“Do I need to read to do well?”: a post-pandemic case study of the reading habits and academic attainment of STEM and non-STEM students

Little is known about the reading habits and interests of college students in the post-pandemic environment, especially those of the understudied student populations of the Middle East. Thus, a convenient sample of female undergraduate students of Saudi Arabian descent was asked to report reading habits and interests. First, the study asked whether the responses of students organized by field of study (STEM versus non-STEM) differed. Then, it assessed the extent to which reading habits (as indexing reading comprehension skills) predicted college attainment (as measured by GPA) in each of the selected subsamples. Although habits and attitudes did not differ between STEM and non-STEM students, attitudes predicted reading habits differently for the two groups. Furthermore, habits and attitudes were better predictors of academic attainment for STEM than non-STEM learners. These findings underscored the need for institutions of higher education to carry out periodical assessments of their students’ reading habits and interests. Results can then inform library policies and instructional interventions that will benefit STEM students the most.


Introduction
College students' reading habits and attitudes toward reading have been frequent topics of surveys in the extant literature [1,2].The reason is simple.Reading is an advantageous skill in a variety of domains across a person's entire lifespan [3,4].For instance, reading is linked to improvements in a variety of performance indicators, such as reading comprehension, quality of writing in different styles, knowledge of vocabulary and grammar, and even spelling skills [5][6][7][8][9][10].In light of the vast amount of information available online and offline, it is not surprising that reading is related to critical thinking skills, such as the selection, analysis, evaluation, and assessment of information students encounter in their academic and personal lives [11].More generally, reading benefits brain connectivity, enhancing cognitive functioning [12] through the deployment of attention to select, direct, modulate, and maintain particular cognitive processes [13].As a result, engagement in academic (i.e., school-related) and extracurricular reading (i.e., independent voluntary reading not related to the demands of school work) is considered a critical factor in students' academic attainment [6,[14][15][16][17].Earlier in life, the frequency of home learning activities, such as book reading, consistently predicts children's cognitive development [18].Too often though research focuses on student populations of the Global North [2,19], thereby neglecting to assess whether student populations of the Global South differ.

Saudi Arabia as a society where youth and change intersect
The population of Saudi Arabia consists of a large number of young people [40].Since its inception, the country has heavily relied on fossil fuels for revenues and a patriarchal social structure for its functioning.Change has been brought about by the implementation of the 2030 Vision.The plan is a socio-economic blueprint of a sustainable economy that is diverse in its content and structure (i.e., a knowledge-and service-based economy), and human capital (i.e., it includes a workforce of both men and women treated as equal).In the push to reverse a rigid patriarchal system, women have been granted agency rights that in the past were relegated to men [41,42].The most notable changes have entailed broad and equal access to academic programs and professional occupations along with freedom of movement without male supervision.Female undergraduate students are the centerpiece of the plan through which a sustainable economy is to be realized.As such, their academic success and the personal and environmental conditions that may sustain or impair it are critical to the implementation of the economic plan [43][44][45][46].
Of course, in light of the 2030 Vision's reliance on innovation, students' attainment in STEM fields has taken center stage [47].Particular emphasis has been paid to female students' choice of STEM programs along with their academic success in such programs [48,49].Women's enrollment in STEM fields has substantially increased, albeit they remain noticeably underrepresented in engineering [50].They are overrepresented relative to men in non-STEM fields, such as business [51].Despite educational reforms and significant allocation of resources, improvement in both female and male students' STEM attainment is far from achieved when national and international indicators are considered [47].

Hypotheses and their rationale
The present study was intended to offer a glance at students' self-reported reading habits and interests in the postpandemic era.Interests were conceptualized as the value that learners attribute to reading (i.e., the relevance of positive attitudes toward it).The research was guided by the expectancy-value framework [52,53], according to which the likelihood of a given habit to develop and persist is shaped by attitudes principally defined by the value (i.e., importance) of the habit, such as reading practices, to the individual.Three key reasons justified the study's focus on the reading habits and interests of female undergraduate students as a function of their chosen academic field (STEM versus non-STEM major): (a) Little is known of Saudi Arabian female undergraduate students' reading habits and interests in STEM and non-STEM fields.(b) Although reading habits are known to be related to academic performance through the underlying reading comprehension skills that such habits are responsible for fostering [54,55], unclear is the relationship between reading habits or interests and academic performance in each field of study.(c) Given that habits developed during the pre-pandemic or pandemic periods may linger or change in the post-pandemic environment, evidence from the past may not inform the present.
Three sets of research questions and hypotheses were tested about the reading habits and interests of the selected student population in the post-pandemic learning environment.Each hypothesis is presented below with its rationale.

Attitudes and behavior: do the interests and habits of STEM and non-STEM students differ?
The extant literature regards STEM and non-STEM programs as qualitatively different educational paths, which tend to attract students with dissimilar habits, interests, and performance attainments [56].For instance, in a study comparing first-year STEM and non-STEM students at a large, 4-year public university, Gansemer-Topf et al. [57] found that the former had higher aspirations and commitment, as well as a higher academic self-efficacy (i.e., confidence in their abilities), academic adjustment (i.e., ability to cope with academic demands), and sense of belonging to the institution in which they were enrolled.STEM students also had a higher GPA at the end of the first year.For cognitive performance, Moè et al. [58] found that STEM students outperformed non-STEM students in mental rotation but were not different in verbal fluency, whereas Hausmann et al. [59] found non-STEM students to have greater verbal fluency.Thus, we hypothesized (H1A) that if students' choices of a STEM or non-STEM field of study underlined broader non-overlapping interests, qualitative differences would emerge between STEM and non-STEM students in their selection of reading materials, views of the activity of reading (e.g., an independent versus a socially influenced activity), and preferences for how reading is executed (e.g., reading versus being read).
Interestingly, successful performance in undergraduate programs heavily depends on students' reading skills [60,61].The reason is simple.Over 80% of learning in college derives from reading text materials independently [62,63].Because reading is key to all undergraduate courses, we expected no differences between STEM and non-STEM learners in the overall frequency of their reading activities (H1B).Yet, reading may be either intrinsically or extrinsically motivated.Reading for enjoyment underlies an attitude that can be conceptualized as one's intrinsic motivation for reading [64,65].In educational settings, intrinsic motivation is often linked to academic achievement, positive emotions and attitudes, and psychological well-being [66,67].Thus, we thought that intrinsic or extrinsic motives could be treated as the basic sources of students' reading habits.We expected (H1C) extrinsically motivated reading to be somewhat less intensely pursued than intrinsically motivated reading due to evidence underscoring that college students do not thoroughly read the textbooks assigned to the courses in which they are enrolled [68,69].However, no differences in the extrinsic motives (e.g., reading guided by external demands) of STEM and non-STEM students were predicted since most college courses contain extensive reading demands (H1D). 1 3

Do interests predict habits?
According to the expectancy-value theory of motivation, the value of reading for individual learners is linked to its frequency.Thus, it was hypothesized (H2A) that positive attitudes toward the activity of reading would predict the frequency with which reading was reported by participants.However, compared with extrinsically motivated reading, intrinsically motivated reading was expected to better predict the frequency with which reading was reported by both STEM and non-STEM students (H2B).Furthermore, if the counter-stereotypical choice of a STEM field underlies motives that align more strongly with behavior [70], the relationship between intrinsic motives and reading frequency was expected to differ between the two groups in favor of STEM learners (H2C).

Do habits and interests predict academic performance?
In the extant literature, reading habits are related to reading comprehension [7,8,71], and thus reflect reading skills [72] in both STEM and non-STEM instructional domains.If so, students' reading frequency should predict academic attainment (as measured by GPA; H3A).However, reading habits might play a greater role in the academic success of STEM learners than non-STEM learners (H3B).The reason behind this expectation resides in the materials to which STEM learners are mostly exposed.Scientific reading materials (e.g., textbooks, articles, etc.) tend to be written in a technical language whose content is concept-dense and abstract [7,8,19].Thus, reading comprehension practice was deemed to be more critical to the academic success of STEM learners.Furthermore, due to the low independent reading rates documented in the literature on college students [69], intrinsically motivated reading was expected to be a better predictor of academic attainment than extrinsically motivated reading [73; H3C].No predictions were made regarding group differences in this matter since the extant literature provided no clear guidance.

Participants
The participants were a convenience sample of 394 female undergraduate students of a Saudi Arabian University that complies with a US curriculum and a student-centered pedagogy.The sample consisted of 163 students enrolled in STEM programs (engineering, computer science, and architecture) and 231 students enrolled in non-STEM programs (business, law, and interior design; [74]).It included all educational levels: 150 freshmen, 62 sophomores, 106 juniors, and 76 seniors.The ages of participants ranged from 18 to 30.Students were Arabic-English bilingual speakers whose proficiency in the second language was attested by standardized English tests before enrollment (IELTS ≥ 6 or TOEFL iBT ≥ 83).

Procedure and materials
Students enrolled in different general education courses were sent an online invitation to complete a survey on reading habits and interests at the end of the third semester of entirely face-to-face courses.At this time, efforts to readjust to face-to-face instruction from online had receded into long-term memory, thereby ensuring that data would reflect the post-pandemic environment rather than acclimatization processes.After providing informed consent, and reporting some demographic information (e.g., student ID, age, major, and academic level), respondents answered 6 questions organized into two categories: reading attitude and reading habits [7,8] (see Tables 1 and 2).A, B, C, and F questions probed their attitudes toward the activity of reading, whereas D and E assessed their reading habits as measured by the frequency of use of different types of materials and the purpose of reading.Thus, reading habits referred to reading behaviors that had become a regular part of students' daily life [75].Because at the selected university English is the primary mode of communication, any distinction between reading in either language was avoided to quell social desirability biases in participants' responses.Thus, each question was written in both English and Arabic to underscore that the survey was about reading regardless of the language used to do it.Furthermore, it was made clear to participants that questions about reading included materials irrespective of the language in which they were presented.
Performance data (i.e., GPA) were collected from the Office of the Registrar at the end of the semester during which information about students' attitudes and habits was collected.As soon as responses to the survey were matched with performance data, identifying information was deleted and random number codes were used to uniquely identify each participant and her responses.After the surveys were completed, a focus group was created that included 4 STEM and 4 non-STEM learners from the same student populations.Each student was given the survey and asked to elaborate on each answer.Answers were organized into reading habits and attitudes by two independent raters.The inter-rater reliability was 98%.
The study was approved by the Deanship of Research.It was deemed to conform to the standards for educational research of the Office for Human Research Protections of the US Department of Health and Human Services.

Results
All results of inferential statistics were treated as significant at the 0.05 level.If multiple tests were carried out on the responses to a question, a Bonferroni correction was applied (i.e., 0.05 divided by the number of comparisons).

Attitudes towards reading: do the interests of STEM and non-STEM students differ?
Table 1 contains the choices made by respondents to questions involving attitudes.Because the option "other" was not utilized by respondents, it was not included in the table.
The answers to Question A allowed us to compute a global index of students' attitudes toward reading (range: 0-2).A chi-squared test conducted on the frequencies of each choice indicated that most students either had a modest preference for reading or did not like it.There were no marked differences between STEM and non-STEM students [χ 2  (2) = 4.84, ns].
Variety may refer to the mode of reading (Question B) and its content (Question C).Each of the answers to Question B illustrated preferences about how reading is carried out: as an independently executed activity ("I prefer to read by myself"), as other-executed activity ("I like listening to stories or being read to"), and/or as an act of communication ("I like talking about what I have read").A chi-squared test carried out on the choices made by participants for each option indicated that although most students perceived reading as an independent activity, there were no differences between STEM and non-STEM students [χs 2 (1) ≤ 3.65, ns].To capture the diverse nature of students' reading engagements, the answers to Question B were added together to create a composite score.A Mann-Whitney U test, which was conducted on the composite scores, yielded group differences.Namely, the preferred modes of reading of STEM students were more diverse than those of non-STEM students [U = 20,662, p = 0.039, with the mean rank equal to 189.55 for non-STEM students and 206.76 for STEM students].
The choices of content made for the list of Question C illustrated the extent to which participants were eclectic readers.The number of choices made by each participant was computed and submitted to a Mann-Whitney U test.No differences were found between the two groups of students [U = 19,340, ns, with the mean rank equal to 195.27 for non-STEM students and 200.65 for STEM students].
Question F measured the extent to which reading is perceived as an independent activity or one that is shaped by social influence.Although reading was mostly seen as an independent activity by both STEM and non-STEM students (i.e., "I choose myself") [χ 2 (1) < 1, ns], social influence in the selection of reading materials was acknowledged.To measure whether there were group differences in reported social influence, the number of times each participant selected the other five choices was computed (range: 0-5).A Mann-Whitney U test indicated that there were no differences in the amount of social influence on reading choices reported by STEM and non-STEM students [U = 19,721, ns, with the mean rank equal to 193.63 for non-STEM students and 202.99 for STEM students].
In sum, H1A was only partially supported when attitudes were measured.Namely, the only difference that was found was in the modes of reading engagement of STEM students, which were more diverse than those of non-STEM students.

Reading behavior: do the reading habits of STEM and non-STEM students differ?
Table 2 contains the choices made by respondents for questions involving behavior.Next to each option, the most frequent choice is listed.The option "other", which was not utilized by respondents, was not included.For simplicity, the answers to "almost every day" and "once in a week" were combined.
We computed an index of reading frequency by adding the numerical answers to all the items of Question D. Total scores could range from 0 to 33.We also computed a reading frequency score for each reason listed as an option to Question E. The score for "fun and relaxation" was treated as an index of internal motivation to read (range: 0-3).The scores for "college assignments" and "need for personal information" combined (range: 0-6) were treated as an index of external motivation to read (i.e., reading arising from environmental demands and often engaged to obtain particular desirable outcomes such as good grades).
Reading frequency scores were submitted to a Mann-Whitney U test, which indicated that there were no differences in the frequency of reading of STEM and non-STEM students [U = 17,339, ns, with the mean rank equal to 203.94 for non-STEM students and 188.37 for STEM students].H1B was supported.
The self-reported incidence of internal motivation did not differ either [U = 18,558, ns], with the mean rank equal to 198.66 for non-STEM students and 195.85 for STEM students].As predicted by H1D, there were no differences between student groups when the incidence of external motivation was reported [U = 19,968, ns, with the mean rank equal to 192.56 for non-STEM students and 204.50 for STEM students].
To examine the relative strength of internal and external motives, a Wilcoxon T-test was used.For both STEM and non-STEM students, internal motives (as measured by reading for "fun and relaxation") counted less than external motives (as measured by reading for "college assignments" or by "need for personal information").Specifically, STEM students reported that they were more likely to read for college assignments [T = 4959, p < 0.001, with ranks for increases totaling 82 and those for decreases totaling 28] or for personal information [T = 3561, p < 0.001, with ranks for increases totaling 68 and those for decreases totaling 31] than for fun and relaxation.Similarly, non-STEM students reported that they were more likely to read for college assignments [T = 10,263, p < 0.001, with ranks for increases totaling 113 and those for decreases totaling 59] or for personal information [T = 7039, p < 0.001, with ranks for increases totaling 93 and those for decreases totaling 49] than for fun and relaxation.In sum, for both groups, external circumstances motivated reading more than the pleasure arising from it.H1C was not supported.

Do interests predict habits?
Spearman's Rho correlations were conducted to examine the relationship between students' interests and selfreported frequency of reading (see Table 3).It was of interest to determine whether our index of students' attitudes toward reading (responses to Question A) predicted the reading frequency scores computed from the answers to Question D. The correlation between the two variables was positive for both non-STEM [r s = + 0.38, n = 231, p < 0.001] and STEM learners [r s = + 0.46, n = 163, p < 0.001].H2A was supported.However, the coefficients of determination, which indicated the percentage of variance in attitudes that explained reading frequency, pointed to differences.Overall favorable attitudes toward reading accounted for its frequency much more for STEM learners than non-STEM learners.
Intrinsically motivated reading was indexed by the participants' selection of reading for "fun and relaxation" in response to Question E. Their reading frequency was predicted by intrinsic motivation in both STEM [r = + 0.65, n = 163, p < 0.001] and non-STEM students [r = + 0.48, n = 231, p < 0.001].Extrinsically motivated reading was indexed by the participants' selection of reading for "college assignments" and "need for personal information" in response to Question E. The combined scores of these two options of question E also predicted reading frequency in STEM [r = + 0.39, n = 163, p < 0.001] and non-STEM students [r = + 0.19, n = 231, p = 0.003].However, the coefficients of the determination illustrated two key issues.First, internal motivation predicted much more students' frequency of reading than external motivation (as predicted by H2B).Second, both types of motivation predicted reading frequency much more for STEM students than non-STEM students.Thus, H2C was supported, but more broadly, including not only intrinsic motives but also extrinsic ones.

Do reading habits and interests predict academic performance?
Students' responses to questions about reading habits (Questions D and E) and attitudes (Question A) were submitted to a Spearman's Rho correlation analysis to determine whether they predicted GPA.Table 4 illustrates the values of the correlations obtained and the corresponding coefficients of determination for STEM and non-STEM students.Habits (Question D) and attitudes (Question A) predicted GPA, supporting H3A.However, they did so to a greater extent for STEM than non-STEM learners if the coefficients of determination are considered, thereby supporting H3B.Coefficients of determination also indicated that intrinsic motives predicted GPA to a greater extent for STEM than non-STEM learners.To a lesser degree, seeking personal information predicted GPA.This extrinsic motive was a better predictor for STEM than non-STEM learners.Reading for college assignments did not predict GPA at all for either group.Thus, intrinsic motives were overall better predictors of academic attainment than extrinsic motives.H3C was supported.Important to note is that GPA did not differ between groups [t(392) = 1.19, ns].The mean of STEM students' GPA was 3.1 (SD = 0.52), whereas the mean of non-STEM students' GPA was 3.39 (SD = 0.50).

Post-facto qualitative evidence
After the survey data were collected, a focus group of 4 STEM and 4 non-STEM students from the same student population was interviewed to gain a deeper understanding of how particular reading habits were related to attitudes.Regardless of prodding, survey responses did not add much to the quantitative findings of the study.Three exceptions to this pattern, however, emerged.First, irrespective of the major, students reported reading mostly online materials.They reported reading often text on instant messaging services (e.g., WhatsApp) and social media posts.Second, students preferred watching videos or listening to audio files (e.g., podcasts, audiobooks, etc.) rather than reading text.There was a stark preference for the auditory modality as the main conveyer of course materials [76], which may be traced to the popularity of oral communication in the Middle East [77].Third, STEM students reported difficulties in reading technical language (e.g., unknown words) more frequently than non-STEM students.

Discussion
The results of the present study can be summarized into two key points.First, there were no marked differences in the reading habits and attitudes of STEM and non-STEM students.However, overall positive attitudes towards reading, as well as intrinsic and extrinsic motives predicted the reading habits of STEM learners more than those of non-STEM learners.These findings suggest that the effectiveness of interventions for at-risk students intended to enhance the frequency of reading to ameliorate reading comprehension skills rests, to a certain degree, on tackling learners' attitudes.Furthermore, our findings suggest that interventions targeting attitudes may be more successful in STEM than non-STEM learners.Second, reading habits and attitudes predicted academic attainment (as measured by GPA) differently for the two groups of learners.Namely, reading habits and attitudes towards reading played a greater role in the academic success of STEM learners than non-STEM learners.The reason may reside in the materials to which STEM learners are mostly exposed.That is, scientific reading materials (e.g., textbooks, articles, etc.) tend to be written in a technical language whose content is concept-dense and abstract [7,8,19].Overall, our results are consistent with those of the extant literature showing a positive link between educational attainment and reading habits and interests [32,[78][79][80].That is, students who spend time reading academic or nonacademic materials tend to have higher grades, perhaps because they become more skillful readers.Our results are also consistent with the findings of Whitten et al. [81], and those of Sappington et al. [82] who reported that leisure reading (e.g., non-academic readings) is related to students' academic attainment.Our findings add to the extant literature by underscoring the importance of reading to students' academic success, especially for STEM learners.
It is important to note that reading habits may be treated literally as the amount of practice with reading that a student puts forth.They can be also viewed as merely a marker for a host of factors, such as effort expended, motivation, etc.One of the most relevant implications of our study is that reading habits predict overall performance much more for STEM than non-STEM students.Thus, reading comprehension may be the underlying factor that produces this group difference.To this end, it is noteworthy to point out that reading a text requires not only the recognition of words but also the understanding of contents.Reading fluency (i.e., one's ability to read quickly and accurately) is different from reading comprehension (i.e., one's ability to extract meaning from a text), albeit both depend on practice.Most importantly, comprehension is not a passive process because readers, at the very minimum, must actively extract information from a given text and integrate it with existing knowledge to develop coherent ideas [83].Readers may also create knowledge by analyzing the text through their critical thinking skills [84].
Another contribution of the findings of our study is to fill the gap in the extant literature by examining the relationship between reading habits or attitudes and academic performance in an understudied student population.The greater link between reading practices and the performance of STEM students may underscore the greater demand for reading comprehension skills in readers who are dealing with materials that are difficult to process.However, other factors may be at work too, such as self-efficacy (i.e., confidence in one's abilities) [85].Another relevant factor may be goal orientation [86,87].Students are learning-oriented if they see academic courses as opportunities to acquire information that is personally meaningful and professionally useful.Instead, grade-oriented students see the pursuit of grades as their ultimate object.As such, they tailor knowledge and skill acquisition activities to the instructors' assessment methods.Concerning the activity of reading, grade-oriented (also labeled performance-oriented) students are motivated by extrinsic forms of reward or punishment (grades) and thus view reading in terms of its implications for course grades.Instead, learning-oriented (also labeled mastery-oriented) students find intrinsic value and satisfaction in the activity of reading.Pilotti et al. [88] found that both orientations coexist in the student population from which the current sample was taken.In that study, a preference for either orientation did not differentiate students' academic performance as measured by grades in assignments and tests.A re-analysis of the data indicated that goal orientation also did not differentiate STEM and non-STEM students.The studies of He, Meece and Miller, and Bernack et al. [89][90][91] found that a learning orientation is linked to optimal reading comprehension.In the current study, we did not directly assess whether learning-oriented students read more or better than grade-oriented students.However, even if such differences existed, they might not overlap with the STEM versus non-STEM distinction.
The present research has limitations that will be addressed in future research.First, the context in which the research occurred defined the limits of the external validity of its findings.To be noted is that data were collected in the postpandemic environment after a period of readjustment to face-to-face instruction had occurred and the implementation of gender equity standards imposed from the top was in progress.Data taken at this particular point in time may be different over time, especially for young women pursuing a college degree in Saudi Arabia.Indeed, gender-equity standards may require time to be adopted into the everyday lives of the recipients.Most importantly, a change of this magnitude is not merely assimilated but rather integrated into the existing tribal and religious customs and traditions of the Gulf, thereby reflecting a lengthy process of accommodation.Second, our findings involve self-reports and are thus liable to be affected by a social desirability bias.Although the bias might have led respondents to report a more favorable view of themselves as readers, there is no reason to believe that it would have affected STEM and non-STEM students differentially.Third, participants were bilingual students for whom English is their second language.Thus, reading involved both English and Arabic texts.We did not separate the two languages to ensure that participants reported their overall reading habits as reliably as possible, thereby minimizing social desirability biases linked to the English language being the primary and preferred mode of communication at the selected university.Most importantly, we did not want participants to underreport their attempts to read materials in Arabic when the corresponding materials in English are unclear.

Table 1
Attitude set:

Table 2
3 = "Almost every day"; 2 = "Once in a week"; 1 = "Once in a month"; 0 = "Never/hardly ever" a Questions for which multiple answers are allowed