Gender, Equity, and Science Writing: Examining Differences in Undergraduate Life Science Majors’ Attitudes toward Writing Lab Reports

Abstract

It is often causally generalized that females naturally excel more at writing than males. Contrastingly, modern conventional wisdom similarly purports that males often innately excel more at science. True or not, both generalizations overlook important gender differences at the intersection of writing in science. This quantitative study investigates undergraduate life science majors’ attitudes to and perceptions of writing lab reports for 294 students who self-identify as either female or male. We used exploratory and confirmatory factor analyses to develop a three-factor scale and provide reliability and validity on several related constructs: confidence, sense of belonging, and persistence as a life science major. Our results indicate males self-report as being significantly more confident at writing lab reports (F (1, 292) = 186.08, p < 0.05) than females. With regard to writing lab reports, no significant differences were found between genders in the constructs of belongingness (F (1, 292) = 1.64, p = 0.20) and persistence as a life science major (F (1, 292) = 0.66, p = 0.42). Understanding attitudes and perceptions toward writing lab reports through an equity lens provides information to help science majors be successful. Our findings add to the literature on gender, equity, and science writing, motivating further exploration into underlying malleable cognitive mechanisms.

1. Introduction

As one of the hallmarks of the scientific enterprise, science is often only considered successful if its methods and results are successfully communicated to the larger community. As one of the basic communication techniques available to scientists, writing is a critical component of ultimately becoming successful in the biological sciences [1] and fully engaging in the process of science. Developing scientific texts is essential to science understanding [2], and it provides a way for students to “think like” and “write like” disciplinary experts [3]. While the components of scientific writing certainly vary by genre and intended audience, the most basic of scientific writing typically includes stating a clearly stated problem through a critical review of the literature (introduction), specifying research methods, objectively relating results, justifying the creation of evidence through systematic analysis of collected data, and providing a thoughtful discussion that explains the relevance and implications of a study’s results. In this way, practicing professional scientists are able to communicate through peer-reviewed journals, agency reports, grant proposal writing, news outlets, and even specialized websites [2]. Taken together, effective science writing is an important skill for future scientists to learn.

Academic writing is often a high-stakes activity, and it establishes the legitimacy of knowledge [4,5] to scientific peers; however, authors examining the subject emphasize how science writing is particularly difficult to master [6]. This is often because of unique cultures, vocabularies, and practices within various STEM fields which impact how students approach challenges throughout their schooling and once (or if) they enter STEM careers [7]. While writing itself is widely known to be a valuable part of the learning process [8,9], writers must have a broad knowledge of the norms and conventions of the specific environment in which they attempt to write [8,10]. As a result, scientific writing is often an arduous task for students, especially when presented with new and complex content [1] that might not yet be fully mastered intellectually.

It is often widely acknowledged that university professors often feel disappointed about their undergraduates’ scientific writing skills and that it is in critical need of amelioration [4,5,11,12]. Nowhere is this more evident in science education than when considering students’ performance on creating a sufficiently detailed lab report. As one of the major writing tasks in almost all science subjects [13], the conventional lab report often contains many of the core elements of a research article, where students transform scientific knowledge and claims into a written narrative, an act which is intended to mimic the frequent task of professional researchers [6]. Yet since science writing is often taught secondarily to science concepts, students often thoughtlessly assemble their assigned lab report the day before it is due [1].

In recent years, the concept of gender has become one of the most extensively studied demographic variables in relation to writing [14], and it is often said in broad generalizations that females naturally excel more than males at writing. Higher writing achievement by females has been widely recognized in many large-scale studies which report females tend to score higher on assessments with written expression [15,16,17], and females tend to have more confidence in their writing abilities from an early age [18]. Along the same lines, the trend of females excelling over males at writing only widens as students progress in their schooling [16].

Gender is also one of the most widely studied demographic variables with regard to STEM education [19,20,21,22]. Even though women are awarded the majority of undergraduate and graduate degrees in the biological sciences in particular [19,20,21,23], many people assume men excel more in the sciences. Despite the report that females dominate the biological sciences by enrollment, men have been identified as having greater exam performance and whole class and group participation and as being more knowledgeable about class content [19,20]. Likewise, longitudinal studies indicate male students have higher self-concept and inclination to pursue STEM careers in college than females [24]. This presents an apparent dichotomy in terms of understanding who is naturally good at science writing and which groups might benefit from more targeted instruction.

Combining gender differences in both writing and science gives rise to the question of if gender differences themselves can account for some of the observed weaknesses in students’ written scientific communication skills in undergraduate biology classes as exhibited by writing lab reports. Since the notions of gender in STEM and gender in writing seem to conflict, we embarked on a survey based on the study of undergraduates driven by an overarching research question: are there differences in attitudes and perceptions of writing lab reports for female and male undergraduate life science majors?

Historically, some studies anecdotally reported that technical writing science assignments like lab reports are dreaded and disliked by undergraduates across a broad swath of introductory science classes [1,25,26]. At the same time, more recent quantitative studies researching students’ affective states towards technical science writing are lacking in the literature. Learning to write academically is influenced by how students perceive the act of writing [5], and positive affective states have been linked with desirable educational outcomes such as achievement [27]. As it turns out, perceptions might largely depend on individual’s judgements of how well they can accomplish a task [28], their confidence [29], and their sense of belonging [30]. Biology majors in particular might also be affected by how writing a lab report influences their attitudes about persisting as a life science major. Researchers have called for more studies investigating students’ affective reactions to writing [31,32], particularly for differences in attitudes and perceptions among females and males relating to lab reports. Such studies are important because studying attitudes and perceptions towards lab reports can help retain both female and male undergraduate science majors. This study is framed theoretically by the effect writing lab reports may have on confidence, sense of belonging, and persistence as a life science major.

1.1. Theoretical Framework

1.1.1. Confidence

Self-confidence is widely considered to be an essential component of self-efficacy [29,33,34,35] which applies to individuals’ attitudes and perceptions about their own writing [34,35,36]. Defined as one’s beliefs in one’s own capabilities to organize and execute actions required to produce a given attainment, self-efficacy is an essential component of Bandura’s social-cognitive theory [35]. A person’s sense of self-efficacy influences their choices, the effort they exert toward a specific task, and their determination in the face of adversity [37,38]. Learning to write is an emergent process, and each new writing task represents a different context where self-efficacy and confidence must be newly foraged [37].

Self-efficacy and confidence are situation- and task-specific [27]. Researchers apply the ideas that self-efficacy and confidence in students’ science writing impact one’s ability to perform certain writing tasks [36]. Students who lack confidence in their writing skills tend to be less inclined to engage in activities where those skills are required and often are more likely to give up in the face of adversity [33].

1.1.2. Sense of Belonging

Like hunger and personal safety, belongingness is ranked as a basic need composing Maslow’s hierarchy of needs [39]. A sense of belonging is a distinct attribute closely related to engagement [40] and can be defined as the extent to which a person believes they are a legitimate and accepted member of a community where their contributions are valued [30]. Since Maslow’s writings, many studies have emphasized the need for a sense of belonging within different contexts [30,40]. Context-dependent belongingness in academics can play a role in predicting students’ level of engagement [40,41] and it has recently been recognized as being critical to academic achievement for students at the classroom, subject, and intuitional levels [30]. Since undergraduate academic experiences are situated within academic majors at the collegiate level, examining the sense of belonging within the context of college classrooms has been gaining attention [40].

A sense of belonging also seems to play a key role in persistence within STEM majors [40,41]. As students progress through a program of study and the academic rigor increases, they face new challenges, and feelings of isolation and uncertainty can sometimes emerge [30]. While a high-achieving student may report low competence despite getting high grades and vice versa, students’ perception of their own academic performance may also impact belonging [41]. Additionally, robust social and academic connections seem to be essential for achievement and persistence for many students, and examining students’ sense of belonging can shed light on how well students are integrated both socially and academically into college life [40,41].

A sense of belonging with regard to various tasks in biology classrooms may be linked with retention [19]. To foster a sense of belonging in the classroom and feel good enough for their desired field, students in biology classrooms must be comfortable offering their ideas through routine communications [19] like writing lab reports. Previous studies have highlighted the impact of gender and a sense of belonging as they relate to various aspects of communication in science classrooms [19,21]. Researchers have identified belonging and gender in STEM classrooms as a gap in the literature worthy of being explored in more depth [41].

1.1.3. Persistence as a Life Science Major

Duckworth [42] described persistence towards long-term goals as a component of academic and professional achievement. Persistence is often paired with passion and working diligently towards goals in the face of adversity. Persistence is a personality trait, and according to Duckworth [42], does not relate to IQ. Authors have argued that students in higher education need high degrees of perseverance to achieve their academic goals [43]. The personality trait of persistence in specific STEM fields can help students transition into the university setting and matriculate through their majors [44].

The persistence of college students as STEM majors often depends on factors such as a supportive learning environment, quality teaching, intuitional conditions, and oftentimes having fellow students who look like them [45,46]. There is a stark difference in STEM degree persistence for undergraduates who are initially interested in STEM careers at the beginning of their undergraduate college career and those who decide to join STEM majors later [47]. For some college students, perseverance towards certain learning experiences is often in flux [48]. Gender disparities associated with college retention and persistence among STEM majors are multi-faceted [20]. Further research regarding underlying factors, such as gender, that might affect persistence and attainment of STEM degrees is warranted [20,47].

Persistence towards degree completion has been identified as paramount to retaining STEM majors [46]. While about one third of undergraduates switch majors [45], the choice of college major is a key stage for undergraduates. According to Sithole et al. [49], the majority of students who enroll in STEM majors do not graduate with a STEM degree. So too, there is a culturally established view that sciences are extra hard subjects for only a select group of students with notable intelligence [45].

2. Materials and Methods

The aims of this study are to investigate measurement validity and reliability of our developed scale and to use the scale to look for gender differences. The Perceptions and Attitudes Toward Writing Lab Reports Scale is intended to examine undergraduate life science majors’ attitudes toward writing lab reports. We use the three-factor instrument as a thematic aid to look for differences in perceptions towards writing lab reports by gender.

2.1. Survey Instrument

To provide an understanding of differences in females’ and males’ attitudes towards writing lab reports, we developed a scale in Qualtrics software with 14 items (Appendix A). All 14 items used a Likert Scale, beginning with strongly disagree and progressing to strongly agree. Participants also answered demographic questions where they self-identified their gender as female, male, or nonbinary/prefer not to say.

2.1.1. Survey Development

Before tackling a research objective to quantify a particular problem, researchers must often identify an instrument, and if they cannot find a suitable instrument, they create their own [50]. Since quantitative studies researching students’ affective states towards writing lab reports are lacking in the literature, we developed a survey that was originally intended to measure attitudes and perceptions relating to writing lab reports, as they relate to confidence and belongingness among undergraduate life science majors. We aimed to develop a short instrument with items that were unambiguous and easily understandable [50]. With some items containing statements of the constructs we wanted to measure, we developed items that would express relevant ideas [50].

Relevant ideas for items were inspired by several surveys which already exist in the literature. Emphasizing the importance of academic writing within highly specialized disciplines as one of the most important and complex activities in higher education, this survey was inspired by work published by Meza and Gonźolas’ [51] to measure self-efficacy for writing certain academic texts. Our instrument development was also influenced by Wortman-Wunder and Wefes’ [52] survey that was designed to measure confidence for specific science writing genera. Belongingness scales used to measure a sense of belonging within specific environments and within higher education also informed the survey [53,54].

2.1.2. Survey Implementation

We obtained approval from the IRB and after piloting the survey with 36 undergraduates in May of 2023; we launched the survey with eight instructors in the summer and fall of 2023. We distributed the survey to five universities and community colleges in the Mountain West and western United States. We selected the majority of the institutions based on the proximity to the research institution overseeing this study.

2.2. Sampling

We partnered with instructors who teach life science majors to distribute the survey to introductory classes specifically designated for life science majors. This study selectively recruited participants who identified as life science majors. Out of the 532 participants who were asked to complete the survey, 327 students responded, for a response rate of 61%. After 26 participants who were not life science majors or who did not complete the survey were excluded from the study, the sample size of the survey used for creating and validating the model was 301 undergraduates who completed the survey from a variety of life science disciplines. The sample included 199 self-identifying females, 95 self-identifying males, and 7 students who identified as nonbinary/prefer not to say. We present the demographic information of the life science majors in this study in

Table 1. Frequency and percent of participants’ self-identified gender, class standing, type of life science major, and type of school.

Demographic Variable	N	Percent
Gender
Female	199	66.11
Male	95	31.56
Nonbinary/Prefer not to say	7	2.32
Class Standing
First Year	36	11.96
Second Year	116	38.53
Third Year	58	19.27
Fourth Year	91	30.32
Type of Life Science Major
Agriculture/Rangeland Ecology	3	1.99
Biology	61	20.27
Biochemistry	3	1.99
Botany	4	1.33
Exercise Science/Kinesiology	43	14.29
Health Sciences/Pharmacy/Pre-Med	72	23.92
Microbiology/Molecular biology	45	14.95
Nutrition/Dietetics	7	2.32
Physiology	44	14.61
Ecology	9	2.99
Neurobiology	6	1.99
Pre-Vet/Animal Science	23	7.64
Zoology	3	1.99
Type of School
University	274	91.03
Community College	27	8.97

Type of life science major exceeds sample size because of double majors.

and descriptives in

Table 2. Descriptive statistics, skew, and kurtosis by item number on the Perceptions and Attitudes Toward Writing Lab Reports Scale.

Item	n	$\overline{\mathbf{x}}$	SD	Median	Min	Max	Skew	Kurtosis	SE
Q3.2	301	2.75	0.96	3	1	5	0.18	−0.61	0.06
Q3.3	301	2.92	1.14	3	1	5	−0.13	−1.16	0.07
Q3.4	301	3.74	0.97	4	1	5	−0.79	0.22	0.06
Q3.5	301	3.82	0.75	4	1	5	−0.61	0.65	0.04
Q3.6	301	3.48	1.00	4	1	5	−0.45	−0.30	0.06
Q3.7	301	2.93	1.01	3	1	5	0.07	−0.69	0.06
Q3.8	301	3.56	1.02	4	1	5	−0.45	−0.45	0.06
Q3.9	301	3.54	1.03	4	1	5	−0.41	−0.52	0.06
Q3.10	301	3.33	0.99	4	1	5	−0.42	−0.61	0.06
Q3.11	301	2.78	0.98	3	1	5	0.16	−0.29	0.06
Q3.12	301	2.55	1.10	2	1	5	0.50	−0.64	0.06
Q3.13	301	3.35	0.83	4	1	5	−0.58	−0.42	0.05
Q3.14	301	3.19	0.92	3	1	5	−0.37	−0.57	0.05

. Since the focus of this study was gender differences, we only included participants who self-identified as female or male for difference testing. We also did not have enough statistical power to include the seven students who self-identified as nonbinary/prefer not to say in the difference testing. Therefore, the one-way ANOVA sample size was 294 participants. We do not report ethnicity because most of the surveys were conducted in rural states with small populations, and we wanted to protect the anonymity of the sample.

2.3. Analytic Approach

In order to assess social constructs, researchers often develop measurement tools to measure phenomena they believe exist because of their theoretical understanding [50]. The objectives guiding this study were to both assess the validity of the survey and investigate demographic differences regarding undergraduate life science majors’ attitudes towards writing lab reports. Before beginning data analysis, we used R statistical software Version 2023.09.0+463 to code positively worded survey items and reverse code negatively worded items. Exploratory factor analysis (EFA) can be used when there is limited knowledge available from past research in underdeveloped substantive areas regarding to the studied phenomena [55]. Since the survey was recently developed and we did not find studies regarding students’ perceptions and attitudes towards writing lab reports in the literature, we used inductive EFA as our first analytic activity.

Since EFA can be used to identify plausible underlying constructs for a set of related items with a larger sample size [55,56,57,58], we used SPSS software Version: 29.0.0.0 [55,57] to conduct an EFA to determine the number of factors that would emerge from the analysis and develop the internal structure of our model [56]. Since researchers must make thoughtful decisions when conducting EFAs, we omitted Q3.1 (Appendix B) before conducting the EFA because items should be selected for their utility as indicators of anticipated factors and not represent variables from unrelated domains [58]. To examine the suitability of the data (for Q3.2–Q3.14) and if the data were factorable, we examined the determinant of the correlation matrix, the Kaiser–Meyer–Olkin (KMO) coefficient, and Bartlett’s Test of Sphericity [55,56].

To consider if a generated hypothesis using EFA is trustworthy regarding the structure under consideration, researchers often conduct a confirmatory factor analysis (CFA) after an EFA [55]. After the three-factor solution emerged from EFA, we conducted a CFA and determined fit indices with the lavaan package [59]. We examined the comparative fit index (CFI), the Tucker–Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardized root mean squared residual (SRMR) fit indices.

One-way analysis of variance (ANOVA) tests can be conducted when there are independent comparison groups and there is one independent and one dependent variable [56]. A one-way analysis of variance (ANOVA) was conducted to test the convergent validity of the latent factors and look for differences between gender in the factors. To ensure we could conduct one-way ANOVAs, we checked the assumptions: independent samples, the homogeneity of variance, and the normal distribution of the dependent variable [56]. Since there were only two means (females and males), post hoc comparisons were not conducted.

3. Results

We conducted three different analytical activities. We first used EFA to test the structure and construct validity of the scale. Next, we used confirmatory factor analysis to test the theoretical structure for construct validity and measures of internal consistency reliability. For the third analytical activity, we conducted an analysis of variance (ANOVA) to test for differences between females and males.

3.1. EFA

We used exploratory factor analysis to test the theoretical structure of the survey (n = 301). During our analysis, we tested several assumptions. We checked the determinant of the correlation matrix, the Kaiser–Mayer–Olkin (KMO) coefficient, and Bartlett’s test of sphericity to determine if we could conduct EFAs with the data [55,56,57]. The determinant of the correlation matrix is an indicator of multicollinearity and was 0.009, which was substantially greater than 0.00001 [57]. The KMO coefficient of sampling adequacy should be greater than 0.7 [55,57], and ours was 0.815. Significance testing with Bartlett’s test of sphericity indicates whether the factor analysis makes sense for a given matrix of variable interrelationships [55]. Our Bartlett’s test of sphericity was significant (X²(78) = 1380.03, p < 0.001), indicating the existence of factors within the data [57]. After comparing our fit indices to those in the literature [55,56,57], we determined the three-factor model had an acceptable fit.

When deciding on the number of factors to retain, researchers must balance parsimony and comprehensiveness [58]. Eigenvalues represent the amount of information captured by a factor, and factors with eigenvalues less than one should not be retained [50,55]. A scree plot graphs eigenvalues associated with successive factors. Researchers should look at the elbow of the scree plot to determine the number of factors [50,55]. We used eigenvalues (

Table 3. EFA total variance explained with eigenvalues for each factor with total, percent of variance, cumulative percent, and total rotation sum of squared loadings.

Factor	Initial Eigenvalue			Rotation Sum of Squared Loadings a
	Total	% of Variance	Cumulative %	Total
1	4.521	34.776	34.776	3.524
2	1.746	13.431	48.207	1.908
3	1.277	9.826	58.033	2.457
4	0.989	7.610	65.643
5	0.728	5.602	71.245
6	0.678	5.218	76.463
7	0.645	4.958	81.421
8	0.614	4.724	86.144
9	0.581	4.470	90.614
10	0.391	3.005	93.619
11	0.301	2.312	95.931
12	0.281	2.162	98.093
13	0.248	1.907	100.000

Extraction method: principal axis factoring. ^a When factors are correlated, sums of squared loadings cannot be added to obtain a total variance.

) and the curve of the scree plot (Figure 1) to establish a three-factor model [50,55]. We used Varimax rotation with Kaiser normalization to interpret the factors and identify factors with higher factor loadings (

Table 4. EFA pattern matrix with factor loadings for each item in the Perceptions and Attitudes Toward Writing Lab Reports Scale. Factor one (f1) is “confidence”, factor 2 (f2) is “sense of belonging”, and factor three (f3) is persistence as a life science major.

Item	Factor Loading
	f1	f2	f3
Q3.14	0.763
Q3.13	0.698
Q3.3	0.667
Q3.5	0.613
Q3.4	0.605
Q3.2	0.586
Q3.12
Q3.10		0.778
Q3.11		0.702
Q3.6		0.533
Q3.8			0.848
Q3.9			0.744
Q3.7			0.421

Extraction method: principal axis factoring. Rotation method: Oblimin with kaiser normalization.

[57]). To determine the construct validity of the scale, we considered the relationships between items and factors. One item (Q3.12) did not load onto the model. All of the items in each factor had a factor loading greater than 0.4 (Table 4), which is often acceptable for high loadings [55,56,57,58].

To see if items conceptually fit together and can be named, researchers should inspect the content of items within each factor [57]. Factors are usually named by their most conspicuous utility of what they have in common [58]. After considering the items on each factor and theoretical framework of the study, we named factor 1 “confidence”, factor 2 “sense of belonging”, and factor 3 “persistence as a life science major”.

3.2. CFA

After using EFA to inductively create the three-factor model, we used CFA to test the factor structure. Confirmatory factor analysis is well suited for hypothesis testing regarding relationships between factors and their indictors [55]. According to DeVellis and Thorpe [50], a CFA can be conducted on the same set of items to confirm a particular pattern of relationships as predicted by theory. Used to test how well the model fits the data and whether the parameters are of significance, CFA can be used after a theory has been developed with EFA reflecting how many factors underlie a set of measures [55].

The sample size of the CFA model (Figure 2) was 301 undergraduate life science majors. To determine the fit of the model, we checked the CFI, TLI, RMESA, and SRMS. As the TLI and CFI approach one, the fit is better [55]. TLI and CFI values greater than 0.9 indicate a good fit. The smaller the RMSEA and SRMR values, the better the fit [55,56]. RMSEA values (which represent the misfit of the model for each degree of freedom) less than 0.05 and SRMR values less than 0.08 indicate a good fit [56]. The CFA yielded the following fit indices: CFI = 0.92, TLI = 0.90, RMSEA = 0.08, and SRMR = 0.06 (

Table 5. The chi-square (χ²), degrees of freedom (df), change in chi-square (Δχ²), RMSEA, CFI, and TLI values of each factor analysis for Perceptions and Attitudes Toward Writing Lab Reports Scale.

	Perceptions and Attitudes Toward Writing Lab Reports Scale
Model	χ2 (df)	Δχ2	RMSEA	SRMR	CFI	TLI
One-factor CFA	498.763	N/A	0.149	0.102	0.674	0.609
	(65)
Three-factor CFA	152.598	346.165	0.081	0.055	0.922	0.899
	(51)

). We compared the fit indices with criteria in the literature [55,56], and inferred that our indices have an acceptable level of fit.

Reliability can be used to indicate the extent to which items are internally consistent with one another [50,55]. Using 0.7 as the cut-off for reliability, we computed the measurement reliability with Cronbach’s alpha for each underlying construct [57]. The alpha for the confidence factor was 0.825, which indicates good reliability [57]. The alpha for the belongingness and persistence as a life science major was 0.743 and 0.710, respectively, which indicates an acceptable internal consistency [57].

3.3. Difference Testing

We used the theoretical structure provided by EFA and CFA to test for differences between females and males. In order to conduct ANOVA testing, we checked several assumptions. Our study design ensured that samples were independent; we also checked skew and kurtosis by gender for each item to determine that the data were normally distributed (

Table 6. Descriptive statistics, skewness, and kurtosis by gender and construct [57] for the Perceptions and Attitudes Toward Writing Lab Reports Scale.

							Skewness		Kurtosis
Gender	Construct	N	Min	Max	Mean	Std. Dev.	Statistic	Std. Error	Statistic	Std. Error
Female	Confidence	199	6	30	19.25	4.211	−0.261	0.172	−0.217	0.343
	Belonging	199	3	15	9.51	2.313	−0.107	0.172	−0.294	0.343
	Persistence as a Life Science Major	199	3	15	9.98	2.497	−0.393	0.172	−0.143	0.343
	Valid N (listwise)	199
Male	Confidence	95	11	30	20.95	3.671	−0.308	0.247	0.192	0.490
	Belonging	95	4	15	9.88	2.458	−0.255	0.247	−0.050	0.490
	Persistence as a Life Science Major	95	5	15	10.23	2.327	−0.184	0.247	−0.658	0.490
	Valid N (listwise)	95
Neither/Prefer not to Say	Confidence	7	14	23	18.57	3.155	0.106	0.794	−0.672	1.587
	Belonging	7	3	13	7.86	3.934	−0.081	0.794	−1.290	1.587
	Persistence as a Life Science Maor	7	7	12	9.14	2.193	0.544	0.794	−1.781	1.587
	Valid N (listwise)	7

). To test for homogeneity of variance, we used Levene’s test, which was not significant for each construct (

Table 7. Levene’s test of homogeneity of variance by gender and construct [57] for the Perceptions and Attitudes Toward Writing Lab Reports Scale.

Item	Levene’s Test of Significance
Confidence	0.158
Sense of Belonging	0.118
Persistence as a Life Science Major	0.883

).

We used an ANOVA to test for significant differences (

Table 8. Fixed-effects ANOVA results for confidence, sense of belonging, and persistence as a life science major for the Perceptions and Attitudes Toward Writing Lab Reports Scale.

Predictor	Sum of Squares	df	Mean Square	F	p	partial η2	partial η2 90% CI [LL, UL]
Confidence
(Intercept)	73,713.07	1	73,713.07	4505.17	0.00
Gender	186.08	1	186.08	11.37	0.001	0.04	[0.01, 0.08]
Error	4777.67	292	16.36
Sense of Belonging
(Intercept)	17,988.26	1	17,988.26	3227.46	0.00
Gender	9.12	1	9.12	1.64	0.202	0.01	[0.00, 0.03]
Error	1627.47	292	5.57
Persistence as a Life Science Major
(Intercept)	19,840.05	1	19,840.05	3322.11	0.00
Gender	3.91	1	3.91	0.66	0.419	0.00	[0.00, 0.02]
Error	1743.86	292	5.97

) between females and males for each construct and to test the convergent validity of the latent factors. The convergent validity and the structure of the model were supported because the three-factor model was statistically different to the one-factor model (F (51, 65) = 346.17, p < 0.05). The sample size for the difference testing was 294 participants (females = 199, males = 95). With regard to confidence, results from the ANOVA show statistically significant differences (F (1, 292) = 186.08, p < 0.05) with small effect sizes (−0.43). Males (M = 20.95, SD = 3.67) reported having more confidence when writing a lab report compared to females (M = 19.25, SD = 4.21). No significant differences were found with writing lab reports between genders with regard to a sense of belonging (F (1, 292) = 1.64, p = 0.20) and persistence as a life science major (F (1, 292) = 0.66, p = 0.42).

4. Discussion

After utilizing EFA and CFA, all scales in the three-factor model show evidence of validity and reliability. The results of the one-way ANOVA on confidence with writing lab reports suggest that males have significantly more confidence writing lab reports than females (F (1, 292) = 186.08, p < 0.05). In contrast, our results from one-way ANOVAs on a sense of belonging (F (1, 292) = 1.64, p = 0.20) and persistence as a biology major (F (1, 292) = 0.66, p = 0.42) indicate there are no significant gender differences with regard to writing lab reports.

Factors are hypothetical constructs that cannot be directly measured [58], and researchers are not always certain how many latent variables will emerge when developing a scale [50]. Even though latent variables take on a specific value, they cannot be measured or directly observed [50]. While more items per factor are preferable, Watkins [58] maintains that at least three measured variables are needed for statical identification of a factor. While our instrument shows strong evidence of reliability and validity and had high factor loadings, two constructs emerge from the EFA (a sense of belonging and persistence as a life science major) with only three items. While three items loading onto a factor is acceptable, EFA tends to perform better when each factor is overdetermined [58].

With 199 females, 95 males, and 7 students who did not identify with either gender (Table 1), our findings support the idea that females outnumber males in the life sciences. Even with unequal group sizes, the ANOVA homogeneity of variances assumption was not violated ([57] Table 7). Due to our large sample size, the gender differences presented in this study may have been more discernable. While we surveyed students’ affective states toward writing lab reports, we did not link academic performance data with the three constructs identified in this study because we did not have access to information regarding students’ grades.

Our findings for undergraduate life science majors’ attitudes and perceptions regarding writing lab reports do not support the oft widespread reports that females tend to have more innate confidence in their writing abilities that widens as students progress through their schooling career [16,18]. Since males in our study exhibited significantly more confidence in writing lab reports compared to females, our results support the idea that there are still equity issues in undergraduate biology, especially when it relates to confidence with writing lab reports. Some of these equity issues may relate to gendered ontologies and epistemologies within scientific inquiry [22], the preposition of STEM as a purely objective discipline unaffected by its gendered history [22], and an incongruence between how some females might view themselves as being capable of becoming a competent biologist (imposter syndrome) [19].

Our findings support other researchers’ work that males tend to have more confidence about their own abilities in life sciences [19,20]. We can infer from this evidence that females could be especially impacted by a decreased confidence in writing lab reports because it could ultimately affect their self-efficacy in written science communication. Gender differences relating to confidence in biology communication (like lab reports) might also transfer to other environments that give students opportunities to practice scientific discourse [19]. Studies indicate males are more likely to be named as more knowledgeable about content in biology classes [21], and instructors need to develop equitable opportunities for students from different genders so they can practice the skills they need [19]. Teaching strategies that might help overcome the gender differences observed in our study might include instructors increasing encouragement and engagement of female students to boost confidence [19], especially as it pertains to their written communication.

While gender gaps have long been a concern in male-dominated STEM fields like physics and engineering, the predominance of female students enrolling in biology classes bolsters the assumption that gender differences do not exist and/or have been ameliorated in undergraduate life sciences [19]. With regard to confidence and science writing, our results suggest equity issues persist in life science. Gender differences regarding confidence writing lab reports are of particular importance because sociocognitive theory suggests writing beliefs and writing performance are related [38]. It seems when students have more confidence, they seem to have a higher corresponding higher writing self-efficacy [29,34,35]. Each new writing task likely represents a different context where self-efficacy must be newly foraged [10,37], and both females and males need to have confidence to succeed in science writing. To boost confidence for both genders, the results of this study indicate that educators implement techniques to improve student writing over several semesters to realize continued improvements [60] in writing lab reports. Such writing techniques include more scaffolding to build rhetorical skills [61,62], group writing projects, breaking the lab report up into manageable chunks [6,63], and providing more positive and specific feedback from the instructor [11].

While other researchers have inferred that increased male confidence in life science classes might inequitably impact their persistence as a major [21], we found no significant differences in our data between females and males in how writing lab reports affects their persistence as a life science major. Our results indicate that it simply may not be true that confidence alone affects persistence as a biology major, and this is an area that requires more research. While this study did not investigate social interactions, other researchers have suggested peer interactions may be a source of gender bias that may influence persistence as a life science major [64].

Socio-cultural norms and aspects of STEM culture and specific disciplines might affect female and male perceptions and attitudes towards writing lab reports differently. The long history of STEM as a male-dominated endeavor [22] and the current predominance of females in biology may have opposing effects. Likewise, the long-held belief that females excel at writing and men excel at science may also have opposing effects with regard to writing in science. Our data focusing on life science majors writing lab reports do not support other researchers’ findings that females report less sense of belonging in STEM fields [41]. This could be because there is general tendency for feelings of belongingness to follow patterns of representation [41], and females are in the majority in the biological sciences [19,20,21]. While a sense of belonging and persistence as a biology major may be influenced by the preponderance of women in the life sciences, an increased confidence in males writing lab reports may be a lingering effect of the long history of males dominating STEM.

Confidence, a sense of belonging, and persistence as a life science major are all important for both female and male undergraduate life science majors. Students must be comfortable and feel competent conducting a variety of academic activities [19] in both classroom and laboratory settings. Just as all females and all males are not the same, not all biology classrooms are the same [19]. Life science students might benefit from being given opportunities to interact with students who are like themselves and practice their skills in low-stakes environments [12,19].

With regard to how science writing affects females’ and males’ confidence, sense of belonging, and persistence as a biology major, our results indicate there are differences that still merit further exploration. Since scientific writing and STEM disciplines are very context-specific, researchers should consider the wide variety of activities and contexts students must be able to navigate to become successful in their chosen field. More research needs to be performed on students’ affective responses to the unique cultures and practices in STEM. Very few studies have been developed to identify belongingness in higher education [54]; this is especially true for students’ experiences within different contexts [53]. Since persistence as a STEM major is often cited in the literature as an area of concern, more studies targeting specific factors that might influence persistence seem worth pursuing if they do indeed impact students switching their majors. More research highlighting how certain aspects of STEM education affect the affective domain of females and males could be fruitful, and researchers might benefit from deciphering equity issues in STEM disciplines by unraveling rather than lumping together the unique and ever-changing cultures, habits, and conventions of science education. The roles of women and men in STEM disciplines and science writing are not fixed and seem to be continuously malleable and developing.