


Paul Joseph Vitta | Last modified date:2023.10.02 |

Associate Professor /
Faculty of Languages and Cultures
Department of Linguistic Environment
Faculty of Languages and Cultures
Department of Linguistic Environment
Faculty of Languages and Cultures
Administration Post
Other
E-Mail *Since the e-mail address is not displayed in Internet Explorer, please use another web browser:Google Chrome, safari.
Homepage
https://kyushu-u.elsevierpure.com/en/persons/joseph-vitta
Reseacher Profiling Tool Kyushu University Pure
Fax
092-1111-2222
Academic Degree
Doctorate - Education/TESOL - Queen's University Belfast - UK - 2019, Masters - TESOL - Sookmyung Women's University - Korea -2012, BA - Sociology & Anthropology - Washington & Lee - USA - 2003
Country of degree conferring institution (Overseas)
Yes Bachelor Master Doctor
Field of Specialization
TESOL, L2 Research synthesis, linguistic complexity, instructed vocabulary (L2), EAP curricula design
ORCID(Open Researcher and Contributor ID)
0000-0002-5711-969X
Total Priod of education and research career in the foreign country
07years00months
Outline Activities
--teach EAP courses in the Q-LEAP3 English Program within University's Kikan Scheme
--research activity here: https://orcid.org/0000-0002-5711-969X
--ad hoc reviewer for 20 SLA/AL/TESOL journals
--*Associate Editor*
Vocabulary Learning and Instruction
*Associate Editor*
International Journal of TESOL Studies
*Editorial Board*
Studies in Educational Evaluation
--research activity here: https://orcid.org/0000-0002-5711-969X
--ad hoc reviewer for 20 SLA/AL/TESOL journals
--*Associate Editor*
Vocabulary Learning and Instruction
*Associate Editor*
International Journal of TESOL Studies
*Editorial Board*
Studies in Educational Evaluation
Research
Research Interests
Membership in Academic Society
- I am interested in second language acquisition, applied linguistics, and TESOL. My research often lies at the intersection of practice and theory and I specialize in research synthesis and quantitative methods in social science.
keyword : TESOL
2012.06~2031.06.
Papers
Presentations
1. | Pablo Robles-García, Christopher Nicklin, Joseph P. Vitta, Jeffrey Stewart, Exploring Teacher Judgements as a Predictor of Students’ Vocabulary Knowledge, American Association for Applied Linguistics (AAAL), 2023.03, [URL], For the last few decades, word-frequency has been widely used to identify which words L2 learners are more or less likely to know (Hashimoto, 2021). However, research indicates that teachers often prefer to rely on their own intuition rather than using corpus-based vocabulary lists for making decisions about the words they want to teach in the classroom (Dang & Webb, 2020; Sánchez-Gutiérrez et al., 2022). Although teacher judgments are a commonly used strategy for vocabulary selection in the L2 classroom, little is known about the accuracy of such judgments when predicting L2 learners’ vocabulary knowledge. This study investigated the effectiveness of word-frequency and teacher judgments in determining students’ vocabulary knowledge and compared the predictive powers of both approaches when estimating word difficulty. Twenty-nine L2 Spanish teachers were asked to predict how likely their students would know words from 3K-LEx (Robles-García, 2020), a 216-word Yes/No test that measures knowledge of the first 3,000 words in Spanish. The accuracy of their responses was compared with the 3K-LEx results of 1,075 L2 Spanish learners. To examine if the results could apply to other L2 settings, 15 L2 English language instructors and 394 L2 English students completed a 70-word Yes/No test measuring knowledge of the first 14,000 words in English. Results showed that for both language contexts, (1) the median teacher rater could assess difficulty with an accuracy roughly comparable to frequency, (2) the combination of teachers’ judgments (minimum of three teachers as determined via bootstrapping) displayed a stronger relationship with word difficulty than frequency, and (3) using teacher judgments and frequency together in multiple regressions did not substantially improve the prediction of word difficulty compared to models with teacher judgments as the lone predictor. These findings suggest the need to develop vocabulary lists that acknowledge teachers’ judgments as a major source of information.. |
2. | Joseph P. Vitta, Aaron Hahn, Christopher Nicklin, Exploring the Sampling Crisis in L2 Quantitative Research: a Predictive Model and Future Directions, American Association for Applied Linguistics (AAAL), 2023.03, [URL], This paper considers the extent to which L2 subfield, author collaboration, and bibliometric and citation analysis metrics predict the observed frequencies of power analysis and multi-site sample use in L2 quantitative research. Recent L2 sampling guidance has called for power analysis derived sample sizes and multi-site samples as a compromise between ideal but impractical randomized sampling and biased single-site samples (Vitta et al., 2021). The predictors were chosen by referencing L2 sampling literature (Morgan-Short et al., 2018) and past research predicting study quality (Al-Hoorie & Vitta, 2019). To investigate L2 sampling practices, 230 papers featuring L2 research and utilizing inferential testing published in 2020 by 46 international journals were coded. Unlike some past sampling reviews that were bound to a handful of journals and/or subfield, the report pool was created to capture a wide-breadth of L2 quantitative research. The search was confined to 2020 as bibliometric and citation analysis metrics were intended as predictors and such metrics are published with a ‘look back’ approach. The results highlighted that the field still overlooks power analysis derived samples with only six of 230 (2.61%) reports considering power in any fashion and thus predictive models involving power as the response variable were omitted. Furthermore, just 36.95% of the reports (k = 85) featured multi-site samples. During the bivariate screening process, bibliometric and citation analysis metrics (e.g., CiteScore) were discounted as significant predictors of multi-site sampling practices. Multivariate generalized modeling (pseudo R2 = .28) demonstrated that instructed SLA-focused reports (OR = .11) and single-authored reports (OR = .36) had a 89% and 64% probability, respectively, of being single-site papers. A sensitivity GLMM analysis was conducted and the clustering (random) effect of journals was found to be inconsequential. Finally, the authors will discuss how to apply the findings to reform future L2 sampling and research practices.. |
3. | Joseph P. Vitta, BAAL Researcher Development Workshop Series: Professional Development in Applied Linguistics for Graduate Students and Early Career Researchers -- Navigating the academic job market, British Association for Applied Linguistics, 2022.03, [URL], Workshop objectives This series aims to serve as a platform to instigate dialogues between doctoral students/early career researchers (ECRs) with senior academics in the field of Applied Linguistics on various important topics relevant to professional development in academia. Specifically, through hosting semi-formal online workshops and informal discussion forums for doctoral students and ECRs, and launching a researcher development website, the organiser aims to: Demystify areas related to professional development in Applied Linguistics which are often hidden and misconceived through experience sharing by senior academics; Provide opportunities for doctoral students/ECRs to network with each other to share information and their own experiences related to different facets of professional development in Applied Linguistics.. |
4. | Joseph P. Vitta, Christopher Nicklin, Simon W. Albright, Academic word difficulty and multidimensional lexical sophistication: A multi-site replication of Hashimoto & Egbert (2019), American Association for Applied Linguistics (AAAL), 2022.03, [URL], Hashimoto and Egbert (2019) demonstrated that multivariate models featuring lexical sophistication predictor predict L2 words difficulty best. Their ‘more than frequency’ conclusion contributed to the ongoing debate regarding the extent to which frequency should be considered the “the single most important characteristic of lexis” (Schmitt, 2010, p. 63). This current study (nwords = 91; npeople = 171) conceptually replicated Hashimoto and Egbert with data from three Asian University EAP sites. This conceptual replication featured two main departures from the original study’s methodology. First, the target words came from Coxhead’s Academic Word List (2000). Second, an alternative testing approach was undertaken featuring a theory-driven selection of predictors and the avoidance of stepwise regression, which applied statistics literature has challenged (Smith, 2018). Like the original study, the replication’s findings favored multivariate models. In the final model, frequency was replaced by range and the predictor accounted for approximately 21% (observed Pratt value) of the total variance predicted in the multiple regression model (R2 = .60). Age of acquisition (AoA: 18%) and word naming reaction time (WN_RT: 15%) were the other significant lexical sophistication predictors. Extending from the original study’s use of fixed-effect modeling, the word difficulty predicted by lexical sophistication hypothesis was addressed with a generalized mixed-effects model, where person and site intercepts were random effects and false alarm rate was added as a fixed-effect covariate. As expected, the model’s ability to predict word difficulty fell (conditional R2 = .45), but AoA, range, and WN_RT remained the strongest predictors with all other predictors having less of an association with the DV than false alarms. Our study’s main implications are that multivariate lexical sophistication models appear ideal for predicting word difficulty across functional domains, and that false alarm rate might be an important covariate in word difficulty research, although it has recently been overlooked (e.g., Hashimoto, 2021).. |
5. | Joseph P. Vitta, Christopher Nicklin, Improving L2 instructed vocabulary experimental designs’ samples: A methodological synthesis with recommendations, American Association for Applied Linguistics (AAAL), 2021.03, In this presentation, we will present the results of a methodological synthesis investigating the state of second language instructed vocabulary (L2 IV) research sampling procedures. A sample of 82 (quasi-)experimental IV reports were systematically collected from five SSCI journals and then synthesized over a two-stage process with the aim of improving future L2 IV research and enhancing its usefulness to researchers and teachers alike. Phase I entailed an analysis of the reports’ sample design choices. The results revealed that none of the reports conducted a priori sample-size planning nor random sampling, and approximately 83% of the studies drew their samples from a single education setting, which potentially affects the generalizability of the results (Morgan-Short et al., 2018). Of the 14 studies employing multi-site samples, none empirically checked for possible clustering effects of the different locations (Al-Hoorie & Vitta, 2019). Half (k = 41) of the reports included random assignment at either the class or participant level. Based on these findings, we suggest more widespread collaboration among colleagues and institutions. We also encourage researchers to engage in random/probability sampling with demographic considerations that are representative of meaningful populations. This addresses the observed over-reliance on smaller samples from a single setting. Additionally, we highlight how multivariate testing can parsimoniously consider random/clustering effects in a sample. Phase I’s finding regarding sample-size planning was addressed in Phase II.The effect sizes observed in our sample were aggregated according to group comparison, as such summative effect sizes are required for sample-size planning. Utilizing effect sizes (between-subject gs = .62 [medium], and .33 [small]; within-subject counterbalancing gav = .25) that operationalized meaningful IV hypotheses (i.e., which teaching method is more effective), power simulations are presented to provide suggestions on the required sample sizes needed to detect the aggregated effects that we observed.. |
- Korea TESOL
- American Association for Applied Linguistics
- Asia TEFL - Life time Member
- JALT
Educational
Social


Unauthorized reprint of the contents of this database is prohibited.
