To account for the specificities of writing, Berninger and colleagues proposed the Not-So-Simple View of Writing (Berninger and Winn, 2006; Berninger and Chanquoy, 2012). According to this model, text generation requires the interaction between transcription and executive functions (EFs). Transcription is the process of transforming oral language into written text and involves handwriting/typing and spelling (Limpo et al., 2020). Correlational and experimental studies showed that handwriting fluency and spelling accuracy are positively associated with writing performance in primary and secondary school students (Limpo et al., 2017; Hurschler Lichtsteiner et al., 2018; Limpo et al., 2020; Skar et al., 2021). Besides transcription, EFs are important contributors to the development of written language (Limpo and Olive, 2021). EFs encompass a diversity of cognitive processes for engaging in goal-directed behavior and effectively addressing problems (Alves, 2019; Willoughby and Hudson, 2021). The EFs components of inhibitory control, planning, and working memory have been linked to writing (Hooper et al., 2011). Empirical findings also showed relationships between transcription and EFs. For example, Vieira et al. (2023) showed that transcription skills and EFs were related with each other in Grade 2 and 1 year later in Grade 3, but prior performance on these skills did not predict later text quality. Despite emphasizing the importance of cognitive processes, these studies did not consider the role of motivation, specifically, self-efficacy and attitudes, which seem to be the most relevant ones for writing (Camacho et al., 2021).

Writing self-efficacy refers to individuals’ opinions regarding their ability to learn and perform writing tasks (Bruning et al., 2013). According to Bandura’s socio-cognitive theory (Bandura, 1997), the level of effort and persistence that students invest in achieving their goals are influenced by these self-perceptions. Those who doubt their abilities tend to show low commitment, frustration, anxiety, and, ultimately, passiveness and avoidance. Conversely, those who believe in their abilities tend to try harder, persist, and engage in adaptive behaviors. Self-efficacy is one of the strongest motivational predictors of writing performance (Zumbrunn et al., 2020; Camacho et al., 2021). Attitudes toward writing are generally regarded as the pleasure or satisfaction writers derive from engaging in writing activities (Ekholm et al., 2018). Specifically, they are characterized by a range of emotional reactions toward writing, encompassing positive and negative emotions (Graham et al., 2007). In line with WWC model, students with positive attitudes toward writing were found to write longer and better texts (Ekholm et al., 2018; Graham et al., 2019; Camacho et al., 2021). This seems to happen because positive affect may facilitate the learning process, and liking writing may translate into more time and energy spent in this activity (Graham et al., 2007). Given the well-documented role of self-efficacy and attitudes in writing, we targeted these two constructs to examine the underexplored link between cognitive and motivational processes.

As described above, there is now substantial theoretical and empirical support for the claim that children with higher levels of cognition and motivation produce better written outcomes. However, little is known about the joint contribution of these cognitive and motivational variables to children’s writing. Two recent theoretical models included both cognition and motivation as central processes in writing.

Under the component of writers’ resources and capabilities, the Writer(s)-Within-Community (WWC) model (Graham, 2018) included not only mental and physical operations (such as transcription) but also control mechanisms (such as executive control) and motivational beliefs (such as attitudes and self-efficacy). These variables are expected to be related to each other. Students who are more motivated to write seem more likely to be attentive and adopt strategic behavior to accomplish writing processes, like putting words onto paper (Graham et al., 2019). Yet, this model did not provide further detail on the link between cognitive and motivational variables. The specification of such a link was advanced by Kim and colleagues in the Direct and Indirect Effects model of writing (DIEW; Kim and Schatschneider, 2017).

The DIEW postulates hierarchical relationships between component skills, namely, transcription skills, higher-order cognitive skills, oral language, motivation, prior knowledge, and EFs (Kim and Park, 2019). Authors hypothesized that EFs and transcription could make direct contributions to writing, as well as indirect, via motivation. Limpo and Alves (2013) found partial support for this claim in middle graders, by showing that self-efficacy mediated the link between transcription and writing. Though similar findings were not found concerning EFs or attitudes, the DIEW model and these results seem to provide sufficient ground to hypothesize that children with better writing-related cognitive skills may have better writing proficiency due to higher motivation. Research into the cognition-emotion link may also help to support this yet-to-be-tested hypothesis. Because a positive emotion may require fewer cognitive resources than a negative one (Pekrun et al., 2002; Coffey, 2020), greater motivation in writing may free up cognitive resources for the task, resulting in better products (Rocha et al., 2019). As positive emotions have been associated with more adaptive forms of cognitive engagement (Isen, 1999), more motivation may also benefit the enactment of cognitive writing processes.

In this study, 157 Portuguese students participated, 92 from grade 3 and 65 from grade 4 (80 boys, 77 girls) with an average age of 8.73 years (SD = 0.61; range from 7.83 to 10.17 years). We used a convenience sample comprised of students from five classes from a public group of schools in the North of Portugal, with which our team has cooperation protocols. The socioeconomic status of the children was determined through their mothers’ education level, which was as follows: 1.3% completed Grade 4, 8.3% completed Grade 5–6, 21% completed Grade 9, 24.8% completed high school, and 37.6% completed college or attained a higher level of education and 7% was unknown.

All students were evaluated in a 60-min group session. Firstly, students were presented with the essay writing topic: “Do you think teachers should give students homework every day?.” Afterward, students had 10 min to plan the essay followed by 15 min to compose it. Secondly, students performed two transcription tasks, in which they were asked to copy a 9-word sentence as quickly as possible during 90 s (Limpo and Alves, 2018), and spell 16 words dictated at intervals of 6-to-10 s (Limpo and Alves, 2018). Thirdly, students filled in the attitudes and self-efficacy questionnaires. The experimenter indicated that, for both questionnaires, there were no right or wrong answers and explained the overall procedure. Items were read aloud, each item at a time, and students were asked to mark their answers individually. Moreover, schoolteachers completed an EF questionnaire about each student.

An overview of all measures is presented on Table 1, including validity and reliability data.

Self-efficacy for conventions refers to writers’ confidence in relying on the use of tools related to the language of writing to express one’s ideas (Bruning et al., 2013). Results showed that about one third of the variability in this dimension was explained by cognitive variables, namely, working memory, handwriting fluency, and spelling accuracy. One of the factors contributing to students’ greater perceived self-efficacy for conventions was working memory. There is a consensus that working memory is fundamental for the production of a good text, including language production. Verhagen and Leseman (2016) showed that, in the first grades, students with greater working memory capacity exhibited greater grammatical and vocabulary skills. Given this association between working memory and language skills, it is not surprising that students with higher levels of working memory may feel more capable of successfully using those skills (Vellutino et al., 2007). Along with working memory, transcription also predicted self-efficacy for conventions. This result aligns with previous findings (Limpo and Alves, 2013), showing that transcription predicted general self-efficacy. Together these findings reinforce the role of students’ handwriting and spelling abilities as a fundamental source of self-efficacy.

Self-efficacy for ideation refers to writers’ confidence in generating ideas for writing (Bruning et al., 2013). Our findings demonstrated that, besides predicting self-efficacy for conventions, working memory also predicted self-efficacy for ideation. This link between higher working memory resources and stronger beliefs on own’s ability to generate ideas reinforces the key role of working memory in writing. Working memory allows students to group active thoughts in memory and link them with information stored in long-term memory (McCutchen, 2000). Thus, students who have greater working memory may perceive it to be easier to remember the objectives of the task, while composing and organizing ideas. Given the importance of working memory to generate ideas (Cordeiro et al., 2019) and formulate language (Verhagen and Leseman, 2016), students with greater working memory resources may feel more competent in producing relevant ideas and choosing appropriate linguistic forms to express them accurately (Martinussen and Major, 2011).

Self-efficacy for self-regulation refers to writers’ confidence to manage their writing behavior and affect (Bruning et al., 2013). Our results revealed that this dimension of self-efficacy was predicted by spelling, suggesting that the more automated children’s spelling is, the greater their perception of behavioral and affective control in writing. When spelling skills become sufficiently accurate, writers may use spare attentional resources for higher-level processes, including self-regulation (Berninger and Winn, 2006; Limpo and Alves, 2013). This may explain why students with better spelling skills may feel more competent to manage writing. The fact that this link only involved spelling and none of the other cognitive variables assessed was unexpected, given past findings showing the importance of EFs for self-regulation (Lyons and Zelazo, 2011). It should, however, be noted that, we measured students’ self-perceived self-regulation skills, which may explain the absence of links. Yet, because our study does not allow us to test this or other hypotheses, more research relating EFs and students’ real and perceived ability to self-regulate writing seems needed.

Although the percentage of explained variance was low (5%), the results demonstrated a relationship between cognitive variables and attitudes. We found that only inhibition and spelling predicted attitudes. It seems that by having stronger abilities to ignore distractions and adopt more thoughtful, less impulsive approaches to writing, children may experience writing as a more pleasant experience. The same relationship was, however, not found for working memory, contradicting theoretical models, such as the WWC or the DIEW, and past findings relating middle graders’ working memory with attitudes toward learning (Jung and Reid, 2009). Despite no prior writing study has explored the link between working memory and attitudes, these later are multifaceted and shaped by a range of factors, including cognitive ones (Camacho et al., 2020). Whether working memory is one of them deserves further attention.

The link between spelling and attitudes may be related to the key role that the transcription has in primary education. Portuguese students struggle to learn spelling and produce several misspellings, which receive repeated corrective feedback from teachers and foster an attitude of demotivation and avoidance toward writing. Therefore, it is essential that teachers adopt a procedural correction approach to writing, in which they value intermediate work, correct multiple drafts, emphasize content and form, and praise writers’ efforts (Cassany, 2000). In this way, teachers will be contributing to the development of students’ positive attitudes toward writing.

We found partial support to the hypothesis that transcription, self-efficacy, and attitudes would predict writing performance, assessed through the amount of writing and the complexity of plans. Concerning text length, results indicated that students with higher handwriting fluency and more positive attitudes toward writing wrote longer texts. These findings reinforce the importance of handwriting for writing performance (Santangelo and Graham, 2016; Limpo and Graham, 2020). Students’ attitudes also influenced the amount of their writing output. Therefore, the more positive the students’ attitudes toward writing, the longer their written productions. In line with the WWC model, studies indicate that the more students enjoy writing, the more often they write and the more energy they dedicate to the task, which is fundamental to producing good writing (Graham and Harris, 2016).

Concerning planning, our findings showed that spelling accuracy and attitudes were significant predictors. Given the multiplicity of simultaneous processes that compete for the focus of writers’ attention during composition, being able to write accurately emerges as a significant advantage for young students (Graham et al., 2002). The present study reinforced the importance of spelling in writing, by showing that these skills not only seem to impact the final product, but also a key processual component of it. This is an important finding because planning is critical for the production of high-quality texts throughout schooling (Graham and Perin, 2007; Graham et al., 2012).

Besides spelling, students’ attitudes also predicted planning. Previous studies already showed that students with more positive attitudes toward writing wrote longer and better texts (Graham et al., 2017; Ekholm et al., 2018). To our knowledge, this is the first study showing a link between attitudes and written planning. In line with the claim that that positive (vs. negative) emotions require fewer cognitive resources (Coffey, 2020) and stimulate more adaptive cognitive engagement (Isen, 1999), this finding suggests that favorable writing attitudes may foster planning by freeing up cognitive resources for relevant processes, such as the generation and organization of ideas.

Contrary to past findings (Pajares, 2003), self-efficacy belief did not predict writing performance. This was a surprising result, which can be explained by the indicators of writing performance we used, planning and text length, whose link with self-efficacy has been barely explored. Typically, self-efficacy has been shown to predict writing performance indexed through text quality. Together, past findings and ours suggest that attitudes, rather than self-efficacy, may have a stronger role in shaping processual measures, such as planning. Further research is needed to support this claim by comparing the differential effects of self-efficacy and attitudes on process and product measures.

Despite the associations between cognitive and motivational processes, as well as between these latter and writing performance, our hypothesis that cognitive processes would be associated with writing performance via motivation was not confirmed. This was the first study testing this claim, making it harder to interpret this lack of mediation. Though it may mean that the link between cognitive processes and writing performance is not mediated by motivation, but by other variables, it may also be a reflex of the variables assessed. Our model left out cognitive processes such as attention, cognitive planning, cognitive flexibility as well as motivational beliefs such as intrinsic and extrinsic motivation derived from Self-Determination Theory (Ryan and Deci, 2000), which have been shown to have a key role in writing. We do not know if targeting other cognitive and motivational variables would have uncovered the motivational mechanisms expected to mediate the link between cognition and performance in writing.