of self-to-prototypematching (Niedenthal, Cantor, and Kihlstrom1985), students compare themselves to their idea ofa typical scientist; those whose self-concepts donot correspond to their perceptions of scientistsexperience self-to-prototype mismatch. Science islargely perceived by students as dull, abstract, andhard to understand. Furthermore, successfulengagement with science culture requires a “cer-tain way of being” in addition to particular person-ality traits (Taconis and Kessels 2009:1130).Congruence between student self-concepts andperceptions of science culture is associated withsuccess in math and science courses, while lack ofcongruence is associated with reduced interest inthose courses (Lee 1998). The perceived matchbetween self and prototype not only influences stu-dents’ affinity for the course but also their intendedcareer choice (Hannover and Kessels 2004).The concept of self-to-prototype matching canbe extended to explain students’ attitudes towardresearch methods courses. Influenced by mediaimages, societal attitudes, and personal percep-tions, social science students compare their self-concepts to their idea of a typical researcher, withmany students experiencing self-to-prototype mis-match as a result. I argue that research self-conceptis an important previously unidentified core con-cept in understanding perceived learning inresearch methods courses.Learning approach, a cognitive factor, predictsperceived learning. A surface learning approach(resistance) is associated with lower levels of per-ceived learning, while a deep learning approach(leaning in) is associated with higher levels of per-ceived learning. The extent of correspondencebetween self and prototype appears to inform stu-dents’ choice of learning approach. Students withmore positive math self-concepts, those who cansee themselves conducting research, and those whoconsider the course useful are more likely to take adeep learning approach. They lean in and adoptlearning behaviors and strategies that result inhigher levels of perceived learning. In contrast, stu-dents with negative math self-concepts, thoseaverse to performing research, and those who con-sider the course useless adopt a surface learningapproach. Since their goal is simply to pass thecourse, they adopt resistant learning strategies,which result in lower levels of perceived learning.The course instructor, an environmental factor,has a substantial impact on students’ perceptions oflearning. Characteristics of good instructorsinclude being student centered, professional, andenthusiastic. Students appreciate instructors whomthey feel are approachable and accessible. As far as

110 Teaching Sociology 45(2)Research culture. Students considered research sepa-rate and unfamiliar territory. Many (172) reportedthat prior to the course, they did not know exactlywhat research was and associated the term researchwith the natural sciences. Few were aware of anyconnection between research and their field of study.Students described the course as different (318) andmore difficult (325) than any they had previouslytaken. The language was “foreign,” and the conceptswere abstract, as one political science major put it: “Ifelt like a ten year old trying to read Shakespeare.”Many students (231) reported that in order to under-stand the material, they had to take a more active andengaged role “compared to being lectured to like inmost normal courses.” One communications majorsummed it up thus: “This course was a real cultureshock. It made my head spin.”Negative hype. Because so many students experi-ence these courses as difficult and they have suchhigh failure and withdrawal rates, they havebecome shrouded in negative hype. Students hadbeen subjected to a discourse of doom that shapedtheir attitudes toward these courses. This sociologystudent described how she felt on the first day: “Ifeel that all the negative publicity about the classprimed me in some ways to expect failure and tohave my world rocked when I stepped in.”These responses reflect the perception thatresearch and research methods are part of a sepa-rate, perhaps hostile, culture to which most socialscience students do not belong but that they mustpass through in order to complete their degree.Anxiety. Most students (454) reported feeling anxiousabout the course prior to taking it, and much of thisanxiety continued as the course progressed. Theiranxiety stemmed from multiple sources but derivedmainly from the negative hype surrounding thesecourses. This psychology major described the sourceof his anxiety: “I had heard a lot about this class. Ithad a 50% drop/fail rate. It would make or break youin this field of study. Most people can’t handle thedifficulty of research methods in psych and changemajors.” Students were worried about the level ofwork required (336) and passing the course (378).One student described the course as “separating thereal Psych majors from the fake ones.”Many students (161) reported feeling not ade-quately prepared for the course, and most (412)reported having some level of math or statisticsanxiety. Once the course began, students experi-enced continued anxiety because they were intimi-dated by the professor (190) and/or worried thatthey would not be able to understand the material(243). “A” students were particularly concernedthat they would not be able to make an A in thecourse and their GPA would suffer.Research self-concept. Since research was consid-ered to be a different world, those who inhabit itwere also considered different. One psychologystudent referred to herself as not being a member of“the nerd club.” This political science studentdescribed his instructor as being incomprehensible:“The main problem I had was with the teacher, Icould hardly understand him. He seemed like hewas from another planet. He just didn’t know howto translate his knowledge into what students couldunderstand.” Additionally, the things that research-ers were assumed to do both at work and in theirspare time seemed peculiar to students, as thiscommunications major related: “It is not somethingthat normal people do, looking at a journal articlewith research and data. That is not the daily readingsomeone would do for fun, or even pick up just tokill time. I can’t imagine doing that.”Many students (236) reported lacking the skillsnecessary to do research and had no desire to obtainthem. Most students (434) had little interest in doingresearch and could not foresee any use the coursemight have for them. Students had conceptualizedthose who do research as a particular type of person,someone who was different from themselves, andsomeone whom they did not want to be like or couldnot be like. They did not see themselves as research-ers; several stated, “I am not a research person” or “Iam just not into research.” These comments reflectanother dimension of academic self-concept,research self-concept, which represents an individu-al’s evaluation of his or her research ability.Learning approach: Resistance versus leaning in. Stu-dents with weak or negative research self-conceptsand those who found the course difficult reacted inone of two ways. One way was to resist learning,adopting what one student referred to as a “just getthe C mindset.” Resisting behavior included notcompleting assigned reading, not paying attentionin class, and not attending class regularly. Resistingstudents adopted surface learning strategies, as thiscommunications major described: “Once you fig-ured out how the teacher tested, the less you had topay attention in class. I only studied what I felt Ihad to.” Retrospectively, many (52), like this politi-cal science major, regretted their resistant behav-ior: “I admit the work I did was subpar. I wish I hadtaken this class more seriously.”

A total of 724 surveys were completed, for aresponse rate of 29 percent. The majority of partici-pants were women (74 percent). Almost two-thirds(63 percent) identified as white, 18 percent asAfrican American, 8 percent as Hispanic, and 10percent as Asian American and other. The mean ageof participants was 22.6. Just about one-fourth weresociology majors; the rest were psychology (23 per-cent), communication (20 percent), political science(17 percent), and criminal justice (16 percent)majors. Successful students were overrepresented;only 14 percent of participants failed to pass theclass with a C or above, compared to 23 percent ofstudents in the target population. See Table 1 for acomparison of course grade distribution betweenparticipants and all students taking these coursesover the three-year period. The sample reflected thedistribution of majors and racial/ethnic makeup ofthe target population, although women were some-what overrepresented (64 percent of the target pop-ulation were women). Participants’ mean GPA was2.98. To adjust for response bias, a weighted samplewas constructed on the basis of final course grade.Parallel sets of analyses were conducted with bothweighted and unweighted samples and yieldedanalogous results. The analyses reported in thisarticle are from the weighted sample.MeasuresTwo measures of achievement were used. One wasfinal course grade. The second, perceived learn-ing, was derived by summing responses to five

rior to taking the course, most students have a poorunderstanding of the nature of research (Bos andSchneider 2009; Earley 2014). They fail to see howresearch connects to their discipline of study (Leston-Bandeira 2013) and may feel forced into taking thecourse (Macheski et al. 2008). Students assume thatproficiency in math is necessary for success in thecourse, although most topics covered in the courseare not directly math related (Murtonen and Lehtinen2003; Parks, Faw, and Goldsmith 2011).Among students in a psychology research meth-ods course, perceived usefulness of the course waspositively associated with achievement as mea-sured by knowledge of course topics (Sizemoreand Lewandowski 2009). Although students’knowledge of research methods increased over thecourse, their perceptions of its utility decreased. Ingeneral, students fail to appreciate the importanceof the skills learned in the course. A majority ofstudents deem the course irrelevant and unneces-sary for their degree or their future careers (Earley2014; Macheski et al. 2008).Math anxiety was negatively associated withcourse grade among psychology students in aresearch design course while self-confidence inmath ability was positively associated with coursegrade (Núñez-Peña et al. 2013). Math anxiety waspositively related to perceived difficulty of coursetopics and negatively related to understanding ofcourse topics in a communications research meth-ods course (Rancer et al. 2013).Math anxiety consists of “feelings of tensionand anxiety that interfere with the manipulation ofnumbers and the solving of mathematical problemsin a wide variety of ordinary life and academic situ-ations” (Richardson and Suinn 1972:551). It is alearned emotional response (Tobias 1993), likely aconsequence of the widely held belief that individ-uals are either “math people” or “not math people.”A significant proportion of social science studentsapproach research methods courses with high lev-els of math anxiety (Bos and Schneider 2009;Macheski et al. 2008), statistics anxiety (Blalock1987; Onwuegbuzie and Wilson 2003), and generalanxiety about the course (Earley 2014).According to the implicit theories model(Dweck, Chiu, and Hong 1995), certain personalattributes, such as intelligence, are considered bymany to be either fixed or changeable. This conceptof trait beliefs has been extended to math as well,with important implications: Women who believedtheir math abilities were fixed were less interestedin math and less likely to pursue a career in mathcompared to women who believed their math abili-ties were changeable (Burkley et al. 2010). Mathself-concept, a dimension of academic self-concept,represents an individual’s evaluation of his or hermath ability (Marsh 1993). Research indicates astrong association between academic domain self-concept (e.g., math self-concept) and achievementin that domain (Marsh and Seaton 2013). In a studyof sociology students, a majority described them-selves as “non-mathematical persons,” concludingthat they could not learn research methods(Murtonen and Lehtinen 2003).Cognitive FactorsStudents experience research methods courses asmore difficult than other content-specific courses(Macheski et al. 2008). They have trouble under-standing abstract concepts and research terminology(Murtonen and Lehtinen 2003). In a study examiningthe effects of peer review on student performance in

In the interviews it became clear that when academics talked about undergraduateresearch experiences, they were talking about different things. When asked to providea definition of undergraduate research, some said undergraduate research referred toeverything that students did at university, whereas others said it referred only to aspecialised set of activities which were available only to a few students. Some academicsfocused on undergraduate research as being closely structured and guided whilst otherssaid that students needed to be doing research independently. Some academics con-sidered that if students were involved in various stages of the research process, forexample, collecting data or engaging in bibliographical exercises, then they would callthat undergraduate research but others considered that students had to be involved inthe whole process of research from setting questions or hypotheses, designing exper-iments or data collection right through to reporting the findings in some kind ofpublication.Another important distinction in undergraduate research conceptions was whetherthe focus was on skills development or on broader development including students’views of the role of research in their future lives. Academics also differed in theirviews of the quality of the research that students conducted. For some, student researchwas inevitably inferior; perhaps replicating existing research but not contributing toknowledge generation. For others, student research was real research which contributedto the generation of publishable new knowledge. Finally, some academics viewed under-graduate research within the confines of students’ courses of study where students weretreated merely as students, whereas for others undergraduate research was viewed ashaving a wider role in preparing future researchers where students were viewed asjunior colleagues and treated as such. Although the language used in different disciplinesvaried, perhaps surprisingly, these definitional differences did not appear to reflect dis-ciplinary differences. Responses are summarised in Figure 2. Implications are discussedbelow.558 A. BREW AND L. MANTAI

Discussion

Perceived academic attitudes, skills and mindsets

Approach

Introduction

The above results suggest quite strong-ly that early participation in undergraduateresearch promotes academic developmentamong social science and humanities stu-dents at Truman State University. Studentswho participate in undergraduate researchearly on report significant gains in the abil-ity to (1) think analytically and logically;(2) put ideas together, and note similari-ties and differences between ideas; (3) learnon their own and to find information theyneed to complete a task. Moreover, it wasfound that early participation in collabo-rative research was of particular benefitfor first-generation college students. Thisfinding seconds that of Nagda, Gregerman,Jonides, von Hippel, and Lerner (1998)who also contend that participation in col-laborative research assists in the academicintegration of "at-risk" students

Table 1 reports the differences in theStudent Independent Analytical Develop-ment Scores for students who reportedparticipating in collaborative research withfaculty versus students who did not. Asindicated in the results only 47.3% (67/129)of the social science and humanities firstor second year students who had not par-ticipated in research reported highdevelopment scores (greater than 2.75) asopposed to 74. 1% (20/27) of the studentswho reported participating in collabora-tive research with faculty. The differenceis statistically significant (chi-square = 6.42p = .01). These results indicate that earlyparticipation in collaborative research withfaculty as a freshman or sophomoreappears to be positively related to selfreported gains in independent analyticaldevelopment.Does early participation in collaborativeresearch with faculty have positive effectson "at-risk" students? Table 2 examinesthe relationship between early participa-tion in collaborative research with facultyand lihe academic development scores forfirst-generation college students who aresocial science or humanities majors. Afirst-generation college student is definedhere as a student where neither parent fin-ished a four-year baccalaureate degree.Indeed, there is a considerable amount ofliterature that points to the difficulties fac-ing first-generation college students. Fromthis perspective, first-generation collegestudents are at particularly high risk toleave school, due to features particular totheri such as the absence of a familial sup-port system familiar with college life, orlower self-esteem and lower sense of self-efficacy (Billson and Terry 1982; Suitor,1987; McCauley, 1988; London, 1989).In particular Billson and Terry, employingTinto's (1975) model of college persis-

MethodologyTo test the proposition that early par-ticipation in collaborative research with afaculty member enhances the developmentof students, a sample of students at TrumanState University was used in this study.Truman State University is the highlyselective public liberal arts and sciencesuniversity of Missouri. With a student pop-ulation of about 6000 (primarilyundergraduate) it is located in the rurallyisolated northeastern part of the state.The primary "independent variable"earlyv student involvement in research wasmeasured by using student responses on theCollege Student Experiences Question-naire (CSEQ) . The CSEQ is an instrumentdeveloped by C. Robert Pace and GeorgeD. Kuh and measures student progress andthe quality of students' experiences insideand outside the classroom. The CSEQ isdesigned to assess the quality of effort col-lege students expend in using the resourcesand opportunities provided by the institu-tion for their learning and development.In turn, it is assumed that the quality ofeffort is a key dimension for understand-ing student satisfaction, persistence. andthe effects of attending college. In partic-ular, the CSEQ asks students whether ornot they had worked with a faculty mem-ber in a collaborative way on a researchproject. The response to this question actsas the primary independent variable in thisstudy.The dependent variable, Student Inde-pendent Analytical Development isconceptualized as the student's ability toact as an independent analytical learner,and is measured by using three questionsthat appear in the CSEQ - the degree towhich students recognized personal gainsin (I) thinking analytically and logically;(2) putting ideas together, seeing relation-ships, and noting similarities anddifferences between ideas; (3) learning ontheir own, pursuing ideas and finding infor-mation they need to complete a task.

Recently, many scholars in higher edu-cation have argued that participation inundergraduate research is of great benefitto students (Boenninger and Hakim, 1999;Spilich, 1997). The many benefits citedinclude students (1) gaining experienceand learn about the research process byworking on an unsolved, open-endedresearch problem; (2) increasing their dis-ciplinary knowledge and theirunderstanding of how that knowledge maybe applied; (3) defining and refining theirresearch and career interests; (4) learnabout the world of academia and graduateschool life; (5) are provided with a forumfor collegial interaction with a facultymember (Alexander, Foertsch, Daffinrudand Tapia 2000; Nagda, Gregerman,Jonides, von Hippel, and Lerner 1998).Indeed, many point to student-faculty inter-actions outside of the classroom thataccompany undergraduate research asplaying a key role in the academic achieve-ment, retention and performance ofundergraduates (Astin, 1993; Pascarellaand Terenzini, 1991; Tinto, 1998). In addi-tion, ]_ddins, Nikolova, Williams, Bushkek,Porter and Kineke (1997), James (1998)Volkwien and Carbone (1994) contend thatparticipation in undergraduate researchmakes it more likely that students mastercomplex scientific concepts and developadvanced critical thinking skills that arebenelficial in promoting a sense of acade-mic integration among students (see alsoSakalys, 1984; Peppas, 1981). Still otherspoint to the individualized attention stu-dents receive from a faculty member whenworking on collaborative projects thatenhance student confidence (Koch andJohnson, 2000; Jacobi, 1991; Blackburn,Chapman and Cameron, 1981). Wubah,Gasparich, Schaefer, Brakke, McDonaldand D)owney (2000) and Gregerman (1999)contend that participation in undergradu-ate research dramatically improves theretention of minority students and studentswith low achievement. Ishiyama and Hop-kins, (2001a; 2001b) and Nnadozie,Ishiyama and Chon (2001) demonstratethat participation in research promotes theretention and timely graduation of first-generation and low-income collegestudents. However, as Spilich (1997,

Our systematic programmatic changes were drivenby a departmental mission for the integration and in-fusion of undergraduate and graduate education withscientific inquiry and scholarship. In faculty researchFigure 1. The effect of implemented changes on the number of students involved in undergraduate research.labs, graduate students mentored and supervised un-dergraduate students. These changes have also led toadditional curricular improvements with the creationof an advanced research experience as a psychologymajor senior capstone course. The capstone studentsare required to complete a research project that resultsin a poster presentation (Brewer et al., 1993).The multifaceted changes we implemented haveaddressed each of the problems we identified. Wehave increased students’ knowledge about undergradu-ate research, have created a more egalitarian environ-ment for students and faculty, and are regularly report-ing the success of student and faculty work. Severalother tangible rewards are evident. Changes in theuniversity culture welcome our departmental actions,and our students have the opportunity to participatein a university-wide research conference each springsemester. Over one third of our graduating seniors en-gaged in undergraduate research experience this past196 Teaching of Psychology

pplication procedure. We posted faculty research in-terests and created an application for students witha well-publicized deadline for participation in Fall2006. The application procedure did not prohibitlate assignments or individual students approach-ing faculty or vice versa, but ensured fair access toall students who met the deadline. After the dead-line passed, faculty discussed the applications andmade decisions about whom they would interview.To date, we have been able to accommodate ap-proximately 95% of those who apply. If we cannotaccommodate a student, we advise him or her toapply again or contact a social research laboratorylocated in the same college.2. Advertisement. We posted faculty research informa-tion and a student application on a dedicated bul-letin board by the psychology office and also onthe departmental Web site. Faculty announced thenew application procedures in their classes. We dis-played posters of previous student work in the hall-ways.3. Assessment and communication with majors. All stu-dents enrolled in statistics and research methodscourses (required for majors) received a brief ques-tionnaire to assess how much they knew about theircourse planning and whether they knew about ourdepartmental Web site and newsletter. The ques-tionnaire also reminded them about the importanceof extracurricular activities, including undergradu-ate research experience and fieldwork.4. Departmental newsletter. Every 2 months, the de-partment published a newsletter online. The goalof the newsletter is to increase a sense of commu-nity among faculty and students by highlighting stu-dent and faculty activities, accomplishments, andawards. Students involved in research are promi-nently featured in each newsletter, with photos,descriptions of research projects, and informationabout conference presentations.5. Restructured teaching assignments. We compensatedall faculty who actively engaged in research witha reassigned teaching load (3–2, instead of a 3–3), leaving eight unstaffed sections per year. Theability to make these changes involved other pro-grammatic changes and administrative support. Forexample, a comprehensive restructuring of our in-troductory psychology courses allowed us to reducethe number of sections offered by increasing en-rollments (from 11 sections to 8 sections a year)and adding four 10-hr teaching assistants. NAU’sCollege of Education supplied us with an advanceddoctoral student to teach four lower division coursesfor our department per year. This negotiated ex-change of resources compensated our departmentfor the resources we provide to their program and af-forded their advanced students vita-building teach-ing experiences in our department. Our graduateprogram restructuring also led to the elimination ofone course. These changes positively impacted fac-ulty morale and communicated that the integrationof teaching and research is an important aspect ofour discipline.

Previous work has documented the benefits ofundergraduate research experiences (Kardash, 2000;Landrum & Nelsen, 2002). Students gain technicaland interpersonal skills (Landrum & Nelsen, 2002),analytical, logic, synthesis, and independent learningskills (Ishiyama, 2002), while improving their abilityto gain entrance into competitive graduate programs(Kierniesky, 2005).Despite the established benefits of undergraduateresearch programs, many psychology programs do notoffer these valuable opportunities and the extent towhich students utilize these research opportunities isunknown. Perlman and McCann (1999) found that63% of psychology programs at bachelor-granting in-stitutions, 67% at master’s-granting institutions, and80% at doctoral-granting institutions have opportuni-ties for students to assist or work collaboratively withfaculty on research projects. Additionally, Perlman andMcCann (2005) reported that 34% of comprehensivecolleges and universities listed research experience asone of the 30 most frequently listed psychology coursesin 1997, as per course catalogs. Given the benefit ofincorporating and enhancing undergraduate researchexperience, surprisingly little discussion has focusedon how departments might better integrate undergrad-uate research into their curriculum and maximize itspositive impact for students, faculty, and the depart-ment as a whole. To address this gap, we describe howwe redesigned and enhanced faculty-led research op-portunities for our students at a midsized psychologydepartment.Undergraduate Research Experience atNorthern Arizona UniversityAt Northern Arizona University (NAU), fordecades, undergraduate students enrolled for 1 to 6credit hr of undergraduate research. Historically, therange of student involvement varied from indepen-dent student-led projects to significant contributionsto faculty-led projects. Although the benefit of engag-ing in research was apparent to involved faculty andstudents, we did little to proclaim the importance of un-dergraduate research participation to our departmentas a whole. In an effort to improve our ability to deliverthis critical dimension to student learning, we appliedsystematic and programmatic changes to improve ourundergraduate research program.We identified problems with our existing system.First, in the annual senior exit surveys students reportedthat they were unaware of undergraduate research op-portunities (lack of student awareness). A second prob-lem was that our lack of a formalized system led to aword-of-mouth system (unequal student access). Somestudents reported feeling too intimidated to approach

Rather, they suggest that it is notuniversally true that all students will benefit or benefit equallyfrom research-related experiences. As one major contribution, thiswork provides an instrument and methodology for probing simi-larities and differences at other institutions. The findings point toa need to look more closely at how to engage average and lowerability students more fully, to be cognizant of potential differencesassociated with gender, and to seek out the means to keep allstudents academically involved through the course of their under-graduate years.

The present findings are limited in several ways. First, the URQscale may be incomplete. There may be additional factors thatcould be added to the scale to better encompass the nature ofundergraduate research experiences. Furthermore, the data clearlyindicate disciplinary differences; however, more thorough inves-tigation and specification of what research means in specificdisciplines, as well as more thorough documentation of the natureand extent of participation, are necessary. Also, the present resultsare based on limited sampling of students from a large publicresearch university. Data from more institutions, including smallliberal arts colleges, may help to establish the generality of thecurrent results. More important, the present data point to poten-tially critical areas related to undergraduate research experiences,and this work provides an instrument and methodology for explor-ing these differences. Additional data will help in filling outdifferences that may exist between and within institutions. Finally,research described in this article developed a general purposesurvey, and one may ask “Is it appropriate?” and “Is it useful?”Factor-based instruments like the URQ provide a simple andefficient means to gauge the impact of academic programs thatsupport undergraduate research. There is, however, also a need topursue methods that more fully allow for understanding the re-search experience, like exit interviews, ethnographic interviews,and surveys.

The inclusion of two factors for which there were no clearpredictions, gender and major, provided additional insights. Gen-der predicted gains in Research Mindset and Research Methods,showing larger benefits for men compared with women. In thepresent data, men and women had similar profiles (see Table 1), soit is not evident why the genders differ. One possibility is thatwomen are treated differently or are mentored differently. Kahle(1990), for instance, found differential expectations for men andwomen in science classrooms and different patterns of interactionsby male and female instructors. This possible explanation, how-ever, goes against some of the research comparing men andwomen, which has suggested that they are treated similarly inresearch settings (Russell et al., 2007; Seymour & Hewitt, 1997).A plausible explanation based on the work of Halpern (2004) isthat men do better because they function more effectively inlearning and problem-solving environments that do not followstandard classroom models.An examination of benefits as a function of college majorshowed that the impact of major could be quite dramatic. Biolog-ical sciences majors showed greater gains in Research Mindset,Research Methods, and Peer Support compared with psychologymajors. This advantage can be explained by appeal to biologicalsciences majors’ significantly greater involvement in research ac-tivities. Consistent with the hypotheses here, more extensive par-ticipation in lab courses, more frequent faculty meetings, and moretime doing research (see Table 1) changes the academic culturein the two disciplines, producing a distinct advantage for one overthe other.The strength of the Taraban et al. (2008) work was in thedevelopment of the URQ through two iterations of data collectionand analysis involving approximately 700 students. That workpresented correlational findings relating the URQ to student be-haviors but did not probe more deeply into the dynamics ofresearch experiences. The present work, by including Gender andMajor variables, and interactions with College Credits and GPA,was able to test the prevailing ideology surrounding researchexperiences and several focused hypotheses.

The strongest predictors for Research Methods were Gender,followed by GPA, Research Hours, College Credits, and FacultyMeetings. The effect for Gender signals that men perceive greatergains in research skills than women, which is consistent with anadvantage for men in the Kardash (2000) investigation reportedearlier. One possibility is that female and male students are taughtdifferently (cf. Kahle, 1990). Differences in instructional methodsmay enhance men’s sense of self-efficacy regarding research skillscompared with women. The advantage for men may also reflecttheir facility in research contexts, compared with more traditionallearning contexts (Halpern, 2004). The association of faculty meet-ings with Research Methods indicates that faculty interactions playa role in acquiring the knowledge and skills associated withconducting research

n independence model 2 (496)
10084.62, p  .001, indi-cated that the hypothesized model was not random and was suit-able for further analysis. In the initial screening of results, theaverage off-diagonal absolute standardized residual value wasexamined. A value of 0 would indicate a perfect fit between thehypothesized and obtained models. The average off-diagonal ab-solute residual value of .048 fell within an acceptable range. Priorto examining the goodness-of-fit statistics, the model was testedfor violations of multivariate normality of variance using Mardia’scoefficient. A normalized coefficient value equal to 54.71 wassignificant. This violation of multivariate normality of variancemotivated using robust goodness-of-fit measures

The URQ. The URQ consisted of the 32 scale items reportedin Taraban et al. (2008), shown in the Appendix. An additional 19items associated with careers and critical thinking were included inthe current data set for exploratory purposes and are not consideredhere.Self-reported academic behavior and demographic ques-tions. In the questionnaires, URQ items were followed byquestions about the frequency or time engaged in research andassociated activities. A list of these questions is presented in Figure1. In our thinking, participants would provide more accurate judg-ments of participation if the timeframe for the questions werelimited. Therefore, the questions were worded with respect to theacademic year and hedged with the qualifier “approximately” (e.g.,“During this academic year, I spent approximately __ hours perweek conducting research”) to evoke responses from participantsthat were generally representative of their experiences within asomewhat broad, but limited, recent time frame. Participants alsoprovided demographic information (gender, age, major, completedcollege credits, completed credits for courses with labs) and theirGPA

n more detail, published research suggests that the majority ofstudents involved in research are high-ability juniors and seniors;however, the question of what happens to average and low-abilitystudents late in their college careers has not been posed. This issueforms the basis of the rival hypotheses that research experiencesare best suited for high-ability students late in their college careers.Based on the published literature, we further hypothesized thatgains from research experiences would be associated with stu-dents’ level of involvement (Bauer & Bennett, 2003). Specifically,we hypothesized that college credits for lab courses and hoursspent doing research would be associated with research benefits.Lab courses were considered important, because in a collegecurriculum, organized lab courses are often students’ first exposureto inquiry and research, and these courses help to build basicresearch skills. Faculty mentors are also important to successfulresearch experiences. When students reflect on why they becameinterested in research, they often present a teacher or mentor as acentral figure (Henne et al., 2008; Seymour et al., 2004). There-fore, we hypothesized that the frequency of faculty meetings, as anindicator of student–faculty involvement, would be associated withresearch benefits. It was less clear whether gender is a factor andwhat the impact of college major is on research benefits. Thus,these variables were examined, but no specific predictions weremade. Studies have also found that research experiences producegains in students’ communication skills (Bauer & Bennett, 2003;Kardash, 2000; Kremer & Bringle, 1990) and clarify students’goals for graduate school (Lopatto, 2004; Russell et al., 2007;Seymour et al., 2004). Therefore, several measures related to theseoutcomes were included in the present study: frequency of com-munication about research through papers, posters, and presenta-tions, and consideration of graduate schools

An interest in assessment has developed in parallel with theemergence of undergraduate research opportunities to test whetherthese academic experiences are successful in achieving their in-tended goals (Ahlm, 1997; Blockus, Kardash, Blair, & Wallace,1997; Hakim, 1998; Kardash, 2000; Spilich, 1997). Kardash(2000) developed a questionnaire to evaluate the research experi-ences of undergraduate interns who participated in mentored re-search in the sciences during a summer session or the academicyear and who were sponsored by the HHMI or National ScienceFoundation. Fourteen research skills were queried, with the largestpre-to-post gains in the ability to interpret data, to relate results tothe bigger picture in the field, and to orally communicate researchfindings. However, interns also expected to gain significantly moreat the beginning of the experience than they reported achieving atthe conclusion of the experience. Simply, the research experiencedid not live up to their expectations. Surprisingly, ratings for fiveefficacy items taken before and after the undergraduate researchexperience showed a significant decrement across the two mea-surements. These items asked students to rate their ability, interest,desire, and motivation to be scientists and their perceived supportfrom faculty. The results suggested that undergraduate researchexperiences can improve students’ ability to design, carry out, andreport on research but can also reduce students’ sense of researchefficacy and turn them away from science as a career. Lopatto(2004, 2009) developed SURE (Survey of Undergraduate Re-search Experiences) to assess HHMI summer research experi-ences. SURE included 20 items to measure learning gains, similarto those in Kardash (2000). The highest rated items related tounderstanding the research process, learning laboratory tech-niques, and understanding how scientists work on problems. Sey-mour, Hunter, Laursen, and DeAntoni (2004) collected interviewdata from undergraduate science majors who participated in aninternship in the summer before their senior year. About 20% ofthe interview comments indicated that the internship clarified,confirmed, or refined participants’ career plans and enhanced their

A substantial body of evidence has been compiled to support thebenefits of participation in undergraduate research (e.g., Taraban& Blanton, 2008). However, there are no known studies that haveattempted to assess the outcomes of research participation whilecontrolling for the effects of academic ability, college level, gen-der, and research discipline. If either of the rival hypotheses issupported, the viability of current initiatives to promote research toall students and to involve them in research at the earliest point intheir academic careers is called into question. If gender and majordiscipline are factors, then curricular changes may need to be madeto ensure that all students benefit from conducting research. Fur-thermore, the sustainability and expansion of undergraduate re-search programs depend in part on identifying the elements thatlead to productive and fulfilling research experiences (Blanton,2008). Once identified, they can be incorporated into curriculamore purposefully so that students may benefit from research

Examinations into gender differences related to undergraduateresearch experiences reveal differences in some areas but not inothers. Research suggests that men and women differ very little inthe variables that have been used to assess outcomes of researchexperiences. Bauer and Bennett (2003) found that male and femalealumni who had participated in undergraduate research programsshowed comparable gains in problem solving skills, literature andlanguage skills, and personal initiative and communication skills,compared with alumni with no programmatic experiences. Lopatto(2004) reported that benefits from research experiences and careerplans were comparable for men and women. Furthermore, femalescience majors generally report no differences in the ways they aretreated compared with men (Seymour & Hewitt, 1997), and men-tors’ appraisals of male and female science interns’ research skillsare similar (Kardash, 2000). Whether women have mentors whoare similar to themselves does not appear to matter (Russell et al.,

Providing research experiences for undergraduates has becomea widely recognized and accepted goal of colleges and universities,and significant effort has gone toward creating and maintainingopportunities for student research. Less attention has been given tothe dynamics of research experiences, that is, to peer and facultyinteractions in research settings, to the academic ability of studentswho get involved in research, to the impact of college curricula ondeveloping a positive attitude toward research and research skills,and to the effects of actual time spent on research. How, inparticular, do academic resources and students’ participationwithin undergraduate curricula shape students’ knowledge, skills,and beliefs about research?

Although self‑report surveys show that participants believe they learned science practices such as labtechniques, ability to analyze data, and skill in oral and written presentation, the studies could bestrengthened by measuring progress directly and by determining whether students have a coherentview of science practices (23,41,47–49). For example, students may develop data collection skills butlack ability to interpret results (25–27,30). One study that combined surveys, interviews, andshadowing of eight undergraduates as they interacted with a research team found that studentsprimarily set up and conducted experiments, rather than understanding the rationale for design or theinterpretation of results (40).These studies do help explain why duration affects outcomes from research experiences (23–25,27,37). During the first year, students, who are typically unfamiliar with the science concepts andtechniques of the lab, need orientation to the specific research project and slowly acquire thisknowledge. In 1 year or less, the duration of most UREs, students learn how to set up experimentsspecific to that lab and collect data but can rarely relate the analyses to a research question (30).Several studies found that adopting the traits, habits, and temperament (patience, perseverance,

These results raise the question of whether recruiting underrepresented students to UREs and CUREscoupled with appropriate mentoring could increase diversity in science. One program, SURE at EmoryUniversity, reports increasing diversity by bringing second‑year community college students to flagshipuniversities for UREs (41). They select students based on math preparation, experience with science,and success in overcoming challenges. Preference is given to first‑generation college attendees andstudents from underrepresented groups. Regression analyses of transcripts revealed that SUREparticipants took more science courses as seniors and earned higher grades in those courses thannonparticipants. A replication with randomized assignment to SURE would strengthen the findings

Typically, mentoring is shared among faculty, postdoctoral researchers, and graduate students withUREs offering an individual relationship with mentors and CUREs requiring students to share one orseveral mentors. Studies indicate that undergraduates interact most frequently with graduate studentsand postdocs, and less with faculty (26).

bout 70% of the research experiences were studied at 4‑year institutions.

Undergraduate research experiences provide a window on science in the making, allowing students toparticipate in scientific practices such as research planning, modeling of scientific observations, oranalysis of data. The experiences are intended to enculturate students into scientific investigation.Faculty, postdoctoral scholars, and other members of the lab mentor students. Ideally, mentors guidestudents to interpret authentic images of scientific research and link their experiences to their ownbeliefs or expectations. Interview studies document the many inconsistent ideas about scientificresearch that undergraduates develop. Many expect scientific research to mimic their college laboratoryexperiments. Others are unprepared for the failure rate in independent research. For example, onestudent said, “I honestly expected it to be like my organic chemistry lab that I just finished last year [...]I’m used to ‘here is the procedure, now get to it,’ and I thought that was what the experience would belike” [(11), p.1084]. In a post–research experience interview another student reflected, “I think thisexperience helped me to really understand that it’s not, like, a magical experiment and you come upwith magical data and some magical conclusion, and that it is frustrating, but you get through it, andyou get over it, and you’ll run it again and if it’s just as frustrating, you’ll do it again” [(12), p.65].To characterize the investigations of research experiences, analyze how they promote integratedunderstanding in science, and recommend improvements, we draw on research in the learning sciences.Specifically, we use the knowledge integration framework that synthesizes extensive research on inquiryscience to identify gaps and conundrums in the research on undergraduate research experiences(13–17). This framework calls for eliciting students’ initial ideas (consistent with hypothesizing) andencouraging students to test them against new ideas (18). To add new ideas, the framework documentsthe value of participating in personally relevant contexts, such as research experiences to make senseof science practices. The framework also highlights the value and importance of dynamic models ofscientific phenomena that reveal insights into unseen processes such as molecular reactions (19).Perhaps most importantly, the framework emphasizes that new ideas can be isolated and forgotten andhighlights the need to guide students to become adept at distinguishing among their initial ideas andthose they encounter in courses or research experiences to build coherent understanding (17). Finally,the framework builds on research showing that learners benefit from reflecting on their investigationsand observations to sort out and consolidate their ideas (20). This framework guides our analysis of theliterature on research experiences and our recommendations for improving them (Fig. 2).Fig. 2. Mentoring to promote knowledge integration.Successful mentors elicit ideas to find out what studentsthink, add relevant new ideas, encourage students to findevidence to distinguish among disparate ideas, and askstudents to reflect and consolidate their experiences.Distinguishing among research experiencesResearch experiences include Undergraduate Research Experiences (UREs) and Course‑basedUndergraduate Research Experiences (CUREs) (21). UREs feature individual students in faculty researchlaboratories and provide the opportunity for one‑on‑one mentoring (Fig. 3). Typically, students spendone or more semesters in labs, although the type of activity and form of mentoring varies substantially.Selection for UREs is highly competitive because few students can be accommodated. Using grades, test

Many claim that undergraduate research experiences improve preparation of the next generation ofscientists and increase persistence in science (1–3). The limited evidence for the impact ofundergraduate research experiences makes it difficult, however, to justify the substantial resources theyrequire. Of the 60 empirical investigations published during the last 5 years, over half rely exclusivelyon self‑report surveys or interviews to document outcomes, although such evidence has serious flaws(4) (Fig. 1). Fewer than 10% of the studies validate self‑reports with analysis of research products (suchas presentations or culminating reports), direct measures of content gains, longitudinal evidence ofpersistence, or observations of student activities. Although researchers often call for betterassessments, valid measures have yet to be designed (5–10). In addition, undergraduate researchprograms often select students who already intend to persist in science and primarily document thatthey continue to major in science. More nuanced indicators of success such as improved use ofscientific practices, increased ability to interpret original sources, or a better sense of possible flaws inresearch designs would strengthen the research base. We draw on the most convincing studies toidentify impacts and opportunities for future investigations. We identify mentoring as essential forsuccessful support of undergraduates considering careers in science. We call for studies that distinguishwhich types of undergraduate research experiences succeed for students with distinct interests,backgrounds, and preparation.Leave a comment (2)Prev | Table of Contents | NextAAAS.ORG FEEDBACK HELP LIBRARIANS All Science Journals Enter Search Term ADVANCEDScience Home Current Issue Previous Issues Science Express Science Products My Science About the Journal