ARE THERE DIFFERENT PATTERNS OF LEARNING STYLES AMONG SCIENCE MAJORS?

Englander, Fred ; Terregrossa, Ralph A. ; Wang, Zhaobo 等

INTRODUCTION

In the last thirty years there has been substantial attention paid to the relative importance of learning styles in accounting for variations in individual student achievement. Advocates of a number of learning style theories have contended that a higher level of cognitive achievement can be realized if the faculty's instructional methods are more aligned to the learning style preferences of their students.

An alternate opportunity for exploiting the potential importance of learning styles in the learning process is to determine whether students' particular learning style preferences are relevant in the students' choice of major (and thus ultimately their choice of career). This study analyzes the behavior of the 1295 incoming freshmen at a major urban university in the northeast region of the United States for the three academic years beginning in 2004-2005. The focus of this study is to determine the extent to which students who selected one of the four science majors (biology, chemistry, pharmacy, physician's assistant) at the outset of their college careers could be characterized as having a pattern of learning style preferences which were distinctive from the patterns of learning style preferences of the alternate three science-related majors. The Building Excellence (BE) survey instrument (Rundle & Dunn, 1996-2009a), based on the Dunn and Dunn learning style model (Dunn, 2000) is the resource employed to measure learning style preferences.

To the extent that a particular pattern of learning style preferences associated with a given academic major can be identified, such an identification could potentially be deployed to advise prospective and existing college students into an academic major that represents a better match of the students' learning style profiles. However, the present authors contend that the research presented in this study is only a necessary but not sufficient condition for initiating a student counselling process in order to encourage students to select a major that better matches their pattern of learning styles. Follow-up research should also be done to determine whether a student who has chosen an academic major congruent with his/her learning style preferences (a) continues in his/her academic program (i.e., advancing the student retention goals of the university), (b) continues in the academic major selected at the outset of the student's academic program, (c) graduates the university within a particular period of time and (d) graduates with a grade point average that reflects well on the student's overall potential. However, this follow-up research, which may offer a university's counselling staff the ability to improve its services in helping students make a more efficacious choice of major, begins with the research pursued in the current study--an exploration of the different patterns of learning style preferences among students in different science-related majors.

LEARNING STYLES

The Dunn and Dunn (Dunn, 2000) learning style model is utilized to determine learning style preferences. This model hypothesizes that each student has a distinct learning style, i.e., approach to process and absorb incremental knowledge or skills. This model also hypothesizes that a student's learning style profile is distinguished by interactions among the six broad categories of learning style preferences (twenty-six preferences in all, technically referred to as 'elements') that influence the student's unique approach to learning. Those six categories of influential elements are described as perceptual, psychological, environmental, physiological, emotional and sociological. (These six categories are, respectively, represented in Tables 3 through 8)

The perceptual category reflects preferences for learning by listening, by seeing images, illustrations or pictures, by reading, through hands-on experience or learning by verbalizing. The psychological category is linked to preferences for either a reflective or compulsive approach to discussing the material, making decisions and solving problems as well as the student's thought processing method, hypothesized to include analytic or global processing methods. The environmental category relates to preferences for background sound versus silence, low or bright light, cool or warm temperature and formal versus informal seating. The physiological category considers preferences for intake of snacks or drinks while learning, the time of day when the student does his or her best work and whether the student needs to be moving while learning. The emotional category is based on preferences for internal versus external motivation, starting and finishing one project at a time, or persistence, conformity to instructional norms and a preference for internal or external structure. The sociological category takes into account the preferences to learn alone, with a partner, or in a group of peers and to learn with an authoritative or a collegial instructor.

The online BE learning style instrument (Rundle & Dunn, 1996-2009a) was used to identify the learning style preferences exhibited by incoming students that chose majors in biology, chemistry, pharmacy and physician's assistant. That instrument is based on the five-point, Likert scale answers to 118 questions. These answers allow for the identification of the twenty-six learning style preferences that comprise the aforementioned six categories.

REVIEW OF THE LITERATURE

Academic interest in the relative importance of utilizing learning style theories as an aid in the education process has been evident for well over three decades. During this time there have been critics of the learning styles concept (e.g., Coffield et al., 2004). However, there have also been research efforts (e.g., Charkins, O'Toole, & Wetzel, 1985; Hawk & Shah, 2007; Terregrossa, Englander, & Englander, 2008) that support the hypothesis that the proper use of learning style theory may play a constructive role in facilitating greater student achievement.

It has also been hypothesized that learning style theory may be a resource in helping students select an appropriate academic major. Honigsfeld and Schiering (2004) analyzed the learning style preferences of education majors and argued that student academic performance could be enhanced by utilizing teaching methods that corresponded to that group's pattern of learning styles preferences. The learning style model of Biggs, Kemper, and Leung (2001) was used by Skogsbert and Clump (2003) who studied the differences in learning style preferences of psychology majors and biology majors. Skoksbert and Clump (2003) found that there were statistically different patterns of learning style preferences between those two groups. They suggested that those differences could be useful in accounting for why students majoring in one of those disciplines might face a greater challenge in learning the content in a course in the alternative discipline. They inferred that such a challenge may be caused by incongruence between, for example, the biology major's typical pattern of learning style preferences and the typical course content presented in a psychology course. This same incongruence might account for why a psychology major would be at a relative disadvantage in learning the content presented in a typical biology course.

A second approach in examining the link between learning styles and choice of academic major would involve an analysis of a larger variety of academic majors. Hunsaker (1981) suggested that efforts along these lines can be traced to the work of Plovnick (1975). Plovnick (1975) found that Kolb's (1974) learning style model was well suited to matching students with different patterns of learning styles into various medical careers. Hunsaker (1981) also states that these findings were soon challenged by Wunderlick and Gjerde (1978) who enlarged Plovnick's (1975) sample and then did not find a statistically significant relationship between the selection of medical career majors and different patterns of Kolb (1974) based learning styles.

Terregrossa, Englander, and Wang (2015) undertook a similar study to examine whether differences in student learning styles would account for differences among students in the selection of one of three encompassing groupings of academic majors (Business, Science and Social Science). Terregrossa et al. (2015) utilized factor analysis to establish groupings of learning style preferences among the twenty-six learning style variables developed in the Dunn and Dunn (Dunn, 2000) learning style model. The factor analysis results found clear distinctions in the relative prominence of these groupings of preferences (i.e., factors) among the three categories of academic majors.

In the present study, by considering differences between individual majors rather than groupings of majors, it can be argued that the hypothesized link between learning style preference patterns and academic majors is being put to a more exacting test. That is, characteristic learning style profiles of students in specific majors might be less evident, i.e., more difficult to identify, among one or more of the disciplines that comprise one of the three broad categories. Analogously, it may not be very difficult to identify visual and behavioral differences between two broad zoological subfamilies such as canines and felines, but more subtle, less easily identifiable differences may exist among different types of felines or different types of canines.

Other studies have not found a significant link between learning styles and the selection of academic majors. Fox (1984) failed to find a statistically significant relationship between educators and health professionals in learning style profiles based on the Kolb (1974) learning style model. Jenkins (1991) utilized the Dunn and Dunn model (Dunn, 2000) to detect differences in learning style preferences relative to choice of academic major among 7164 freshmen enrolled in an historically Black, public university in Mississippi. Her analysis did not reveal any statistically significant differences.

One prominent approach for delineating learning styles is the Myers-Briggs Type Indicator (MBTI) (Myers & McCaulley, 1985). Hardigan and Cohen (2003) utilized the MBTI and found that among seven health care related occupations, there were observable differences in the learning style profiles for those individuals within each of the relevant occupations. However, Hardigan and Cohen (2003) posed the question of how much can be reasonably inferred from the existence of these different learning style profiles from one occupation to another. More specifically, they asked if it is pedagogically effective for professors to teach the relevant courses to students training to be, for example, optometrists in such a manner as to appeal to the dominant learning style of the trainees. Hardigan and Cohen (2003) asked whether students in a given program would soon become bored with the content material if such concepts were always being presented to them in the same manner (consistent with the dominant learning style of students in such a program). Further, even if we logically accept the idea that a given learning style profile is prevalent within a given occupation, Hardigan and Cohen implicitly wondered how important learning style profiles may be in determining the success of practitioners within that occupation. They stated that differences in innate ability or biology, which determine learning style preference, are not destiny.

Hardigan and Cohen (2003, p.6) even make reference to the developers of the MBTI instrument in stating, "All types (of workers) can and do perform in varying ways, depending upon the situation, the opportunity and the motivation to do so." This argument clearly suggests that even though learning style preferences may predict choice of academic major or occupation, a given pattern of learning style preferences may not necessarily predict a successful academic performance in that major or successful career in that occupation.

Stratton et al. (2005) also utilized the MBTI to test whether learning style preferences were helpful in establishing congruence of students within individual career paths. They found evidence that medical students with particular learning style profiles were more likely to gravitate to some medical specialties.

In their study of a sample of 458 freshman at a large Midwestern public university, Pulver and Kelly (2008) administered to the study's subjects a survey instrument based on the Myers-Briggs learning style model (Myers & McCaulley, 1985) and a separate survey of student interests. Students' native interests were measured for various occupational and non-occupational content areas. After accounting for the influence of the Strong Interest Inventory survey, the Myers-Briggs learning style model did not account for any significant additional effect in explaining the students' choice of major.

As indicated, an important research question underlying the current study is to examine the extent of differences among the learning style preferences among students who have chosen different academic majors within the realm of science (i.e., biology, chemistry, pharmacy and physician's assistant). In this study, it is hypothesized that students with a given pattern of learning styles may be expected to choose particular academic majors that focus on cognitive content that is congruent with those patterns of learning styles. However, the present authors believe that this emphasis on learning styles as a tool to guide choice of major decisions would only be logical if a second, related research hypothesis is also tested. That is, are students who select a major that reflects cognitive content that is congruent with their learning style preferences able to experience a higher level of academic success? This hypothetical success might be reflected in a number of ways--a higher grade point average within the major and, perhaps, across all courses, greater longevity in the college/university where the student began his/her studies and a reduced inclination for the student to change major during the student's undergraduate tenure.

It is important to note, however, that while past research has examined the role that learning styles may play in the selection of a college major, a careful search of the literature has not revealed an analysis of this logical next step in the process. That is, do students with learning styles profiles congruent with their major attain greater academic and career achievement? Shouldn't this second, related hypothesis be tested before we expect college counsellors or others who might offer advice to students about their major to rely on learning style analysis in offering such counsel? It is also noted that the present study does not examine whether students with congruent majors experience better academic or career outcomes. However, the present authors believe that a recommendation for such a follow-up analysis as a condition for the use of learning style analysis in the student counselling process represents a meaningful contribution.

There are two studies which appear to be somewhat relevant to the two-stage task of establishing criteria for sorting students into majors based on their characteristics and then determining whether students who chose majors that were congruent with their particular characteristics were able to perform better in such majors. In one study, Wessel, Ryan, and Oswald (2008) utilized a survey instrument to measure the students' 'perceived fit' between their academic major and their individual pattern of interests. Relative to the learning styles discussion presented earlier, this perceived congruence, or fit, between the pattern of individual interests and the selection of academic major is only roughly analogous to the individual learning styles approach to measuring congruence.

However, an important distinction between the learning styles framework discussed earlier (e.g., McCaulley & Martin, 1995; Hardigan & Cohen, 2003; Stratton et al, 2005; Pulver & Kelly, 2008) and the framework applied by Wessel et al. (2008) is that the learning style framework conceives of congruence based on the use of learning style preferences which are determined by students' responses to a particular learning style survey. That is, the process of establishing the student's trait profile and then utilizing the trait profile of students to sort them into various academic majors is based on particular criteria that are in place before the sorting. Alternatively, the Wessel et al. (2008) approach to determining congruence between a student's psychological preferences or traits (i.e., traits that are only very broadly akin to learning style preferences) and that student's choice of major is to offer the student an ex poste survey which gauges reactions or perceptions of that student to his/her existing major. It can be argued that if congruence is to be used as a tool to mentor students regarding whether they have made an appropriate choice of major, it is more appropriate to use a measure of congruence between that student's ex ante preferences and choice of major rather than between a student's choice of major and an ex poste measure of the student's preferences or traits.

The Wessel et al. (2008) study considered students who had selected one of twenty-two academic majors. The authors hypothesized that perceived fit of the students in their respective majors would be statistically, positively linked to academic performance, as measured by grade point average and student satisfaction, and inversely related to negative student behaviors such as academic withdrawal, likelihood of a change of major and avoidable student absences. However, Wessel et al. (2008) report that the statistical evidence that they examined offered very little to support their a priori hypotheses.

A second study that attempted to use student characteristics in order to establish within-major congruency and then determine whether such congruency would effectively predict academic performance was undertaken by Pozzebon, Ashton, and Visser (2014). This study examined the marginal contribution of the congruency between vocational interest (again, based on psychological attributes that are distinct from learning style preferences) and choice of academic major (among four groupings of academic majors--arts/humanities, business, science and helping/child oriented) to academic performance. Pozzebon et al. (2014) determined that although there were discernable distinctions in personality characteristics among the students in the four groupings of academic majors, a greater congruency of students' vocational interests to their chosen major grouping did not effectively predict student academic success after controlling for personality traits and cognitive ability.

The studies by Wessel et al. (2008) and Pozzebon et al. (2014) have been given particular attention in the present study. Although neither of these studies are based on an application of a learning style model to determine whether students with particular learning style preferences appear to be best suited to a particular choice of major, both studies do use alternate measures of student characteristics in order to establish congruency between those characteristics and choice of major. There are other comparisons between these two studies and the analysis undertaken in the present study. The present study argues that establishing a congruency between student characteristics and a particular major (as Wessel et al., 2008) sets a higher standard for differentiation than establishing a congruency between student characteristics and a broader, grouping of majors (as Pozzebon et al., 2014, did because of sample size limitations). The present study also argues that if the purpose of developing measures of congruency is to provide a basis for advising students as to appropriate choices of academic major, it is more useful to utilize a measure of congruency based upon characteristics that can be observed before the actual choice of major decision is made (Pozzebon et al., 2014) rather than base the measure of congruency on student characteristics that are not observable before the student's actual choice of major decision (as Wessel, et al, 2008, did).

However, it should be emphasized that both the Wessel et al. (2008) and Pozzebon et al. (2014) studies have recognized, as no previous studies utilizing various learning style theories as a vocational guidance resource have recognized, the importance of validating a measure of congruence between student characteristics and choice of academic major by testing whether that measure of congruence is statistically associated with greater academic success. The relevance of this validation step is reinforced by the fact that although both the Wessel et al. (2008) and Pozzebon et al. (2014) studies were able to establish congruence between student characteristics and choice of academic major, neither study was able to find that such congruence was associated with a higher level of academic performance.

Having made this criticism against earlier studies of learning style theories and choice of academic major, it should be pointed out that the execution of the validation step is a formidable task at most colleges and universities. Securing the permission and cooperation of college or university staff and administrators with responsibility for maintaining and protecting the accuracy and privacy of student records for a period of several years is not an easy or simple undertaking. The present authors hope that the present research will serve as an argument to persuade the appropriate administrators to grant us the proper cooperation and access. However, the present authors also believe that part of the contribution of the present study is to bolster the statistical, academic and logical case to help garner such administrative support.

METHODOLOGY

Sample and Data Collection

The sample for this study consists of 1295 incoming freshmen students at a major urban university in the northeast region of the U.S. for the academic years beginning in 2004, 2005 and 2006. These are students who selected one of the following four science majors in which there was an adequate sample size (n): biology (n=341), chemistry (n=49), pharmacy (n=712) and physician's assistant (n=193).

The BE survey (Rundle & Dunn, 1996-2009a) was utilized to identify all twenty-six learning style preferences that comprise the six preference categories of the Dunn and Dunn learning style model (Dunn, 2000) for each student in each major. The BE survey, taken online, is comprised of 118 questions answered on a five-point Likert scale. The proprietary BE scoring for each of the 26 learning style preferences includes (a) negative scores (indicating an inverse relationship between the Likert scale based response and the relevant preference), (b) positive scores (indicating a direct relationship between the Likert scale based response and the relevant preference), or (c) zero (the lack of a systematic relationship between the Likert scale based response and the relevant preference). These BE scores for each of the twenty-six learning style preferences range from negative fifty to positive fifty. Thus, there were a total of 8,866, 1,274, 18,512 and 5,018 learning style preferences measured for the biology, chemistry, pharmacy, and physician's assistant majors, respectively. The total number of learning style preferences for a major was analyzed by exploratory factor analysis independently of the other majors.

Data Analysis

Exploratory factor analysis was utilized to examine whether there are distinctive learning styles among the cohort of students majoring in one of the alternative science disciplines--biology, chemistry, pharmacy and physician's assistant. The VARIMAX method of orthogonal rotation was utilized to differentiate the patterns of learning style preferences that characterized the students who chose to major in one of the alternative science disciplines. In this way, the large number of collinear learning style preferences for a cohort of students were separated into a reduced number of independent, (i.e., non-collinear) common factors. Although, the same number of common factors was extracted for each major, the composition of each common factor was unique to each major. The factor loadings of the common factors for a specific major summarize the unique, multivariate dimensions of the students' leaning styles in a meaningful way. Meaningful in the sense that the common factors are congruent with a specific preference category, a thought processing method or some other unique learning style profile.

The exploratory factor analysis was conducted in several stages. First, each learning style preference was decomposed into a linear combination of common factors and a component unique to the specific preference. Standardized partial correlation coefficients, or factor loadings, between the common factors and each of the specific learning style preferences were estimated via a linear regression model. The estimated total variance of all the learning style preferences was composed of the variance of the common factors, or communality, and the variance of the unique component of the specific preference.

A conventional approach is to extract a factor that has an eigenvalue that is no less than one. Instead of explicitly utilizing a specific cut-off eigenvalue, an alternative approach commonly used is to set a specific limit to the number of common factors extracted. In this study, the limit was set at four factors. However, analogous to the coefficient of determination in regression analysis, there is no specific value that determines the actual number of common factors to be extracted.

In this study, four common factors were extracted for each major based on the criteria commonly utilized in factor analysis that prudently limits the total number of extracted factors to a manageable and informative few. Manageable in the sense that the reporting of the total number of common factors extracted, their eigenvalues, their factor loadings, or their correlations with the learning style preferences, is limited by space. Informative in the sense that only those factors that provided additional, meaningful insight that differentiated the learning styles of the alternative science majors were extracted, analyzed and reported.

In the next stage, the extracted factors were rotated in such a way as to assure that each factor is uncorrelated with the other factors (orthogonal rotation), that each factor is identified by only some of the learning style preferences and that each factor is associated with a distinguishing characteristic of the learning style model. More specifically, the orthogonal VARIAMAX rotation method was utilized which maximizes the sum of the variances of the factor loadings and results in factor loadings that correlate highly with only one extracted factor and do not correlate meaningfully with the other factors. Consequently, each extracted factor was distinguished by its high correlation with distinctive learning style preferences. For example, if the extracted factor correlates highly with preferences for background noise, soft lighting and informal seating, then the factor is distinguished by the environmental learning style category. Tables 1 and 2 summarize the eigenvalues and the factor loadings of the extracted factors, respectively, for the biology major. These two tables are included primarily to link the results of the factor analysis to the rankings of the alternative preferences that compose each of the six learning style categories as reported in Tables 3 through 8 for each of the four science majors.

The methodology described above may potentially provide an answer to the question that is the focus of this study: Are there distinctive learning styles among students majoring in one of several science disciplines--biology, chemistry, pharmacy and physician's assistant? The exploratory factor analysis with orthogonal rotation automatically orders the relative importance of the extracted factors based on the proportion of total variance of all the learning style preferences that is explained by the extracted factor. Furthermore, the standardized coefficients of the partial correlation between the learning style preferences and the extracted factor (i.e., factor loadings) automatically order the relative importance of the learning style preferences that comprise each extracted factor. Finally, the sign of each factor loading indicates the positive or negative relationship between the extracted factor and the learning style preference for each major.

The results associated with each of the six learning style categories (perceptual, psychological, environmental, physiological, emotional and sociological) are reported in Tables 3 through 8, respectively. The results summarize the relative order of importance of both the alternative extracted factors and the highly correlated learning style preferences that comprise each factor, thus identifying the distinguishing learning style profile of each major.

Table 3, for example, reports the rankings of the preferences in the perceptual category including, learning by listening (auditory), by seeing images, illustrations or pictures (visual pictures), by reading (visual word), through hands-on experience (tactile/kinesthetic) or learning by verbalizing (verbal/kinesthetic) for each of the four majors. Organizing the results in this way provides a direct comparison of the relative importance of the alternative learning style preferences in each category that characterize and distinguish the learning style profile for the alternative majors.

In Tables 3 through 8, the first column reports the learning style preferences that comprise the specific learning style category. The second, fourth, sixth and eighth columns report the extracted factor (Fi:j), where (i) goes from 1 to (f), representing the ith extracted factor, and (j) goes from 1 up to (l), representing the jth leaning style preference that highly correlates with the extracted factor, both (i) and (j) are numbered in descending order of importance. The third, fifth, seventh and ninth columns report the sign of the correlation coefficient between the extracted factor and the learning style preference, indicating the direction of the learning style preference in accordance with the Dunn and Dunn learning style model (Dunn, 2000). The same ranking method is included in Table 2, which reports the extracted factors, their associated learning style preferences and their factor loadings for the biology major. The ranks of the four extracted factors are reported in parentheses next to the factor loadings. These ranked factors correspond to the results reported in Tables 3 to 8 for the biology major.

For example, the result, + F3:3, reported for the biology major's environmental preference for sound in Tables 2 and 5 indicate. This result indicates a positive preference (+) for background sound when learning, that the importance of this extracted factor (F) in distinguishing the biology major's learning style profile was ranked third (F3) out of the four extracted factors, and that the preference for sound was rated third (F3:3) relative to the other six learning style preferences that correlated highly with the third extracted factor. By comparison, the result, + F1:1, reported for the physician's assistant major's environmental preference for background sound in Table 5. This result indicates a positive (+) preference for background sound when learning, that the importance of this extracted factor (F) in identifying the physician's assistant major's learning style profile was ranked first (F1) among the four extracted factors and that the preference for sound was ranked first (F1:1) relative to the other learning style preferences that correlated highly with the first extracted factor. Clearly, these results indicate that the environmental preference for sound is a substantially more important, distinguishing characteristic of the learning style profile for students in physician's assistant major than it was for those students in the biology major.

RESULTS OF THIS STUDY

Table 1 presents the eigenvalues of the correlation matrix for the biology major. The factors, their eigenvalues, the incremental difference between consecutive factors, the proportions of the variance of the common-factor components that is explained by the factor, and the cumulative proportions are reported in columns 1 through 5, respectively.

The eigenvalue for the fourth factor for biology majors equaled 1.817, increased the explained proportion of the variance of the common factor components by .07 and increased the cumulative proportion of variance explained by all four factors to .38, or approximately forty percent of the total variance of the communality of the learning style preferences for the biology major. The eigenvalues for the fourth factor was 2.8 for chemistry, 1.6 for pharmacy and 1.5 for physician's assistant majors. The eigenvalues for the biology, pharmacy and physician's assistant were similar, and the eigenvalue for the chemistry major is noticeably higher. In this case, the fourth factor for the chemistry major explained the variance of almost three learning style variables.

As described above, the orthogonal VARIMAX rotation method maximizes the sum of the variances of the factor loadings and results in factor loadings that correlate highly with only one extracted factor and do not correlate meaningfully with the other factors. Consequently, the extracted factor is distinguished by the highly correlated learning style preferences.

Table 2 reports the four extracted factors and their factor loadings for each of the 26 learning style preferences for the biology major. The relative importance, or rank, of the preference is reported in parentheses. The six learning style preferences that ranked first for the biology major, including four from the sociological category and two from the physiological category, are label F1:1 to F1:6 in Table 2 above and concomitantly in Tables 8 and 6, respectively. For example, the positive factor loading for the learning style preference for learning with a team is ranked first (has the greatest correlation with factor 1) and therefore is labeled F1:1 in Table 2 above and in row 3, column 2 of Table 8, and the positive sign is reported in the adjacent cell, row 3-column 3, for the biology major. This result indicates that biology major is characterized by a strong preference for learning with others.

Likewise, the eight learning style preferences that ranked second, including five in the perceptual category, two in the emotional category and one in the sociological category, are labeled F2:1 to F2:8 in Table 2 above and concomitantly in Tables 3, 7 and 8, respectively. For example, the positive factor loading for the verbal/kinesthetic learning style preference (learning by reading aloud) is ranked first for the second factor (has the highest correlation with factor 2) and therefore is labeled F2:1 in Table 2 above and concomitantly in row 5, column 2 in Table 3 with a positive sign in the adjacent cell, row 5, column 3. This result indicates that the biology major prefers to read aloud when learning.

The six learning style preferences that ranked third, including four in the physiological category and two in the environmental category, are labeled F3:1 to F3:6 in Table 2 above and concomitantly in Tables 6 and 5, respectively. As an example, the negative factor loading for the learning style preference for seating is ranked fifth for the third factor (has the fifth highest correlation with the third factor) and therefore is labeled F3:5 in Table 2 above and concomitantly in row 4, column 2 of Table 5. The negative sign (-) of the factor loading is indicated in row 4, column 3 of Table 5. This result indicates that the biology major prefers to learn while seated in an informal setting, such as sitting on a couch or lying in bed while learning.

The six learning style preferences that ranked fourth, including two from the environmental category, two from the emotional category and two from the psychological category, are labeled F4:1 to F4:6 in Table 2 above and Tables 5, 7 and 4, respectively. For example, the positive factor loading for the learning style preference for persistence, or starting and finishing one task at a time, is ranked first for the fourth factor (has the highest correlation with the fourth factor) and therefore is labeled F4:1 in Table 2 and concomitantly in row 2, column 2 in Table 7. The positive correlation is indicated by the + sign in the adjacent cell, row 2, column 3.

This method of organizing the results of the factor analyses in terms of the relative importance of the twenty-six preferences that compose the six learning style categories to distinguish the learning style profile of the biology major was applied to the chemistry, pharmacy and physician's assistant majors. In this way, the learning style profiles of the alternative majors were differentiated. These results are reported in Tables 3 through 8.

The results reported in Table 3 indicate that the perceptual category played a unique role in the learning style profile of the pharmacy major. Four of the five preferences that comprise the perceptual category are ranked first and one preference is ranked second for the pharmacy major. The signs of the perceptual preferences indicate that students majoring in pharmacy prefer to learn with hands-on experience, and prefer not to learn by listening to lectures, by reading text, by viewing graphs, diagrams or pictures or by reading aloud. None of perceptual preferences were ranked first for the biology, chemistry and physician's assistant majors.

The results reported in Table 4 indicate that the psychological category played a distinctive role only in the learning style profile of the physician's assistant major. Both preferences that comprise the psychological category ranked first for the physician's assistant major. The signs of the preferences indicate that the physician's assistant major was characterized by the global cognition method, or reasoning from a general conclusion to an understanding of the underlying principles.

The results for the environmental category reported in Table 5 indicate that the preference for light when learning ranked first for students majoring in both chemistry and physician's assistant. The negative signs of the preference for light for both majors indicate that the students prefer soft, low or dim light when learning. The environmental preference for sound ranked first only for the students in the physician's assistant major, and the positive sign indicates that those students preferred background sound when learning. None of the environmental preferences ranked first for the biology and pharmacy majors.

The results for the physiological category reported in Table 6 indicate that the time-of-day preference ranked first for the biology, chemistry and physician's assistant majors. The signs of the time-of-day preference indicate that the biology major had a strong preference for learning earlier in the day, the chemistry major had a strong preference for learning in the afternoon and the physician's assistant major had a strong preference for leaning in the evening. The preference for mobility also ranked first for the physician's assistant major. The positive sign of the preference indicates that the physician's assistant major learns best when moving while learning. Although, the preference for intake ranked first for both the physician's assistant and pharmacy majors, the signs of the preference indicate that the physician's assistant major prefers to snack or drink while learning, but that the pharmacy major does not.

According to the results reported in Table 7, the emotional category played an important role in the learning style profiles of the chemistry, pharmacy and physician's assistant majors. The preference for conformity ranked first for both the chemistry and pharmacy majors, but in contrary ways. The negative sign indicated that the chemistry major preferred the opportunity and flexibility to express what he/she has learned in an independent and unique manner.

Alternatively, the positive sign indicated that the pharmacy major readily conforms to the instructions of the instructor in expressing what he or she has learned. The preference for motivation ranked first only for the pharmacy major, and the positive sign of the preference indicates that the pharmacy major is motivated externally, perhaps by a parent or teacher. The preference for persistence ranked first for the physician's assistant major, and the negative sign indicates that the physician's assistant major prefers to work on several tasks or projects simultaneously. None of the emotional learning style preferences ranked first for the biology major.

The results reported in Table 8 indicate that the sociological category played an important role in the learning style profiles of the biology, chemistry and pharmacy majors. Four of the five learning style preferences that comprise the sociological category ranked first for the biology major. The preference for variety ranked first, and the negative sign indicates that the biology major prefers a routine method of instruction.

The three preferences indicating with whom the students prefer to learn all ranked first, and the signs of the preferences indicate that, in descending order of importance, the biology major prefers to learn as part of team, in a small group or with a partner. The same three preferences ranked first for the chemistry major, and the signs of the preferences indicate that the chemistry major also prefers to learn with others as opposed to learning alone.

The preference for variety also ranked first for the pharmacy major, but, contrary to the biology major, the sign of the preference indicates that the pharmacy major prefers a variety of pedagogical methods. The type of instructor, i.e., an authoritative or collegial teacher, also ranked first for the pharmacy major, and the sign of the preference indicates a preference to learn with an authoritative teacher. None of the sociological preferences ranked first for the physician's assistant major.

LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH

Among the many studies that have examined whether there are systematic relationships between choice of academic major and learning style preferences, the underlying assumption has been that if such relationships exist, then knowledge of such relationships could be used by academic counsellors at the high school and college level (and others seeking to advise students who are entering or already enrolled in college) to guide students as to what majors would be most appropriate to a given student. The results presented in this study for four separate science majors may potentially be applied in such a manner.

However, it has been argued in this study that establishing statistical relationships between students' learning styles and their choice of an academic major represents a necessary but insufficient condition for guiding a student's choice of a major. It remains to be tested whether students who have chosen a major that is congruent with those students' learning styles perform better academically. The formidable task of collecting the relevant performance data for such a validation is a direction for the authors' research.

CONCLUDING REMARKS

The results of the exploratory factor analysis utilized in this study provide ample evidence that supports the hypothesis that the alternative science majors can be distinguished by a unique learning style profile. The extracted factors and the signs of their factor loadings clearly differentiated the learning style profiles of the biology, chemistry, pharmacy and physician's assistant majors.

The biology major was distinguished from the chemistry, pharmacy and physician's majors in three key ways. Only the biology major exhibited a strong positive preference for learning earlier in the day. Contrary to the pharmacy major, the biology major had a strong preference for learning with a routine method of instruction. And although at least one of the preferences that comprise the emotional category played an important role for the chemistry, pharmacy and physician's assistant majors, none figured prominently in the learning style profile of the biology major.

The chemistry major was differentiated from the other three majors in one vital way. The physiological preference for time-of-day indicted that only the chemistry major had a strong preference for learning later in the afternoon. This time-of-day preference contrasts with both the biology and physician's assistant majors.

The pharmacy major was distinguished from the other three majors in several essential ways. Considering that four of the five preferences that comprise the perceptual category ranked first for the pharmacy major but none ranked first for the other three majors, this category played a distinguishing role only for the pharmacy major. In direct contrast to the physician's assistant major, regarding the physiological preference for intake, the pharmacy major preferred not to snack while learning. Contrary to the chemistry major, the pharmacy major exhibited a strong emotional preference to conform to the instructor's directions. And only the pharmacy major had a strong positive sociological preference to learn with an authoritative instructor.

The physician's assistant major was distinguished from the biology, chemistry and pharmacy majors in several important ways. It was the only major that exhibited a strong, positive preference for the global learning style, a component of the psychological category. It was the only major that revealed strong, positive physiological preferences for mobility when learning, for learning in the evening and for intake, or snacking, while learning. And it was the only major with a strong, positive environmental preference for background noise when learning. Finally, unlike the other three majors, none of the sociological preferences rated first for the physician's assistant major. The summary and analysis of the results highlighted the more salient distinctions between the learning styles that characterized the alternative science majors. An expanded, more thorough analysis would allow additional comparisons to be made. Nonetheless, it should be clear that the results reveal meaningful differences in the learning styles among students who chose each of the four majors considered here.

REFERENCES

Biggs, J., Kember, D., & Leung, D. Y. P. (2001). The revised two-factor Study Process Questionnaire: R-SPQ-2F. British Journal of Educational Psychology, 71, 133-149.

Charkins, R. J., O'Toole, D. M., & Wetzel, J. N. (1985). Linking teacher and student learning styles with student achievement and attitudes. Journal of Economic Education, 111-120.

Coffield, F., Moseley, D., Hall, E., & Ecclestone, K. (2004). Learning styles and pedagogy in post-16 learning: A systematic and critical review. London: The Learning and Skills Research Center. Retrieved from: http://sxills.nl/lerenlerennu/bronnen/Learning%20styles%20by%20Coffield%20e.a..pdf,

Dunn, R. (2000). Capitalizing on college students' learning styles: Theory, practice and research. In R. Dunn, & S. Griggs (Eds.), Practical approaches to using learning styles in higher education (pp. 3-19). Westport, CT: Greenwood Publishing Group, Inc.

Fox, R. (1984). Learning styles and instructional preferences in continuing education for health professionals: A validity study of the LSI. Adult Education Quarterly, 35(2), 72-85.

Hardigan, P., & Cohen, S. (2003). Comparison of learning styles among seven health professions: Implications for optometric education. The Internet Journal of Allied Health Sciences and Practice, 1(1), 1-7.

Hawk, T. F., & Shah, A. J. (2007). Using learning style instruments to enhance student learning. Decision Sciences Journal of Innovative Education, 5, 1-19.

Honigsfeld, A., & Schiering, M. (2004). Diverse approaches to the diversity of learning styles in teacher education. Educational Psychology, 24(4), 488-507.

Hunsaker, J. (1981). The experiential learning model and the Learning Style Inventory: An assessment of current findings. Journal of Experiential Learning and Simulation, 2, 145-152.

Jenkins, C. (1991). The relationship between selected demographic variables and learning environment preferences of freshman students of Alcorn State University, retrieved from ProQuest Dissertations and Theses, (Order No. 9216065).

Kolb, D. A. (1974). On management and the learning process in organizational psychology: A book of readings. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall.

McCaulley, M. H., & Martin, C.R. (1995). Career assessment and the Myers-Briggs Type Indicator. Journal of Career Assessment, 3(2), 219-239.

Myers, I. B., & McCaulley, M. H. (1985). Manual: A guide to the development and use of the Myers-Briggs Type Indicator. Palo Alto, CA: Consulting Psychologists Press.

Plovnick, M.S. (1975). Primary care career choices and medical student learning styles. Journal of Medical Education, 50, 849-855.

Pozzebon, J.A., Ashton, M.C., & Visser, B.A. (2014). Major Changes: Personality, ability and congruence in the prediction of academic outcomes. Journal of Career Assessment, 22(1), 75-88.

Pulver, C.A., & Kelly, K.R. (2008). Incremental validity of the Myers-Briggs Type Indicator in predicting academic major selection of undecided university students. Journal of Career Assessment, 16(4), 441-455.

Rundle, S., & Dunn, R. (1996-2009a). Building Excellence[R] (BE) Survey. Retrieved From: www.learningstyles.net.

Rundle, S., & Dunn, R. (1996-2009b). Building Excellence (BE) Survey 2000 Research Manual. Retrieved From: http://www.asb.dk/fileexplorer/fetchfile.aspx?file=7783.

Skogsberg, K., & Clump, M. (2003). Do psychology and biology majors differ in their study processes and learning styles? College Student Journal, 37(1), 27-33.

Stratton, T., Witzke, D., Elam, C., & Cheever, T. (2005). Learning and career specialty preferences of medical school applicants. Journal of Vocational Behavior, 67, 35-50.

Terregrossa, R. A., Englander, F., & Englander, V. (2008). A Production function for teaching microeconomics principles: What categories of learning style variables matter? International Journal of Education Research, 3(3), 11-20.

Terregrossa, R., Englander, F., & Wang, Z. (2015). Learning style theory as a potential tool in guiding student choice of college major. International Journal of Education Research, 10(1), 24-38.

Wessel, J., Ryan, A., & Oswald, F. (2008). The relationship between objective and perceived fit with academic major, adaptability, and major-related outcomes. Journal of Vocational Behavior, 72, 363-376.

Wunderlich, R. P., & Gjerde, C. L. (1978). Another look at learning style and medical career choice. Journal of Medical Education 53, 45-54.

Fred Englander

Fairleigh Dickinson University

Ralph A. Terregrossa

St. John's University

Zhaobo Wang

Fairleigh Dickinson University

About the Authors:

Fred Englander is a Professor of Economics at Fairleigh Dickinson University in Madison, New Jersey. He received his Ph.D. in economics from Rutgers University. Dr. Englander has published articles in the Southern Economics Journal, Business Ethics Quarterly, Journal of Academic Ethics and the Journal of Education for Business.

Ralph A. Terregrossa is an Associate Professor of Economics at St. John's University in New York City. He holds a Ph.D. in economics from Binghamton University. Dr. Terregrossa has published articles in The Quarterly Review of Economics and Finance, International Advances in Economic Research and Educational Review.

Zhaobo Wang is a Professor of Production and Operations Management at Fairleigh Dickinson University, Madison, New Jersey. Dr. Wang received a Ph.D. in operations research from Rutgers University. He has published articles in the Journal of Educational and Behavioral Statistics and the Journal for Economic Educators. Table 1 Eigenvalues of the Correlation Matrix: Biology Factor Eigenvalue Difference Proportion Cumulative 1 3.8922 1.5810 0.1497 0.1497 2 2.3112 0.2674 0.0889 0.2386 3 2.0437 0.2220 0.0786 0.3172 4 1.8217 0.2830 0.0701 0.3873 5 1.5387 0.1843 0.0592 0.4464 6 1.3545 0.1470 0.0521 0.4985 7 1.2075 0.0685 0.0464 0.5450 8 1.1390 0.1180 0.0438 0.5888 9 1.0210 0.0377 0.0393 0.6281 10 0.9833 0.1169 0.0378 0.6659 11 0.8664 0.0195 0.0333 0.6992 12 0.8469 0.1010 0.0326 0.7318 13 0.7459 0.0432 0.0287 0.7605 14 0.7027 0.0406 0.0270 0.7875 15 0.6621 0.0219 0.0255 0.8130 16 0.6402 0.0642 0.0246 0.8376 17 0.5760 0.0078 0.0222 0.8597 18 0.5682 0.0329 0.0219 0.8816 19 0.5353 0.0346 0.0206 0.9022 20 0.5007 0.1040 0.0193 0.9214 21 0.3967 0.0080 0.0153 0.9367 22 0.3888 0.0266 0.0150 0.9516 23 0.3622 0.0386 0.0139 0.9656 24 0.3235 0.0027 0.0124 0.9780 25 0.3209 0.0700 0.0123 0.9904 26 0.2508 0.0096 1.0000 Table 2 Factor Patterns--Biology Major Preference Factor1 Factor2 Team 0.80947 (F1:1) 0.00126 Small Group 0.79201 (F1:2) -0.00006 Alone/Pair 0.69854 (F1:3) 0.11958 Morning 0.50991 (F1:4) 0.10358 Late Morning/Early 0.49909 (F1:5) 0.09994 Afternoon Variety -0.34865 (F1:6) -0.20139 Verbal/Kinesthetic 0.10596 0.63389 (F2:1) Tactile/Kinesthetic 0.23165 0.59373 (F2:2) Auditory 0.02585 0.55757 (F2:4) Visual Word 0.11635 0.54351 (F2:5) Visual Picture -0.00174 0.52617 (F2:6) Conformity -0.07256 -0.41858 (F2:8) Authority 0.10462 -0.46436 (F2:7) Motivation -0.3017 -0.57732 (F2:3) Evening -0.13195 0.2258 Late Afternoon 0.07318 0.16806 Sound -0.1443 -0.19046 Intake 0.27634 -0.04241 Mobility 0.02142 -0.17788 Seating -0.18663 -0.04324 Persistence 0.06543 -0.00969 Light 0.10484 -0.07585 Structure 0.42752 0.29543 Temperature 0.08353 0.0371 Impulsive/Reflective 0.1915 0.16304 Global/Analytic -0.05224 -0.34068 Preference Factor3 Factor4 Team 0.14975 0.00666 Small Group 0.18817 -0.05047 Alone/Pair 0.1396 0.01643 Morning -0.29301 -0.26402 Late Morning/Early -0.17088 0.00514 Afternoon Variety -0.12598 -0.21234 Verbal/Kinesthetic 0.15061 0.11797 Tactile/Kinesthetic 0.08482 0.04875 Auditory -0.24244 -0.09226 Visual Word -0.21001 -0.01209 Visual Picture 0.3075 0.08905 Conformity -0.3557 0.00366 Authority 0.02361 0.10979 Motivation -0.04741 0.14387 Evening 0.67171 0.05316 (F3:1) Late Afternoon 0.62129 0.06685 (F3:2) Sound 0.48936 -0.22117 (F3:3) Intake 0.41916 -0.23112 (F3:4) Mobility 0.34996 -0.0889 (F3:6) Seating -0.41129 -0.0983 (F3:5) Persistence -0.09333 0.76447 (F4:1) Light 0.04043 0.44883 (F4:4) Structure -0.04574 0.43733 (F4:5) Temperature -0.02802 -0.14541 (F4:6) Impulsive/Reflective 0.04062 -0.59166 (F4:3) Global/Analytic 0.13508 -0.67501 (F4:2) Table 3 Factor Patterns of Perceptual Learning Style Category for Alternative Science Majors Biology Chemistry Learning Style Factor Sign Factor Sign Preference Pattern Pattern Auditory F2:4 + F3:4 - Visual Picture F2:6 + F2:2 + Visual Word F2:5 + F3:3 - Tactile/Kinesthetic F2:2 + F2:3 + Verbal/Kinesthetic F2:1 + F2:1 + Pharmacy Physician's Assistant Learning Style Factor Sign Factor Sign Preference Pattern Pattern Auditory F1:8 - F2:4 + Visual Picture F1:6 - F2:7 + Visual Word F1:7 - F2:5 + Tactile/Kinesthetic F2:4 + F2:1 + Verbal/Kinesthetic F1:5 - F2:3 + Table 4 Factor Patterns of Psychological Learning Style Category for Alternative Science Majors Biology Chemistry Learning Style Factor Factor Preference Pattern Sign Pattern Sign Analytic/Global F4:2 - F3:1 + Reflective/Impulsive F4:3 - F3:7 + Pharmacy Physician's Assistant Learning Style Factor Factor Preference Pattern Sign Pattern Sign Analytic/Global F3:2 + F1:3 + Reflective/Impulsive F3:4 + F1:4 + Table 5 Factor Patterns of Environmental Learning Style Category for Alternative Science Majors Biology Chemistry Learning Style Factor Sign Factor Sign Preference Pattern Pattern Sound F3:3 + F3:5 + Light F4:2 + F1:3 - Temperature F4:5 - F4:2 - Seating F3:5 - F2:4 - Pharmacy Physician's Assistant Learning Style Factor Sign Factor Sign Preference Pattern Pattern Sound F3:3 + F1:1 + Light F3:7 - F1:7 - Temperature F4:5 + F2:8 - Seating F2:5 - F3:5 - Table 6 Factor Patterns of Physiological Learning Style Category for Alternative Science Majors Biology Chemistry Learning Style Factor Sign Factor Sign Preference Patten Pattern Intake F3:4 + F2:5 + Early Morning F1:4 + F4:5 - Late Morning/Early F1:5 + F3:8 Afternoon Late Afternoon F3:2 + F1:6 + Evening F3:1 + F3:2 + Mobility F3:6 + F2:6 + Pharmacy Physician's Assistant Learning Style Factor Sign Factor Sign Preference Pattern Pattern Intake F1:9 - F1:6 + Early Morning F4:2 - F2:6 + Late Morning/Early F4:4 F3:4 + Afternoon Late Afternoon F4:3 + F4:2 + Evening F4:1 + F1:5 + Mobility F3:6 + F1:8 + Table 7 Factor Patterns of Emotional Learning Style Category for Alternative Science Majors Biology Chemistry Learning Style Factor Sign Factor Sign Preference Pattern Pattern Motivation F2:3 - F4:6 + Persistence F4:1 + F3:5 - Conformity F2:8 - F1:4 - Structure F4:5 + F4:4 - Pharmacy Physician's Assistant Learning Style Factor Sign Factor Sign Preference Pattern Pattern Motivation F1:2 + F2:2 - Persistence F3:1 - F1:2 - Conformity F1:4 + F4:5 - Structure F3:5 - F4:3 + Table 8 Factor Patterns of Sociological Learning Style Category for Alternative Science Majors Biology Chemistry Learning Style Factor Sign Factor Sign Preference Pattern Pattern Alone/Pair F1:3 + F1:2 + Small Group F1:2 + F1:1 + Team F1:1 + F1:5 + Authority F2:7 - F4:1 + Variety F1:6 - F4:3 + Pharmacy Physician's Assistant Learning Style Factor Sign Factor Sign Preference Pattern Pattern Alone/Pair F2:3 + F3:1 + Small Group F2:1 + F3:2 + Team F2:2 + F3:3 + Authority F1:1 + F4:4 - Variety F1:3 + F4:1 -