The communicative practices of contemporary society are shaped by the widespread presence of computer-mediated communication (CMC). This communicative reality is configured through computer- mediated discourse (CMD), understood as “the communication produced when human beings interactwith one another by transmitting messages via networked or mobile computers” (Herring & Androutsopoulos, 2015, p. 127). The main feature of computer-mediated discourse (CMD) is multimodality (Herring, 2019), by which CMC can support the simultaneous presence of different modes of transmission including not only text, but also voice, audio, video, and graphics. As Candefors Stæhr et al. (2019, p. 171) highlight, “people are already socialized in a communicative reality with many possibilities for interpersonal communication across online and offline contexts through different modes of interaction”. In this sense, smartphone messaging is one of the most popular and extended ways of interaction within the large spectrum of possibilities offered by CMC (Yus, 2022).

According to data from IAB Spain-Elogia (2022), 92% of adolescents between 12 and 17 years of age regularly use WhatsApp as an instant messaging (IM) program. WhatsApp has thus become the most used social media in this age group, surpassing other platforms such as Instagram (83%), TikTok (77%) or Facebook (39%). This is the first fully connected generation, halfway between the so-called Generation Z and Generation Alpha, which has a complete mastery of digital interactions and for whom the use of technological devices is connatural to their personal and cognitive development, leading them to show a “fascination with electronic devices” (Castro et al., 2020).

The wide diffusion of this type of instant messaging programmes among teenagers has favoured the creation of a new written code called digitalk (Turner, 2010) or textese (Johnson, 2015), characterised by the fact that it differs from the standard writing norm. Digitalk combines elements of written discourse with features of orality in an attempt to reproduce the “voice of the speaker” (Turner, 2010, p. 43) through creativity and communicative language proficiency. In fact, it has sometimes been defined as “oralised writing” (Martín Gascueña, 2016), “oralised written text” (Yus, 2010) or “written spoken language” (Mancera Rueda & Pano Alamán, 2013).

One of the main characteristics of digitalk is the use of textisms, “contractions and non-standard spellings specifically developed to reduce the length of words for fast and cost-effective text messaging” (De Jonge & Kemp, 2012, pp. 49-50). Textisms can also include features specific to the digital medium, such as the use of multimodal elements like emoticons, emojis, images, videos, or stickers, among others (Aull, 2019; Cantamutto & Vela Delfa, 2019; König, 2019), which fulfill different pragmatic and stylistic functions in CMC (Padilla, 2023; Sampietro, 2019; Vela Delfa, 2020). The digitalk is especially sensitive to technological advances, so that “it becomes quicker and easier to type messages” without any changes in attitudes about the appropriateness or otherwise of the use of textisms (Kemp & Grace, 2017, p. 96).

In relation to the Spanish language, digitalk has been described in detail over the last decade. Numerous studies detail the characteristics of digitalk, both with respect to European Spanish (Calero, 2014; Caurcel Díaz et al. 2013; Ibarra Murillo, 2019; Mancera Rueda, 2016; Mancera Rueda & Pano Alamán, 2013; Mas Álvarez & Zas Varela, 2012; Vázquez-Cano et al, 2015) as well as its use in some of the varieties of American Spanish (Cantamutto, 2018, 2019; Cantamutto & Vela Delfa, 2018; Flores-Salgado & Castineira-Benítez, 2018; Giraldo Giraldo et al., 2018). According to Mancera Rueda (2016), digitalk is characterised by the

presence of elements that can be classified into two main categories: antiorthography (Palazzo, 2005) and heterography (Martínez de Sousa, 2004) or dysorthography (Gómez-Camacho, 2007). The first group includes unintentional deviations from the orthographic norm, i.e. spelling mistakes (errors in accentuation or in the use of letters) caused by ignorance or insufficient knowledge of the linguistic norm on the part of the speaker. For their part, heterographies or disorthographies are phenomena specific to digitalk, insofar as they are intentional and conscious deviations from the orthographic norm with an expressive function in the digital communicative context.

Gómez-Camacho et al. (2018, p. 94) offer a categorization of textisms in Spanish. This classification can be organized into three levels: graphophonemic, lexical-semantic, and multimodal textisms, although these textisms can appear simultaneously in a single element (Table 1). The graphophonemic level includes those textisms based on the discrepancy between phoneme and grapheme, such as the phenomena of word shortening or reduction, the use of non-normative graphemes, or the repetition and omission of graphemes, punctuation marks, or capital letters. Textisms at the lexical-semantic level include the use of dialectalisms, the creation of new words, and the use of foreign words. Finally, the multimodal level contains the use of emoticons, images, videos, audios, or stickers. Table 1 shows the reformulation of Gómez-Camacho et al. (2018) types of textisms with examples extracted from the corpus analized.

The spread of digitalk has generated extensive debate about its interaction with the standard writing norm and its possible negative influence on the literacy of children, adolescents, and adults (Zebroff, 2018). Research shows that the influence of digitalk and textisms on literacy can be associated with certain factors and does not affect all language skills equally. Bernicot et al. (2014) argue that factors that may influence the way in which digitalk interacts with grammatical competence are: the type of spelling (normative or usage), the level of written competence (skilled vs. less-skilled), the duration of practice with digitalk, or the type of textisms used. This would justify the very disparate and contrasting results observed in the literature.

In addition, the age and gender of users also determine the way in which digitalk is used; in this regard, several studies show that younger users use more textisms than older users (Hilte et al., 2017; Verheijen, 2018) and that women incorporate more textisms and multimodal elements in their digital communications (Grace & Kemp, 2015; Rosen et al., 2010; Tossell et al., 2012).

On the one hand, numerous studies show a negative correlation between the use of textisms and certain aspects of communicative competence, such as morphological awareness (De Jonge & Kemp, 2012), reading accuracy (Drouin, 2011; Drouin & Driver, 2014) or vocabulary (Drouin & Driver, 2014).

Other research, however, has clearly shown a limited or even positive influence between the use of textisms and literacy skills in different languages, highlighting that “the use of grammatical violations does not appear to be linked to changes in grammatical skills over time” (Wood et al., 2014, p. 427); these studies consider that digitalk is not a threat in the language skills of children, adolescents, or adults. Van Dijk et al. (2016) observe that word omission in digital writing is closely linked to grammar performance in Dutch, in that “the more words children omitted in their text messages, the better their grammar performance” (Van Dijk et al., 2016, p. 16). Similarly, a positive correlation has been observed between the use of instant messaging applications and higher spelling performance in Dutch (Verheijen et al., 2020; Verheijen & Spooren, 2021) and English (Plester et al., 2009; Wood et al., 2011). Lanchantin et al. (2015, p. 21) highlight, in their study with French speakers, that “adolescents rely on their literacy knowledge to notice what could be deleted or replaced in a DWIM [Digital Writing in Instant Messaging] production”, confirming that the participants in their research showed a high morphemic awareness that induced them to delete morphemes without phonetic value in their digitalk and to keep them in their traditional writing.

Finally, in terms of the relationship between textisms and spelling competence, Bernicot et al. (2014) showed how the type of textism used became a determining factor. Thus, textisms and traditional spelling could be negatively, neutrally, or positively correlated, depending on whether the textism broke or coincided with the relationship between graphemes and phonemes established by the standard writing norm. Based on these studies, we can conclude that IM app users are aware that digitalk and standard writing are two different codes, “whose acquisition is dependent on the same symbolic abilities” (Bernicot et al., 2014, p. 570), but which allow “a willing playful engagement with language” (Plester et al., 2011).

In the context of the Spanish language, research has focused on the analysis of students’ or preservice teachers’ perception of the influence of instant messaging on their own or adolescents’ spelling competence. Although the results show, in general, a negative perception of digitalk over standard writing (Cremades et al., 2021; Giraldo Giraldo et al., 2018), some studies reveal a greater tolerance among young IM users toward three types of textisms that do not contradict the academic norm, i.e., textisms based on multimodal elements (Gómez-Camacho et al., 2018), those that do not alter the relationship between phonemes and graphemes (Hunt-Gómez et al., 2020) and those based on the creation of neologisms (Núñez-Román et al., 2021).

Despite a conspicuous amount of research in other languages, in the context of the Spanish language, research comparing real texts from the secondary school classroom and interventions taken directly from the WhatsApp chats of 14-16 years-old is scarce. Therefore, the aim of this research is to describe the characteristics of the digital norm of Spanish adolescents in their interactions through WhatsApp and to explore the relationship between the use of textisms and the normative orthography of their academic texts. It will also analyze the existence of possible differences in relation to the user’s gender. Thus, it is a novel study of particular relevance in the field of Spanish. To this end, a secondary school involved in educational

research projects at the university, national, and European Union level was selected for the case study. It is also involved in educational action research projects in which secondary school teachers participate, and in educational innovation projects to incorporate ICT into language teaching. The research questions to be addressed are as follows:

In this study, the Secondary School IES Jesús del Gran Poder located in Dos Hermanas (Seville, Spain) was selected on the basis of the following criteria:

The sample of participants, which makes up the case study, consists of 43 students, of which 75% are female and 25% are male. Participants are enrolled at IES Jesús del Gran Poder in Dos Hermanas (Seville, Spain). Specifically, 70% of the sample is in the third year of compulsory secondary education and 30% is in the fourth year of compulsory secondary education, levels attended by students aged 14 to 16.

The data were collected through real texts provided by the students participating in the study collected, on the one hand, from instant messaging applications on their smartphones (in this case WhatsApp) and, on the other hand, from academic texts produced as an activity in the secondary school classroom.

A research team composed of 30 active language teachers of secondary education belonging to the pro- ject “Digital writing of secondary education students in Andalusia. Educational implications of instant mes- saging” collected the texts as a classroom activity and as part of the syllabuses of secondary education lan- guage subjects1. The participating students selected an uncorrected handwritten text that they considered representative of their way of writing in class (notes, essays, etc.). At the same time, the participating students selected WhatsApp messages, also uncorrected and unmodified, which they considered representative of their way of writing through this application, and which had been sent in the context of digital communication with other adolescents, previously eliminating any personal reference or confidential data.

The interventions in WhatsApp of the 43 participants in the study were analyzed, considering as interven- tion the text written by the subject preceding the click on the Send button in the App. An average message- intervention in the social network of 20.14 (SD=2.406) was obtained. The message rate ranged from a mini- mum of 13 interventions to a maximum of 33. In total, 866 interventions were collected.

Every WhatsApp single intervention was coded by two evaluators from the project “Digital writing of adolescent students in Andalusia. Instant messaging and its educational implications”, according to the textisms established by Gómez-Camacho et al. (2018) (Table 1) for the Spanish language in a previous study of this research. These categories were validated by six researchers who independently coded the collected material, unified the criteria, and made decisions related to divergent elements, thus establishing the final categories for the analysis (Table 2). The results obtained were pooled, discrepancies were resolved according to the unified criteria established by the six researchers, and a single final coding was carried out.

The academic texts were coded according to the orthographic categories established by the Royal Academy of the Spanish language in the Orthography of the Spanish Language (Real Academia Española, 2010, pp. 72-160) (Table 3).

To process the data, two record sheets are created in GoogleForm, one for the analysis of textisms that appear in student messages posted on the WhatsApp social network and the other for the analysis of spelling errors that appear in academic texts.

All variables are scale variables. In addition, a total variable is generated in each of the cases that corresponds to the total number of spelling errors made in academic texts and the total number of textisms in WhatsApp messages.

The textisms and misspellings of each participating student were analysed by two evaluators from the project “Digital writing of adolescent students in Andalusia. Instant messaging and its educational implications”, according to the categories described in Table 2 and Table 3.

To answer the first and second research question, statistical analyses were performed for the descriptive results of central tendency (mean) and dispersion (standard deviation, hereafter SD; and minimum and maximum). Bivariate correlation analyses were also carried out by applying Pearson’s coefficient to check the degree of relationship between the scale-type variables analysed.

Subsequently, the textism density parameter coined by Drouin (2011) and De Jonge and Kemp (2012) was calculated by dividing the total number of textisms by the total number of words in the sample texts (Kemp et al., 2021). Every textism appearing in each intervention was quantified, except for textisms from the category Emphatic Repetitions (Table 1), which were computed as a single occurrence in each intervention. The same procedure was used to calculate the density of spelling errors in academic texts.

To answer the third research question, inferential analyses were also carried out according to gender. The KS and Shapiro Wilk test determined the normality of the dependent variables, Student’s t-test (previous Levene’s test to determine equality of variance) was applied for parametric variables and the Mann Whitney U-test for nonparametric variables. IBM SPSS v.26 software was used for statistical analysis and GPower software was used to calculate the effect size.