What do (some of) our association measures measure (most)? Association?

This paper discusses the degree to which some of the most widely-used measures of association in corpus linguistics are not particularly valid in the sense of actually measuring association rather than some amalgam of a lot of frequency and a little association. The paper demonstrates these issues on the basis of hypothetical and actual corpus data and outlines implications of the findings. I then outline how to design an association measure that only measures association and show that its behavior supports the use of the log odds ratio as a true association-only measure but separately from frequency; in addition, this paper sets the stage for an analogous review of dispersion measures in corpus linguistics.

The Effects of Two Association Measures on L2 Collocation Processing

The influence of association measures has been little examined in research on L2 collocation processing. For this reason, the present study replicated Öksüz et al. (2020) experiment on intermediate L2 learners of English to determine whether the association measure mutual information (MI) is a stronger predictor of L2 performance than the Log Dice measure. Twenty-two intermediate Arab learners of English completed a timed acceptability judgment task on the online Gorilla platform. The task included (1) high-frequent collocations (e.g., bad news), (2) low-frequent collocations (e.g., only friend), and (3) non-collocates (e.g., true news, wrong friend) which had differing MI and Log Dice scores. Mixed-effects models were built to analyze the participants’ reaction times to the three conditions. The results showed that the frequency of the collocation (operationalized as item type) and its length significantly influenced reaction times, while both MI and Log Dice scores did not surface as significant predictors. This suggests that intermediate English L2 learners are not sensitive to corpus-based association measures. The results have important implications for L2 teaching and testing and may indicate that it is not worthwhile to determine which collocations to include in the materials based mainly on the strength of the association.

Objective and subjective collocational frequency

Research states that when deciding what target lexical items to focus in their syllabi, textbook designers tend to use their intuition and establish their framework for vocabulary teaching accordingly rather than using a systematic method. With the advent of computational methods for language analysis, contemporary EFL publishers also claim they have begun to design corpus-informed textbooks reflecting actual language use in their products as evidenced in a representative corpus. In an attempt to offer a more rational method to form corpus-oriented and pedagogically convenient target vocabulary lists, the present study, exploiting different association measures in a representative corpus, seeks to detect the collocational strength of 50 target word combinations presented in the two EFL textbooks as an initial step. Additionally, inspired by Ellis et al.’s (2008) research, the current study aims to investigate if/to what extent the association measures indicating collocational strength correlate with EFL instructors’ intuitions regarding collocational frequency. The results indicate that EFL instructors’ collocational frequency intuitions correlate strongly with an objective collocational frequency measure (t-score). The findings are likely to guide decision makers in tertiary level schools in constructing their vocabulary syllabi and designing materials for teaching collocations in particular.

Regression models for interval censored data using parametric pseudo-observations

Background Time-to-event data that is subject to interval censoring is common in the practice of medical research and versatile statistical methods for estimating associations in such settings have been limited. For right censored data, non-parametric pseudo-observations have been proposed as a basis for regression modeling with the possibility to use different association measures. In this article, we propose a method for calculating pseudo-observations for interval censored data. Methods We develop an extension of a recently developed set of parametric pseudo-observations based on a spline-based flexible parametric estimator. The inherent competing risk issue with an interval censored event of interest necessitates the use of an illness-death model, and we formulate our method within this framework. To evaluate the empirical properties of the proposed method, we perform a simulation study and calculate pseudo-observations based on our method as well as alternative approaches. We also present an analysis of a real dataset on patients with implantable cardioverter-defibrillators who are monitored for the occurrence of a particular type of device failures by routine follow-up examinations. In this dataset, we have information on exact event times as well as the interval censored data, so we can compare analyses of pseudo-observations based on the interval censored data to those obtained using the non-parametric pseudo-observations for right censored data. Results Our simulations show that the proposed method for calculating pseudo-observations provides unbiased estimates of the cumulative incidence function as well as associations with exposure variables with appropriate coverage probabilities. The analysis of the real dataset also suggests that our method provides estimates which are in agreement with estimates obtained from the right censored data. Conclusions The proposed method for calculating pseudo-observations based on the flexible parametric approach provides a versatile solution to the specific challenges that arise with interval censored data. This solution allows regression modeling using a range of different association measures.

Collocational Development during a Stay Abroad

The purpose of the current study was to explore if and how additional-language learners may show changes in phraseological patterns over the course of a stay in a target-language environment. In particular, we focused on noun+adjective combinations produced by a group of additional-language speakers of French at three points in time, spanning 21 months and including an academic year in France. We extracted each combination from a longitudinal corpus and determined frequency counts and two strength-of-association measures (Mutual information [MI] score and Log Dice) for each combination. Separate analyses were conducted for frequency and the strength-of-association measures, revealing that phraseological patterns are significantly predicted by adjective position in the case of all three measures, and that MI scores showed significant change over time. We interpret the results in light of past research that has reported contradictory findings concerning change in phraseological patterns following an immersion experience.

Словацкие наречия группы «часто» на фоне венгерских по данным мер ассоциации

In the report, the most frequent Slovak and Hungarian temporal adverbs with the meaning of high repeatability of the situation are considered on the corpus material. The compatibility of lexemes is analyzed according to the indicators of association measures, on this basis, the main trends characteristic of the structuring of the semantic group in each of the two spatially bordering languages are identified. It is shown that the Slovak language, more than Hungarian, tends to semanticize lexemes within a group, as well as to mark non-standard, according to the speaker, situations.