Extramural activities (EA) are of particular interest to language learning researchers because learners engage in them so frequently and readily – which also makes them especially difficult to research. De Wilde et al. (2020, p. 180) found three activities to be especially fruitful: “use of social media in English, gaming in English, and speaking English”. They belong to the more interactive activities, which require the learner to use English, rather than just consume it. Consuming English-language media, in turn, has been shown to be beneficial for understanding the form-meaning connection of words and for language comprehension in general (cf. Peters, Heynen & Puimége, 2016; Peters & Webb, 2018). Social networking services (SNS) such as Instagram, Facebook, or Twitter have furthermore been explored and show benefits for learners’ willingness to communicate in the foreign language as well as their writing fluency (cf. Dzion, 2016) and overall language confidence (cf. Bailey & Almusharraf, 2021). However, learners express themselves differently on the internet than they do in “instructionally designed language teaching and learning practices in schools” (Lantz-Andersson, Vigmo & Bowen, 2013, p. 294). Users often sacrifice proper grammar and spelling to get messages across in fewer characters (cf. Sirucek, 2009). Since it is still questioned whether the language used online, also called netspeak or Internetese (cf. Angel, 2022), is beneficial to the language used within the classroom, this study was conducted with German EFL, connecting their in-class writing to their EA. The hypothesis is that the English learned online (through Extramural English (EE)) is a different variety of English than the one used in the classroom and thus, no large effects on productive vocabulary and lexical diversity can be found in the learners’ in-school writing. This study is based on a detailed survey on EE and a corpus collected from texts written for an exam. In addition, the informants fulfilled one discourse completion task (DCT) asking them to write a text message to a friend, informing them of a secret. A total of 43 informants (age m=17.77) participated in the study. They will all graduate from a German high school (Gymnasium) and have received the same type of formal education. On average, they spend 2.3 hours per day on TikTok, 1 hour on Instagram, and even more time streaming content in both English and German on platforms such as Netflix, YouTube, Twitch, and the like (cf. poster). In particular, more than half of the participants use Twitter, TikTok, and Instagram in English, while chat platforms such as WhatsApp are almost exclusively used in German. All participants consume English audio and visual content daily and thus, the participants overall consume a lot of English-language content in varying intensity. However, they rarely produce much of it themselves. To explore the productive vocabulary, lexical frequency profiles (LFP) were used. Laufer and Nation (1995) view the LFP as an advanced method for measuring vocabulary size, as it is independent of a curriculum and can be used to compare learners from different backgrounds. In this study, the LFPs of the exam texts show no significant correlation between time spent online in English, participation in EA, and lexical diversity in writing. As the pie charts show, the German pupils as a whole perform significantly worse than a control group of native speakers writing on similar questions. Even if the amount of EE is considered, the group participating intensively in EE does not produce significantly better LFP. The (intensive) participation in EE does therefore not directly predict good LFPs. Thus, the question is whether the language consumed online is beneficial for Standard English expected in school writing. Since internetese is a variety of English, competence in internetese should be tested to find out if EE has an effect on its use. A first attempt at such a competence test was made by using the DCT to prompt the production of a text message (cf. adjacent poster). Of the 37 students who produced a text message, 20 participants demonstrated very confident use of internetese by using acronyms, internet slang, and sentence structures typical of the variety. The more proficient users of internetese have several things in common (cf. poster): they indicated high confidence in their English skills, they game less but watch more English content, they spend more time on average on SNS and text more in English, and they read more, especially fanfiction. Overall, there seems to be a relationship between consuming English media, engaging in online use of English, and the competences in the English internet variety. However, the use of EA does not significantly predict the use of lexically diverse vocabulary in school writing. Further studies should aim to create a proper test of internetese competence to explore this further. Additionally, looking at other skills besides writing might show which skills improve by the use of the internet.