Second-language writing anxiety and its correlates: A challenge to sustainable education in a post-pandemic world

Abstract

One of the challenges of sustainable education in the post-pandemic world is students’ writing anxiety. The present study aimed to answer two key questions about writing in a second language for Arabic-English speakers enrolled in a written communication course after a return to on-campus instruction. First, it examined whether early exposure to English would predict diminished anxiety. Then, it assessed whether anxiety predicted specific writing deficits in the quality and quantity of students’ writing. Students completed a writing anxiety questionnaire, answered questions about their exposure to English and their attitudes toward English writing, and responded in writing to the query “Who am I?”. In this correlational study, overall anxiety increased with students’ later exposure to English and decreased with exposure to English-speaking media, and favorable attitudes toward English writing. Overall anxiety was linked to particular writing qualities, such as increased use of concrete words, and brevity of exposition. Yet, somatic anxiety, appraisal concerns, communication apprehension, and avoidance behaviors were differentially related to measures of English exposure and attitudes as well as to writing quality and quantity. For instance, late exposure to English writing was related to increased levels of appraisal concerns, somatic anxiety, and communication apprehension, but not to avoidance behaviors. Except for appraisal concerns, all other forms of anxiety were accompanied by decreases in word and sentence outputs and increases in the use of concrete words. Appraisal concerns were related to a decreased use of low-frequency (i.e. unfamiliar) words. Applications and implications for sustainable education in the post-pandemic world are examined.

Keywords:

• writing
• writing anxiety
• second language
• Middle East
• college students

1. Introduction

Writing is a demanding activity, especially for second-language speakers (Zailaini et al., 2015). Nevertheless, it is a set of skills that they have to master to a certain extent to do well in college. General education courses in written communication are intended to foster practice with such skills to ensure attainment in the courses of students’ chosen majors. Predictably, low performance in communication courses is often a harbinger of future academic difficulties (Gladys & Angeles, 2019; Lei & Lei, 2019; Millea et al., 2018).

A considerable amount of research has been devoted to the relationship between anxiety and cognitive performance. In such research, it has often been reported that anxiety impairs performance, particularly when the task being performed is complex and demanding (Derakshan & Eysenck, 2009; Naveh-Benjamin, 1991). According to the interference account, low performance reflects anxiety-producing difficulties in retrieving information needed for task completion (Derakshan & Eysenck, 2009; Naveh-Benjamin, 1991). Central to this assumption is working memory, which is the “sketchpad of conscious thought” where information from the environment and long-term memory is kept to be processed through limited attentional resources (Miller et al., 2018) (p. 463). In one’s working memory, the experience of anxiety involves the presence of various task-irrelevant thoughts (e.g., worries), which reduce the attentional resources available to be allocated to task-relevant activities (Sari et al., 2017; Stout et al., 2017) such as writing. Consequently, anxiety is expected to be negatively related to the quality and quantity of the writer’s output, especially when the task to be performed is complex and attention-demanding (as per writing in a second language) (Saeedi & Farnia, 2017).

For many students in countries around the world, their ability to write in a second language is a key competency (Bowles & Murphy, 2020). Indeed, writing is one of the four language competencies, in addition to listening, speaking, and reading, through which students express their thoughts and viewpoints, thereby representing one of the pillars upon which academic success rests (Grabe & Kaplan, 2014). As such, anxiety in second-language writing can be described as one of the foes of the sustainable development goal of ensuring a quality education (4.0) for all students. It is important to note that sustainable education is aimed at the development of a solid foundation of knowledge and competencies, both of which can be transferred to different task environments in response to the changing professional and labor demands of a country (Geitz et al., 2019). This type of education is competence-based and intended to develop critical thinking skills, broaden and integrate knowledge, and foster a desire as well as a capacity to apply that knowledge (e.g., through problem-based learning) (Chen, 2021; Dubey et al., 2017; Geitz et al., 2019). A subset of this aim is represented by education for sustainable development, which is meant to foster the acquisition and development of knowledge and competencies, as well as reflection on the consequences of actions and decisions in a forward-oriented, global, and responsible manner (Barth et al., 2007).

Within the broader framework of sustainable education, the present study offers a test of the interference assumption by examining the extent to which writing anxiety predicts the quantity and quality of the written output of second-language learners. Instead of examining particular errors, this study focuses on writing performance as measured by quantity measures (i.e., the number of words and sentences produced and the average number of words per sentence), and quality measures (i.e., poor-quality indices such as a preference for concrete and high-frequency words, and low sentence complexity) (Güvendir & Uzun, 2023). To control for test anxiety during writing, students are asked to respond in writing to a personally relevant prompt (i.e., “Who am I?”) unrelated to the course assessment protocol. Furthermore, the study examines the precursors of writing anxiety. Namely, it surveys the extent to which exposure to a second language, and the personal recognition of the relevance of the language for learners and their parents may predict writing anxiety. The target population is college students for whom Arabic is the first language and English is the language to be used in a compulsory written communication course in which they are enrolled. Considerable differences exist between Arabic and English in syntactic, morphological, phonological, lexical, semantic, rhetorical, and orthographical domains. These linguistic differences make learning English by Arabic speakers particularly challenging due to negative transfer, whereby knowledge of the Arabic language interferes with the acquisition of English (Thompson‐Panos et al., 1983).

The study takes place in the post-pandemic environment at a university that has returned to completely face-to-face courses from asynchronous online learning during the COVID-19 pandemic. Considerable evidence exists that increased anxiety has been experienced by college students during the pandemic (Bailey & Almusharraf, 2022; Khoshaim et al., 2020; Wang et al., 2021). Writing anxiety may have been aggravated in learners by the pitfalls of online instruction, such as technical difficulties and physical distance (Russell, 2020). Yet, little is known about the writing anxiety of Arabic-English speakers in the post-pandemic world, especially those who can be classified as modest or competent users according to the International English Language Testing System (i.e., IELTS) standards. Whenever earlier findings are available, they often do not apply to the post-pandemic period. Furthermore, participants’ foreign language competence level, as measured by standardized tests, such as the IELTS or the Test of English as a Foreign Language (TOEFL), is underreported along with other critical factors. For instance, Aloairdhi’s study (Aloairdhi, 2019) focuses on the pre-pandemic period. She found moderate levels of writing anxiety, which college students attributed to their fear of evaluation, time pressure, doubts of competence, and low self-confidence. Her study, however, was limited to the assessment of anxiety with no direct measure of writing performance or assessment of its relationship with anxiety levels. Al-Ahdal and Abduh (Al-Ahdal & Abduh, 2021) found overall elevated anxiety levels among college students purportedly during the pandemic. They reported that writing anxiety was negatively related to writing strategies (as measured by a questionnaire) and writing achievement. The latter, however, was vaguely defined as the precise and accurate written presentation of one’s ideas in a foreign language, without a clear operational definition that ensured replicability. Alfarwan (Alfarwan, 2022) reported that several participants from either a preparatory course or a first-year course of their English major experienced moderate levels of anxiety which were perceived as having a positive impact on their English writing. Unclear, however, is the timing of the data collection.

Based on the interference hypothesis, the following predictions are made:

H1:

Earlier exposure to English or English writing in particular (i.e., age of acquisition as a measure of one’s familiarity with a language) will be associated with lower students’ overall writing anxiety.

H2:

Greater students’ self-reported practice with the English language through media exposure and travel will be linked to lower overall writing anxiety.

H3:

Students’ recognition of the importance of English in their lives will be associated with higher overall anxiety. Similarly, parents’ recognition of the importance of English will be linked to higher overall anxiety.

H4:

Higher students’ overall anxiety will be accompanied by written outputs that will be brief and of poor quality. In the present study, brevity is measured by the number of words, number of sentences, and number of words per sentence in students’ written output. Poor quality is indexed by students’ more frequent use of concrete and high-frequency words, and lower sentence complexity.

In our study, we acknowledge that writing anxiety is not a unidimensional concept. Rather, the feelings of fear, dread, and concern, known as writing anxiety, comprise both psychological and physiological states. They include somatic anxiety (e.g., self-reported symptoms of increased autonomic arousal and nervousness, such as hyperventilation, sweating, trembling, palpitations, etc.), and cognitive manifestations that reflect thought processes (e.g., worry, intrusive thoughts, and lack of concentration), and avoidance behaviors (e.g., withdrawing, procrastinating, and evading) (Cheng, 2004). As writers with anxiety view writing as an unrewarding activity, avoidance behaviors exemplify their attempts to circumvent situations where writing is demanded (Daly & Miller, 1975). Cognitive manifestations can be broadly differentiated into appraisal concerns (i.e., fears of being negatively evaluated) and communication apprehension (i.e., fears of not being able to communicate effectively) (Pilotti et al., 2023). Guided by the interference hypothesis, the following predictions involving distinctive types of anxiety are made:

H5: Different types of writing anxiety will be related to quality and quantity measures of writing outcomes depending on the extent to which they interfere with word retrieval. Interference may have a selective impact (e.g., fostering the choice of concrete and high-frequency words, and reducing sentence complexity). Alternatively, it may have a global impact (e.g., reducing the number of words and sentences produced, and the number of words per sentence). Selective interference may be expected to be linked to appraisal concerns and communication apprehension, whereas global interference may be expected from somatic anxiety and avoidance behaviors.

H6: Precursors will be related to different types of writing anxiety depending on the extent to which they exemplify global rather than localized experiences. Namely, age of acquisition (i.e., a measure of one’s foreign-language familiarity) covers a variety of experiences of second-language learning. As such, it is likely to be associated with all types of anxiety. Instead, exposure to English media and travel to English-speaking countries may be mostly linked to reduced communication apprehension. Although such apprehension is known to be experienced while interacting with others, travel and media experiences usually entail leisure activities that are likely to make appraisal concerns, somatic anxiety, and avoidance behavior largely moot.

2. Literature review

In the extant literature, the relationship between anxiety and writing output has mostly focused on errors. Overall, the output of anxious writers compared to that of less anxious writers has been reported to be of worse quality, have more errors, and be briefer, underdeveloped, and syntactically more rudimentary. For instance, Türnük and Aydin (Türnük & Aydin, 2020) focused on interlanguage errors, which learners commit when writing in a second language that differs from their first language in terms of grammar and vocabulary. They found writing anxiety to be related to increased interlanguage errors. Sabti et al. (Sabti et al., 2019) reported that the higher the writing anxiety, the poorer the writing performance. In their study, writing performance was measured on five criteria, including, content, organization, grammar, vocabulary, and mechanics. Scores on the five criteria were combined to yield a measure of students’ overall English writing ability. Liu and Ni (Liu & Ni, 2015) also found differences among students based on their second language proficiency measured holistically through their performance on an argumentative essay writing task. More ambitiously, Saeedi and Farnia (Saeedi & Farnia, 2017) reported that writing anxiety is negatively associated with measures of students’ writing outputs. Such measures included complexity (i.e., the number of clauses relative to the number of utterances “consisting of an independent clause, or sub-clausal unit, together with any subordinate clause(s) associated with either”; Foster et al., 2000; p. 365), accuracy (i.e., the percentage of error-free clauses relative to the total number of clauses), and fluency (i.e., the number of syllables divided by the total number of seconds taken to answer a written prompt multiplied by 60).

Other studies have focused on the sources of writing anxiety. For instance, Rabadi and Rabadi (Rabadi & Rabadi, 2020) found that learners who experience tension in writing tend to attribute it to a variety of factors, including linguistic difficulties, inadequate practice, low self-confidence in their writing abilities, and fear of writing tests. Contrary to causal attributions collected from learners via qualitative methodologies, other research has focused on learners’ age of acquisition as an index of their familiarity with a second language. Although the variable age of acquisition has traditionally been seen as a predictor of second language performance in naturalistic contexts (Abrahamsson & Hyltenstam, 2008), in limited-exposure programs, research has indicated little advantage for an early start. In immersive language programs, mixed findings have been reported. Pfenninger (Pfenninger, 2020), for instance, found that age of acquisition was related to second language performance for written fluency and accuracy, whereas there was no relationship for written lexical richness, complexity, or mean length of utterance. Yet, one important aspect of students’ exposure to the English language is the extent of their overall practice with writing activities. Namely, more exposure to writing activities, tends to be linked to lower students’ self-reported writing anxiety (Rabadi & Rabadi, 2020; Schweiker-Marra & Marra, 2000).

By and large, for either the output or source measures, the relationship with writing anxiety is not always uniform in its intensity and direction. Comparisons across studies are also difficult given the variety of first and second languages, which may entail differential negative transfer, competency levels, and methodological choices for the assessment of English writing. Our study is born out of the acknowledgment that the assessment of writing anxiety’s relationship with potential precursors and output measures rests on specific populations of students (as defined by language type and competency). Specificity at the expense of generality enhances the value of predictions regarding writing anxiety for educators, counselors, and administrators who are on the ground working with such populations.

Even when a particular population is considered, findings are not uniform. Consider, for instance, evidence regarding female students’ second language writing anxiety. Min and Rahmat (Min & Rahmat, 2014) reported lower somatic and cognitive anxiety, as well as lower behavioral avoidance in female students. Similarly, Liu and Ni (Kirmizi & Kirmizi, 2015) found female students to exhibit overall lower writing anxiety, and less evaluation apprehension as well as less dislike for writing in a second language. Conversely, Kirmizi and Kirmizi (Liu & Ni, 2015) found no gender differences in cognitive anxiety and avoidance behavior, but greater somatic anxiety in female students. Quvanch and Na (Quvanch & Si Na, 2022), instead, reported no gender differences. They found cognitive anxiety as the dominant type among all students.

Of particular interest to us are female undergraduate students of a traditionally patriarchal society that is being re-engineered from the top to integrate women into the workforce. To this end, Saudi Arabia is chosen as the context of the study. Its Vision 2030 is a socio-economic plan to convert an economy based on fossil fuels into a knowledge- and service-based economy largely reliant on renewable energy (Allmnakrah & Evers, 2020). The entire plan rests on young people, including women and men, who possess the right skills and knowledge to contribute to the envisaged society (Pilotti, 2021). Tangible evidence of success, such as good academic performance, is at the core of the entire plan as it is expected of both men and women in all facets of life. Equal educational opportunities for women and men have suddenly opened to female students in career paths that were previously forbidden. At the same time, gender equity has placed a great deal of pressure on women to succeed (Pilotti & Elmoussa, 2022). Thus, female college students of Saudi Arabian descent are of particular interest to us as they are the main subjects of a field experiment choreographed from the top. Do they approach the challenges of writing in a second language with a threat framing, including negative emotions, such as anxiety (Pilotti et al., 2021; Vandewalle et al., 2019)? Alternatively, do they approach such challenges with an opportunity framing, including constructive thoughts about the value of additional effort when setbacks occur, which considerably minimizes anxiety?

3. Materials and methods

3.1. Participants

The participants were 409 female undergraduate students enrolled in a written communication course of the general education curriculum. To enroll in courses of the general education curriculum, students obtained at least a 6.0 IELTS overall score and a 5.5 IELTS writing score (range: 0–9). Thus, the selected sample included a variety of speakers from modest to competent users, which are often neglected in the extant literature in favor of introductory-level learners. The course was offered to freshmen by a Saudi Arabian University that complies with a US curriculum and advocates a student-centered pedagogy. The curriculum and instruction of the selected course were approved before utilization by the Texas International Education Consortium (TIEC). The students’ average age was 19.27 (range:18–30). Of the 409 participants, 49.14% chose a STEM major (i.e., engineering, computer science, or architecture), and 50.86% a non-STEM major (i.e., business, law, or interior design).

3.2. Materials and procedure

Participants performed two tasks after giving informed consent. First, they answered a series of demographic and attitudinal questions and were asked to write a brief paragraph in response to the question “Who am I?”. A loose 10-minute timeline was allocated to the latter task. Table 1 illustrates demographic and attitudinal questions related to students’ exposure to English. Then, students completed a revised version of the Second Language Writing Anxiety Inventory (SLWAIr) of Cheng (Cheng, 2004), which was utilized to assess students’ writing anxiety. The development of SLWAIr, which entailed the assessment of face and content validity, and test-retest reliability (DeVellis, 2017; Mellinger & Hanson, 2020; Taherdoost, 2016) was guided by the need to adapt the scale to the student population of the selected university. SLWAIr consisted of 22 items in which surface changes in the wording of some items ensured second-language speakers’ intuitive comprehension. For instance, the phrase “write English composition” was changed to “write assignments in English”. Participants were asked to report their answers on a 5-point scale from strongly agree (+2) to strongly disagree (−2) with 0 serving as the neutral point. Cronbach’s alpha was 0.85. Items were organized into four dimensions of anxiety based on prior Principal Component Analysis (PCA) (Pilotti et al., 2023): somatic anxiety, appraisal concerns, avoidance behaviors, and communication apprehension (see Appendix 1).

Table 1. Mean (M) and standard error of the mean (SEM) of self-reported exposure to English, anxiety levels, and measures of writing

Measures of Exposure to English	M	SEM
Age at which I was first taught English in school.	13.29	0.32
Age at which I began writing in English.	11.80	0.18
I believe that it is important to write well in English. (−2 – +2)	+1.70	0.03
My parents believe that it is important that I learn to write well in English. (−2 – +2)	+1.68	0.03
I regularly travel to English-speaking countries. (−2 – +2)	+0.89	0.06
I regularly watch TV, movies, and videos in English. (−2 – +2)	+1.56	0.04
Writing Anxiety Measures (−2 – + 2)	−0.38	0.03
Somatic Anxiety	−0.86	0.04
Appraisal Anxiety	−0.13	0.04
Avoidance	−0.43	0.04
Communication Anxiety	−0.19	0.03
Students’ Writing—Quantity Measures (0–100)
Word Count	42.82	1.26
Sentence Count	2.96	0.09
Word per Sentence	16.81	0.53
Students’ Writing—Quality Measures (0–100)
Word Unfamiliarity	48.10	0.94
Word Concreteness	55.40	0.83
Syntactic Complexity	41.08	1.06

The study was conducted in adherence with the ethical guidelines for educational research of the OHPR of the U.S. Department of Health and Human Services as well as those of the American Psychological Association. The research was approved by the Deanship of Research at the selected institution before execution.

3.3. Data analyses

The scoring of the paragraphs written by students in response to the “Who am I?” question was organized into a set measuring quantity and a set measuring quality through the use of the TextEvaluator® software developed by the Educational Testing Services (ETS). Quantity measures included the number of words used, the number of sentences written, and the average number of words per sentence. These measures were intended to test whether anxiety is related to diminished word output.

At the word level, quality measures included a word unfamiliarity index and a word concreteness index. Word unfamiliarity (1–100) was computed from two corpora, one of 400 million words developed by ETS, and one of 17 million words developed by Zeno et al (Zeno et al., 1995). Word concreteness (1–100) relied on the concreteness ratings developed by Coltheart (Coltheart, 1981). Higher scores reflected a higher proportion of concrete words and a lower proportion of abstract words. At the sentence level, an index of syntactic complexity was utilized (1–100), which incorporated the average sentence length, the average number of modifiers per noun phrase, the average number of dependent clauses per sentence, and a measure of sentence depth, which reflects the memory load imposed by sentences with varying syntactic structures (Yngve, 1960). Taken together, quality measures were intended to test whether anxiety is related to lower word unfamiliarity, greater concreteness, and reduced syntactic complexity.

Pearson correlations were computed to determine the extent to which measures of students’ exposure to English or attitudes would predict each anxiety type (somatic anxiety, appraisal concerns, avoidance behaviors, and communication apprehension). Pearson correlations were also computed to determine whether each type of anxiety would predict specific features of students’ writing products, including quality measures, and quantity measures.

4. Results

Table 1 displays the descriptive statistics for the following variables: students’ self-reported exposure to the English language, attitudes towards English writing, their anxiety levels organized by type, and quantitative and qualitative writing performance measures. Inferential statistics are arranged by the question they sought to answer.

4.1. Does exposure to English or its perceived relevance predict different forms of anxiety?

Table 2 shows the extent to which different forms of anxiety (or overall anxiety) were predicted by the amount and type of exposure to English or view of the relevance of English. Unsurprisingly, the older students were when they started learning English or, more selectively, writing in English, the higher their overall writing anxiety. However, exposure to English and English writing selectively predicted different types of anxiety. Namely, a student’s exposure to English predicted appraisal concerns and avoidance behaviors, whereas practice with English writing predicted all types of anxiety except avoidance behavior. Furthermore, travel to English-speaking countries was accompanied by a decline in somatic anxiety and appraisal concerns. Exposure to English-speaking media was also accompanied by a decline, but more widespread, of all forms of anxiety, except appraisal concerns.

Table 2. Pearson correlation coefficients and coefficients of determination (%) illustrating the extent to which writing anxiety is predicted by exposure to English and predicts word and text properties

Measures of Exposure to English	Writing Anxiety	S	Ap	Av	C
The age at which I was first taught English in school.	+0.197 4%	+0.045	+0.177 3%	+0.401 16%	−0.019
The age at which I began writing in English.	+0.213 5%	+0.258 7%	+0.139 2%	+0.045	+0.211 5%
I regularly travel to English-speaking countries.	−.093	−0.132 2%	−0.101 1%	+0.026	−0.069
I regularly watch TV, movies, and videos in English.	−0.192 4%	−0.193 4%	−.085	−0.217 5%	−0.118 1%
Attitude Measures
I believe that it is important to write well in English.	−0.232 5%	−0.189 3%	−0.131 2%	−0.237 6%	−0.181 3%
My parents believe that it is important that I learn to write well in English.	−0.108 1%	−0.046	−0.070	−0.105 1%	−0.120 1%
Students’ Writing—Quantity Measures
Word Count	−0.194 4%	−0.186 4%	−0.062	−0.195 4%	−0.175 3%
Sentence Count	−0.151 2%	−0.118 1%	−0.051	−0.176 3%	−0.143 2%
Word per Sentence	−0.031	−0.067	−0.023	−0.004	+0.001
Students’ Writing—Quality Measures
Word Unfamiliarity	−0.098	−0.065	−0.108 1%	−0.067	−0.063
Word Concreteness	+0.150 2%	+0.163 3%	+0.078	+0.102 1%	−0.126 2%
Syntactic Complexity	−0.098 1%	−0.124 2%	−0.031	−0.084	−0.072

Note: S = Somatic Anxiety; Ap = Appraisal Anxiety; Av = Avoidance; C = Communication Anxiety. Values in bold are significant at the .05 level.

Contrary to our predictions, the more students believed writing in English was important, the less anxiety they felt across all four types. Parents’ views had a more limited reach. Namely, the more their parents believed being able to write in English was important, the lower the frequency of avoidance behaviors and communication concerns students reported. Yet, it is noteworthy to consider that coefficients of determination (i.e., the percentage of variance in measures of exposure or attitudes that accounted for reported anxiety) were rather small. Thus, other factors might contribute to writing anxiety that exposure or attitude measures were unable to capture.

4.2. Do forms of anxiety predict deficiencies in writing quality and quantity?

Table 2 also shows the extent to which overall anxiety or different forms of anxiety might predict specific writing deficiencies. For qualitative deficiencies, overall anxiety predicted increased use of concrete words and decreased sentence complexity. However, when specific forms of anxiety were examined, all predicted concrete word use except for appraisal concerns. The latter predicted a decline in the use of low-frequency (unfamiliar) words, whereas somatic anxiety was linked to a decline in the complexity of sentences.

For quantitative deficiencies, overall anxiety predicted a decline in the number of words and sentences produced. When specific forms of anxiety were examined, all predicted a decline in the number of words and sentences produced except for appraisal concerns. Again, the coefficients of determination (i.e., the percentage of variance in self-reported anxiety that accounted for particular writing deficiencies) were rather small. Thus, other factors besides anxiety might contribute to deficiencies in second-language writing.

4.3. How do female students experience writing in a second Language?

If female students approached the challenges of writing in a second language predominantly with a threat framing, they would exhibit on average writing anxiety scores above 0 on the scale from −2 to + 2. Alternatively, if they approached such challenges predominantly with an opportunity framing, their writing anxiety scores would be on average near or below 0. Table 1 reports the average values for each type of writing anxiety. A single-sample t-test performed on each of the four types of anxiety indicated that writing anxiety among the participants was rather a null or minor occurrence [ts(408) ≥ 3.63, p < .001].

5. Discussion

The results of the present study can be summarized into four points. First, late exposure to English in general and specifically to English writing was related to increased levels of overall writing anxiety, thereby supporting H1. However, H6 was not uniformly supported. Late exposure to English in general and specifically to English writing was related to increased levels of appraisal concerns. Late exposure to English also predicted avoidance behaviors, whereas late exposure to English writing predicted somatic anxiety and communication apprehension (in addition to appraisal concerns). It is important to note here that the question about the acquisition of English was interpreted by students as referring to a complex and varied set of skills that cover real-time and offline communication, either oral or written, in both informal and formal settings. Enhanced speakers’ appraisal concerns and avoidance behaviors might take a prominent role when the skills to be acquired are complex and diverse with a wide range of applications. Instead, the question about the acquisition of written English referred to a subset of communication skills that are seen as highly challenging, requiring considerable practice to master, and that are often associated with unavoidable tasks (e.g., course assignments). As such, avoidance behaviors might play less of a role while other forms of anxiety take center stage (e.g., appraisal concerns, communication apprehension, and somatic anxiety).

Second, exposure to English-speaking mass media (TV, movies, and videos) predicted a decline in overall anxiety, thereby supporting H2. Again, H6 was partially supported. Exposure to English-speaking mass media was linked to a decrease in all forms of anxiety except appraisal concerns, whereas travel to English-speaking countries was linked to a decrease in somatic anxiety and appraisal concerns. Traveling to English-speaking countries generally entails leisure activities through which interactions with native speakers occur. As such, heightened physiological arousal and fears of being judged might be largely reduced due to the recreational atmosphere of the interactions. Instead, exposure to English-speaking media entails mostly passive reception of communication (e.g., watching TV, movies, and videos) that makes appraisal concerns largely irrelevant.

Third, personal attitudes toward writing in English mattered. Students interpreted the question regarding the importance of English (or English writing) as an index of their positive attitudes toward the language. Thus, H3 was not supported. Instead, their recognition of the importance of the English language was related to a reduction in all manifestations of anxiety rather than an increase. Parents’ recognition of the importance of the English language was related to reduced avoidance behaviors and communication apprehension. During debriefings, students valued English as the lingua franca for employment in all types of occupations in and outside Saudi Arabia. They reported that parents who were supportive of their learning English would make encounters with English through school activities, mass media, and travel more pleasant and broader in scope, thereby reducing their communication apprehension and avoidance behaviors.

Fourth, according to MacIntyre and Gardner (MacIntyre & Gardner, 1991), anxiety may affect different stages of information processing. The results of the present study are consistent with the notion that anxiety is linked to deficits in written output that can be traced to disruptions of working memory. For instance, the declines in the number of words and sentences produced as writing anxiety increases are consistent with Zabihi’s finding (Zabihi, 2018) that writing is dependent on the writer’s working memory and that anxiety disrupts the writer’s ability to retrieve linguistic information, such as the lexical items and syntactic frames to be used in working memory to build sentences.

Specifically, we found that the higher the overall anxiety experienced by students, the less they wrote and the more their writing relied on concrete words and simplified sentence structures. This evidence provides partial support for H4. H5 garnished mixed support. Except for appraisal concerns, all other forms of anxiety were linked to decreases in word and sentence outputs and increases in the use of concrete words. Appraisal concerns were related to decreased use of low-frequency (i.e., unfamiliar) words. Taken together, these relationship patterns suggest that most forms of anxiety may deplete the cognitive resources demanded by word search processes, leading to brevity and concreteness of exposition. Furthermore, fear of judgment may bias word search processes towards familiar lexical items.

Güvendir and Uzun (Güvendir & Uzun, 2023) found that writing anxiety predicts diminished sentence complexity. In our study, somatic anxiety, in particular, was related to diminished syntactic complexity, suggesting that the perception of symptoms of increased autonomic arousal and nervousness (e.g., hyperventilation, sweating, trembling, palpitations, etc.). This evidence may be accompanied by a reduction in the writer’s ability to keep information in working memory, leading to short sentences with no embedded clauses. The broader connection of anxiety with writing deficiencies uncovered by Zabihi (Zabihi, 2018) (including writing complexity, fluency, and accuracy) might be linked to the higher anxiety experienced by his sample of participants, as well as to the challenges of the writing task (11 minutes to write a story based on cartoon strips with no planning time). That is, the importance of the contextual frame in which a writing assignment is performed, including time available and topic, should not be discounted as capable of modulating the strength of the relationship between writing anxiety and writing deficiencies. Yet, the findings of Zabihi’s study and ours underscore that the availability of working memory resources during writing is key. It is important to note though that as for the predictability of exposure-to-English measures, the coefficients of determination, which illustrated the percentage of variance in self-reported anxiety that accounted for particular writing deficiencies, were rather small. Thus, although most hypotheses were supported, other factors could be considered as having a role in predicting writing anxiety and writing outputs.

5.1. Limitations

First, our study focused on preconditions for writing anxiety, such as the age of exposure to the English language, and relied on anxiety in its different forms to predict particular measures of writing output. The study did not focus on traditional measures of writing quality, such as misspellings or grammatical construction errors.

Second, it did not examine the specific disruptions of working memory brought about by anxiety. According to the control theory of attention (Miyake et al., 2000), working memory entails three key functions: inhibition (i.e., resisting disruption of interference from task-irrelevant materials), shifting (i.e., remaining focused on task-relevant materials of current importance), and updating (i.e., performing the transient storage of information through updating and monitoring of working memory contents). Performance in a conceptually complex task, such as writing, involves several different forms of processing. Such complexity makes it difficult to identify the particular process or processes most affected by each of the given types of anxiety.

Third, our study entailed learners who were classified by their performance on the IELTS as ranging from modest to competent users of English. Although this choice allowed us to focus on quality and quantity measures instead of the traditional error measures, it might have reduced the variability of the scores and their range, thereby shrinking the size of the correlation coefficients. A broader range of English writing competencies might have yielded more robust values. Of course, students’ first language (i.e., Arabic), their cultural background, which exhibits a mixture of collectivistic and individualistic motives (Hamann et al., 2022; Jiang et al., 2018), and particular personal experiences with English might also contribute to writing anxiety.

Fourth, our results referred to students’ writing in response to a personally relevant query (“Who am I?”) without pending course assessment. In this context, the levels of anxiety reported by the participants were rather minor. It follows that writing in response to less relevant instructor-assigned topics, and under stringent time constraints and testing conditions might have yielded higher reports of anxiety as well as stronger links between writing anxiety, its different forms, and output measures.

Fifth, writing in response to a personally relevant query (“Who am I?”) was a one-time event without much of an opportunity for careful planning and revising. Evidence exists that revising is a key determinant of writing quality (as measured by the organization, content, word choice, and language use of the writer) and a potential contributor to the reduction of anxiety (Bayat, 2014; Chung et al., 2021; Tsiriotakis et al., 2017). The self-reported levels of anxiety of our study might have been more likely to reflect participants’ natural approach to writing if revising was given a greater opportunity to be performed through additional time. The latter would have offered writers the chance to reflect on their work.

Sixth, the possibility exists that our participants underreported anxiety levels due to the context in which writing occurred (e.g., the absence of course assessment). Real-time measures of sympathetic nervous system activity, such as electrodermal activity and heart rate variability, might have offered a more accurate assessment of the actual levels of anxiety experienced (Kantor et al., 2001).

5.2. Applications, and implications

Anxiety is a complex phenomenon with several facets: somatic, behavioral, and cognitive (appraisal concerns and communication apprehension). Interventions targeting each specific form of anxiety may be effective in weakening the relationship that links it to output measures. Writing strategy-based facilitation (Tsiriotakis et al., 2017), relaxation techniques, and group discussions in which students can realize that their anxiety is not a unique feature of their existence may be helpful. Interventions targeting self-efficacy (i.e., confidence in one’s abilities) (Pilotti et al., 2023; Sabti et al., 2019) and metacognition (Allmnakrah & Evers, 2020; Yeh, 2015) may also be helpful in addition to a heavier reliance on computational thinking skill development (Peng et al., 2023).

Notwithstanding limitations, the findings of the present research have implications for practice. To this end, it is important to note that our research was guided by the assumption that although good writing performance in general education courses at the start of a student’s academic journey is positively linked to retention and persistence (Webb-Sunderhaus, 2010), writing anxiety has the opposite relationship (McAllister et al., 2017; Stearns, 2023). Thus, our study focused on the relationship between writing anxiety and precursor factors (such as exposure and attitudes) or output measures. As such, the evidence collected adds to the findings of research on predictors of academic attainment. In this type of research, individual differences, which are assumed to be markers of students’ persistence and retention, inform admission, instruction, advising, and counseling practices. Although most of the research has targeted pre-college measures (e.g., high-school GPA and SAT scores) (El-Moussa et al., 2021), and first-year performance (Jackson & Kurlaender, 2014), basic competencies, such as those upon which writing rests, are also recognized as markers of persistence and retention in students’ academic pursuits (Nicholes & Reimer, 2020; Simmons, 1994).

6. Conclusion

Sustainable education in a traditionally patriarchal society, which is being reengineered from the top to embody a knowledge- and service-based economy, requires that both men and women contribute. Thus, in Saudi Arabia, gender equity starts with equal opportunities and ends with performance outcomes that fulfill the expectations that the envisioned society has for both females and males. In this socio-cultural context, female students of college age are the fulcrum of the engine of transformation, thereby making their professional success a necessary step for the success of the 2030 Vision.

Our study was prompted by a broad question about how female college students of Saudi Arabian descent would respond to the pressure to succeed. We believed that a particular skill that is key to their academic and professional success (i.e., writing in a second language) would adequately illustrate their responses. Namely, did they approach the challenges of writing in a second language with a threat framing, including negative emotions, such as anxiety (Pilotti et al., 2021; Vandewalle et al., 2019)? Alternatively, did they approach such challenges with an opportunity framing, which would considerably minimize anxiety? As a group, such young women did not report experiencing much writing anxiety (as illustrated by the descriptive statistics of Table 1). Thus, we concluded that for the selected female sample the prevailing approach to the challenges involved in second-language writing was opportunity framing. Such an approach induces one to see gains as more salient than losses. In contrast, threat framing induces students to perceive losses as more salient. Thus, it is not surprising that opportunity framing is ideal for dealing with change, whereas threat framing encourages the avoidance of change (Sanner & Evans, 2019; Taylor & Armor, 1996). Given the societal changes that the women of the selected sample have faced in their recent memory and probably will continue to face in the future, the adoption of opportunity framing may secure their success. Whether they will be able to sustain such an approach as challenges mount is a matter to be investigated further.

Acknowledgments

We would like to express our appreciation and gratitude to the students who participated in the study and to the members of the Undergraduate Research Society who contributed to the data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Appendix 1

The 22 Items of the Second Language Writing Anxiety Inventory (SLWAIr). An R in parentheses refers to items that require reverse scoring.

Somatic Anxiety:

I often feel panic when I write assignments in English under time constraints.

I tremble or perspire when I write assignments in English under time pressure.

I feel my heart pounding when I write assignments in English under time constraints.

I usually feel my whole body to be rigid and tense when I write assignments in English.

I freeze up when I am asked to write assignments in English.

Appraisal Concerns (Cognitive Dimension):

I’m afraid that other students would not like my assignments written in English if they read them.

I’m afraid of my assignments written in English being chosen as samples for discussion in class.

If one of my assignments written in English is to be evaluated, I would worry about getting a very poor grade.

While writing in English, I feel worried and uneasy if I know my work will be evaluated.

While writing in English, I often worry that the ways I express and organize my ideas do not conform to the norm of English writing.

While writing in English, I often worry that what I say will not be understood by others, such as my instructors.

Avoidance Behavior:

Whenever possible, I would use English to write assignments. (R)

I usually seek every possible chance to write in English outside of class. (R)

I often choose to write down my thoughts in English. (R)

I usually do my best to avoid writing in English.

I do my best to avoid situations in which I have to write in English.

Communication Apprehension (Cognitive Dimension):

My thoughts become jumbled when I write assignments in English under time constraints.

My mind often goes blank when I start writing an assignment in English.

While writing in English, I’m not nervous at all. (R)

I usually feel comfortable and at ease when writing in English. (R)

When I write in English, my ideas and words usually flow smoothly. (R)

When I write in English, my mind is usually very clear. (R)