LATIC–A linguistic analyzer for text and item characteristics

Abstract

Analyzing texts and items regarding their linguistic features might be important for researchers to investigate the effects of the linguistic presentation as well as for practitioners to estimate the readability of a text or an item. The Linguistic Analyzer for Text and Item Characteristics (LATIC) is a software that enables users to analyze texts and items more efficiently. LATIC offers a multitude of features at three different reading levels and can be used for texts and items in four different languages: English, French, German, and Spanish. It is open source, free to use and designed to be user-friendly. In this study, we investigated LATIC’s performance: LATIC achieves highly accurate results, while being extremely time saving compared to human raters. While developing LATIC, the respective features are tested continuously to ensure a high accuracy of results in the future.

Lingualyzer: A computational linguistic tool for multilingual and multidimensional text analysis

“…It is, however, important to note that tools can be used for multiple purposes. For example, T-Scan (Pander Maat et al, 2014) and LATIC (Cruz Neri et al, 2022b) have been designed to quantify both text characteristics and complexity, while Coh-Metrix, for example has been used to characterize variation in language registers (Louwerse et al, 2004) and authorship attribution (McCarthy et al, 2006).…”

“…Most tools only support English, and some are even more specific, focusing only on written English (Crossley et al, 2016) or even written English as a second language (Kyle, 2016;Lu, 2010). There are, however, some tools that cover more than just the English language (Bengoetxea & Gonzales-Dios, 2021;Brunato et al, 2020;Cruz Neri et al, 2022b;Dascalu et al, 2013). These tools may support one other language, but with the exception of Profiling-UD (Brunato et al, 2020), the languages beyond English are only supported by a subset of the measures.…”

“…This means that all measures are available in all languages. Consequently, all values are calculated using exactly the same computations, and annotations are performed based on the same strategies and schemes, which is not the case for most other multilingual quantitative text analysis tools, such as ReaderBench (Dascalu et al, 2013), MultiAzterTest (Bengoetxea & Gonzales-Dios, 2021) and LATIC (Cruz Neri et al, 2022b). This means that the output of all the measures are comparable across languages.…”

“…Part-of-speech annotations in LATIC were manually verified in English and German through human annotations on a small sample from corpora containing fiction and news articles respectively (Cruz Neri et al, 2022b). Next, using five short texts, taken from introductory texts of questions from a science assessment, the measures were calculated by human annotators and correlations between the annotators and the output from LATIC were computed.…”

The effect of word transpositions on grammaticality judgements in first and second language sentence reading

“…The average SUBTLEX Zipf word frequency (van Heuven et al, 2014) was M = 5.5 ( SD = 1.3) in the English version of the sentences, and the SUBTLEX log frequency (Brysbaert et al, 2011) was M = 3.9 ( SD = 1.2) in the German version. Readability scores were generated for the sentences using the LATIC application (Neri et al, 2022) indicating moderate difficulty for the English (Flesch score = 75.1) and German (Flesch score = 77.3) sentences.…”