The dissertation examines Finnish upper-secondary school second language use from four perspectives: the development of lexical knowledge in relation to extramural activities (Article I); the role of lexical recognition in searching online sources for words and information (Article II) digitally identified cohesive features in writing (Article III); and lexical sophistication as examined by traditional lexical tests and digital analyses (Article IV). The study is situated in the context of second language acquisition with implications for teaching, and assessment building on previous research by examining lexical recognition skills across the high-, mid-, and low-frequency bands, triangulating information from questionnaires and video-observations with data elicited from traditional lexical tests and digital analyses (TAACO and TAALES). The studies in Articles I and IV were conducted longitudinally, while Articles II and III focus on examinations conducted in the second year.

Article I demonstrates that using English in cognitively demanding extramural activities, such as reading and gaming, develops both overall lexical recognition skills and recognition of infrequent lexis. In the second year, these activities collectively explained 45% of the variance in the scores for infrequent lexis. Article II focused on using online sources in diverse indirect writing tasks. The results indicated that consulting online dictionaries and informational web-pages required rapid lexical recognition, multiple reading strategies and adequate digital skills to formulate queries and evaluate the search results, and that these abilities were directly associated with the participant’s lexical recognition skills. According to the analyses, participants scoring less than 60‒64% in the lexical recognition test (the VLT), did not benefit from consulting online dictionaries and other sources in diverse writing tasks.

In Articles III and IV, writing skills were also examined using digital analyses. Article III investigated cohesive devices in essays written in the second year. Digital analyses (TAACO) showed that using diverse referential devices, e.g., adverbs and connectors, explained 37% of the variance in the essay ratings. In Article IV, written production was examined longitudinally using traditional lexical tests and a digital tool (TAALES). Digital analyses combining three indices (function word frequency, infrequent content words and typical English two-word combinations) collectively explained 46% in the variance of the first-essay scores and 44% of the variance in the second-essay scores. Corresponding traditional tests on associative word knowledge and recognition of two-word combinations explained 56% in the variance in the first essay scores and 61% of the second year scores. Thus, the results elicited by traditional tests were in line with the digital results. This finding implies that traditional tests for recognition and associative skills can also be used to assess lexical sophistication in language teaching.