Effects of Age and Familiarity on Visual Word Recognition in Korean: Evidence from a Web-based Large Scale Lexical Decision Task

Abstract

Purpose One of the most studied and verified psycholinguistic phenomena is a word frequency effect, which refers to the tendency for more frequent and more familiar words (e.g., life) to be recognized more efficiently than less frequent and less familiar words (e.g., dove) (Brysbaert et al., 2011; Brysbaert et al., 2018). However, the influence of aging on the word frequency effect is still under discussion. The frequency effect may remain unaffected by age in some cases (Cohen-Shikora & Balota, 2016), or it may grow stronger (Rayner et al., 2006) or weaker (Gollan et al., 2008) with age. While most of these earlier studies aimed at Roman-script languages such as English or German, relatively little research has examined age effects in Korean word recognition. Moreover, a number of attempts have recently been made in behavioral sciences to collect data through web-based experiment platforms, which enables more efficient data collection with no face-to-face meeting with every participant (e.g., Aguasvivas et al., 2020). Thus, the present study investigated age differences in familiarity effects in Korean visual word recognition through a web-based large scale lexical decision task. Method A lexical decision task on 120 Korean words varying in frequency and 120 nonwords was built using PsychoPy3 and made available online at Pavlovia.org. Fifty-eight adult Korean speakers, whose ages were evenly distributed from the 20s through the 60s, rated subjective familiarity of each word on a 7-point scale. Another group of 497 adult Korean speakers were recruited from an online survey respondent panel. This group also consisted of similar numbers of respondents across five age groups: 20s, 30s, 40s, 50s, and 60s. They participated in the lexical decision task using a URL that they received via email. They then took the Korean Author Recognition Test and the Comparative Reading Habits survey as a measure of reading skills. The effects of familiarity, age, and reading skills on lexical decision accuracy and latency were analyzed using mixed-effects regression models.  Results and Discussion Overall, both lexical decision accuracy and RT increased with age. More familiar words were recognized more accurately and more rapidly, which replicates the word frequency effect found in previous lab-based studies. Higher reading skills were associated with shorter RT, and this effect was stronger among younger participants. Crucially, there was a significant interaction between age and familiarity, which was mainly driven by younger participants’ noticeably less accurate and slower responses to low-familiarity words. This result is in the same vein as Gollan et al.’s (2008) results and supports the Lexical Quality Hypothesis (Perfetti & Hart, 2002), which posits that accumulated language experience improves the quality of lexical representations of words and that this learning effect is more prominent in low-frequency words than in high-frequency words due to a ceiling level in lexical quality. Finally, unlike in previous studies (Kuperman & Van Dyke, 2011), there was no significant interaction between familiarity and reading skills.