How do Spanish speakers read words? Insights from a crowdsourced lexical decision megastudy

Abstract

Vocabulary size seems to be affected by multiple factors, including those that belong to the properties of the words themselves and those that relate to the characteristics of the individuals assessing the words. In this study, we present results from a crowdsourced lexical decision megastudy in which more than 150,000 native speakers from around 20 Spanish-speaking countries performed a lexical decision task to 70 target word items selected from a list of about 45,000 Spanish words. We examined how demographic characteristics such as age, education level, and multilingualism affected participants' vocabulary size. Also, we explored how common factors related to words like frequency, length, and orthographic neighbourhood influenced the knowledge of a particular item. Results indicated important contributions of age to overall vocabulary size, with vocabulary size increasing in a logarithmic fashion with this factor. Furthermore, a contrast between monolingual and bilingual communities within Spain revealed no significant vocabulary size differences between the communities. Additionally, we replicated the standard effects of the words' properties and their interactions, accurately accounting for the estimated knowledge of a particular word. These results highlight the value of crowdsourced approaches to uncover effects that are traditionally masked by smallsampled in-lab factorial experimental designs.

Which words do English non-native speakers know? New supernational levels based on yes/no decision

“…Indeed, many of these words are likely to be cognates of words in their L1 or even in another L2. Keuleers, Brysbaert, Stevens, and Mandera (2015) pointed out that the same type of indirect vocabulary likely explains why Dutch speakers’ L1 vocabulary increases with the number of L2s known (for a similar finding with Spanish L1 speakers, see also Aguasvivas et al., 2020). More in general, this suggests that a part of a multilingual’s vocabulary can be considered as being ‘pan-language’.…”

Words in the CLOUD: How orthographic similarity and bilingual experience facilitate foreign vocabulary learning

“…Notably, the vocabulary size used in the present modeling work represents an almost ten-fold input size increase from previously reported models, containing around 1,700 words at most (Li & Grant, 2019). Our choice of vocabulary aligns with a real-world scenario considering that the average young adult controls more than 25,000 words (Aguasvivas et al, 2020;Brysbaert et al, 2016).…”

“…Empirical evidence from a Spanish lexical decision megastudy suggests that monolingual and bilingual participants do not differ in their LDT accuracy (Aguasvivas et al, 2020). Therefore, the different versions and runs of the CLOUD models needed to have comparable performance on this adapted Spanish LDT.…”

Effects of Print Exposure on an Online Lexical Decision Task: A Direct Replication Using a Web-Based Experimental Procedure

“…In the case of recruiting participants from heterogeneous groups using a crowdsourcing method, concerns have been raised that the participants will be easily distracted and the data quality may be poor (Woods et al, 2015 ). However, recent web-based crowdsourced experiments have successfully replicated standard lexical processing effects, as long as outliers are removed and other data-quality practices are adopted (Aguasvivas et al, 2020 ; Mandera et al, 2020 ). This all suggests that reliable data can be obtained using less-controlled web-based platforms, and further studies should be conducted using this approach.…”