Breno B. Silva

Breno B. Silva

Auto-generated profile, based on publicly available information

Breno B. Silva

University of Warsaw

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 June 2022 in Corpus-Based Research in Legal and Institutional Translation
Verified
Journal Article
Closed access
BSBreno B. Silva
Breno B. Silva
DODavid Orrego-Carmona
David Orrego-Carmona
ASAgnieszka Szarkowska
Agnieszka Szarkowska
In this paper, we aim to promote the use of linear mixed models (LMMs) in eye-tracking research on subtitling. Using eye tracking to study viewers’ reading of subtitles often warrants controlling for many confounding variables. However, even assuming that these variables are known to researchers, such control may not be possible or desired. Traditional statistical methods such ast-tests or ANOVAs exacerbate the problem due to the use of aggregated data: each participant has one data point per dependent variable. As a solution, we propose the use of LMMs, which are better suited to account for a number of subtitle and participant characteristics, thus explaining more variance. We introduce essential theoretical aspects of LMMs and highlight some of their advantages over traditional statistical methods. To illustrate our point, we compare two analyses of the same dataset: one using at-test; another using LMMs.

13 May 2017 in International Review of Applied Linguistics in Language Teaching
Verified
Journal Article
Closed access
PDF file
BSBreno Silva
Breno Silva
AOAgnieszka Otwinowska-Kasztelanic
Agnieszka Otwinowska-Kasztelanic
Laufer and Hulstijn (2001) suggest that the motivational-cognitive construct ofinvolvementmay explain and predict different levels of effectiveness for vocabulary-learning tasks. Drawing on their original work and on later research on theinvolvement load hypothesis(ILH), this study set out to compare the effectiveness of carefully-designed tasks for incidental vocabulary acquisition in children. Thirty-eight EFL elementary-level 10-year-olds from a public school in Warsaw, Poland, participated in the experiment. Divided into three groups, the participants performed three different sequences of tasks, each sequence inducing similar levels of involvement load. In order to measure receptive lexical learning and retention of meaning and spelling, the children were tested with an orthography test, an L2-L1 translation test, and a multiple-choice test immediately after the treatment, and one week later. In support of the ILH, the MANOVA results showed no significant differences between the treatments (irrespective of them being input- or output-based) in any of the test measurements, either in the immediate or in the delayed posttest. We discuss the results in light of the ILH, and outline some limitations and possible implications for pedagogy.

26 April 2024 in Journal of Audiovisual Translation
Verified
Journal Article
Open access
PDF file
ASAgnieszka Szarkowska
Agnieszka Szarkowska
VRValentina Ragni
Valentina Ragni
DODavid Orrego-Carmona
David Orrego-Carmona
SBSharon Black
Sharon Black
SSSonia Szkriba
Sonia Szkriba
BSBreno Silva
Breno Silva
We present results of a direct replication of Liao et al.’s (2021) study on how subtitle speed and the presence of concurrent video impact subtitle reading among British and Polish viewers. Our goal was to assess the generalisability of the original study’s findings on a cohort of Australian English. The study explored both subtitle-level and word-level effects, considering the presence or absence of concurrent video and three subtitle speeds: 12 characters per second, 20 cps, and 28 cps. Overall, most of the original results were replicated, confirming that the presence of video and the speed of the subtitles have a measurable impact on processing across different viewer groups. Additionally, differences in how native and non-native speakers process subtitles emerged, in particular related to wrap-up, word frequency and word length effects. The paper describes the replication in detail, presents the findings, and discusses some of their implications. Lay summary In our study we were interested in the effects that the presence of video and various subtitle speeds have on how viewers watch subtitled videos and how they understand them. We also wanted to know if the previous results obtained in a study by Liao et al. (2021) in Australia hold true for other viewers living in different locations. With this goal in mind, we repeated Liao et al.’s (2021) study on British and Polish viewers. The study explored both subtitle-level and word-level effects, considering the presence or absence of video and three subtitle speeds: 12 characters per second, 20 cps, and 28 cps. Overall, most of the original results were confirmed, showing that the presence of video and the speed of the subtitles have an impact on processing across different viewer groups. Additionally, differences in how native and non-native speakers process subtitles emerged, in particular related to well-known linguistic effects from reading studies, such as wrap-up, word frequency and word length effects. The paper describes the replication in detail, presents the findings, and discusses some of their implications.

15 June 2021 in Language Teaching Research
Verified
Journal Article
Closed access
PDF file
BSBreno B. Silva
Breno B. Silva
KKKatarzyna Kutyłowska
Katarzyna Kutyłowska
AOAgnieszka Otwinowska-Kasztelanic
Agnieszka Otwinowska-Kasztelanic
The involvement load hypothesis (ILH), which predicts the lexical learning potential of tasks, assumes that writing sentences (SW) and compositions (CW) using novel target words (TWs) lead to similar lexical gains. However, research on the issue is scarce and contradictory. One possibility is that the higher cognitive load of CW hinders learning relative to SW. To verify the learning potential of SW and CW, we selected 20 English academic TWs and conducted a pretest–posttest quasi-experiment with Polish advanced learners of English. First, all participants wrote a control essay (without TWs), then SW participants wrote sentences and CW participants wrote two essays, each with 10 TWs. Generalized linear mixed models revealed higher gains in breadth and depth of knowledge for SW than for CW, which contradicts the predictions of the ILH. Furthermore, to detect signs of cognitive load, we derived three task-based performance measurements from the compositions: holistic scores, number of errors, and words per minute. The measurements found that the control essay and essays with TWs were of similar quality (holistic scores), but that the control essay was written faster and with fewer errors than the other two. Concluding, using TWs in essays probably increased learners’ cognitive load, slowing down their writing, generating more errors, and ultimately, decreasing learning of the TWs.

1 April 2019 in English for Specific Purposes
Verified
Journal Article
Closed access
BSBreno B. Silva
Breno B. Silva
AOAgnieszka Otwinowska-Kasztelanic
Agnieszka Otwinowska-Kasztelanic

3 December 2021
Verified
Journal Article
Closed access
ASAgnieszka Szarkowska
Agnieszka Szarkowska
BSBreno Silva
Breno Silva
DODavid Orrego-Carmona
David Orrego-Carmona
How much time do viewers spend reading subtitles and does it depend on the subtitle speed? By posing these questions, in this paper we re-analyse previous data to address this issue while promoting two methodological advancements in eye-tracking audiovisual research: (1) the use of proportional reading time (PRT) as a metric of time spent on subtitle reading and (2) the analysis of data via linear mixed models (LMMs). We tested 19 Polish L1 viewers with advanced English proficiency watching two clips with English soundtrack with Polish subtitles. First, we compared PRT at two different subtitle speeds: 12 characters per second (cps) and 20 cps. Then, we used actual subtitle speed rates to better understand the speed-PRT relationship. The results showed a significantly higher PRT for 20 cps compared to 12 cps, with the models predicting a PRT of 45.24% at 20 cps. We have also found strong evidence of the advantage of LMMs over more commonly used statistical techniques.