@inproceedings{correa-etal-2026-short,
title = "When Short Responses Are Hard to Distinguish in Translation Audits",
author = "Correa, Jo{\~a}o Victor M. and
Shiozawa, Pedro and
Ferreira, Victor Hierro Moraes and
Marinho, Emanoel Cristhian Le{\~a}o",
editor = "Wang, Zimu and
Jiang, Yiwen and
Shen, Yaling and
Zhao, Xiangyu and
Fong, Stephanie and
Sun, Xin and
Zhou, Yi and
Ji, Shaoxiong and
Desimone, Robert and
Nelson, Barnaby and
Dwyer, Dominic and
Ge, Zongyuan",
booktitle = "Proceedings of the 1st Workshop on Multimodal, Multilingual, and Multicultural Mental Health and Psychotherapy ({M}ulti{P}syche)",
month = nov,
year = "2026",
address = "Hengqin, Zhuhai, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.multipsyche-1.2/",
pages = "15--25",
ISBN = "979-8-89176-448-4",
abstract = "Reference-free translation audits have little to work with when a survey answer is only one or two words long. A translation may be reasonable and still be nearly impossible to reconnect to its source among thousands of similar answers. We examined this problem in 44,118 Portuguese{--}English response pairs from 14,706 survey-specific respondent records across four workplace surveys. One third of the pairs addressed mental health or well-being. Our analysis kept duplicate-aware retrieval, COMETKiwi quality estimates, language identification, and affective-lexicon coverage separate because they answer different questions. The retrieval encoder found the paired target or an exact normalized duplicate at rank one for 71.80{\%} of English queries and 73.37{\%} of Portuguese queries overall. Response length changed the picture. Rank-one recovery ranged from 8.39{--}12.96{\%} for one- or two-word answers to 98.45{--}99.08{\%} for answers longer than 20 words. COMETKiwi scores did not rise steadily with length, and about one quarter of responses in either language contained no emotion term covered by the lexicons. A failed match therefore does not, by itself, identify a poor translation. For brief and repetitive survey text, reference-free audits should report how distinguishable the response is and how much of it the diagnostic resources cover."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="correa-etal-2026-short">
<titleInfo>
<title>When Short Responses Are Hard to Distinguish in Translation Audits</title>
</titleInfo>
<name type="personal">
<namePart type="given">João</namePart>
<namePart type="given">Victor</namePart>
<namePart type="given">M</namePart>
<namePart type="family">Correa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pedro</namePart>
<namePart type="family">Shiozawa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Victor</namePart>
<namePart type="given">Hierro</namePart>
<namePart type="given">Moraes</namePart>
<namePart type="family">Ferreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Emanoel</namePart>
<namePart type="given">Cristhian</namePart>
<namePart type="given">Leão</namePart>
<namePart type="family">Marinho</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 1st Workshop on Multimodal, Multilingual, and Multicultural Mental Health and Psychotherapy (MultiPsyche)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zimu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yiwen</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yaling</namePart>
<namePart type="family">Shen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiangyu</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Stephanie</namePart>
<namePart type="family">Fong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xin</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shaoxiong</namePart>
<namePart type="family">Ji</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Robert</namePart>
<namePart type="family">Desimone</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Barnaby</namePart>
<namePart type="family">Nelson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dominic</namePart>
<namePart type="family">Dwyer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zongyuan</namePart>
<namePart type="family">Ge</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, Zhuhai, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-448-4</identifier>
</relatedItem>
<abstract>Reference-free translation audits have little to work with when a survey answer is only one or two words long. A translation may be reasonable and still be nearly impossible to reconnect to its source among thousands of similar answers. We examined this problem in 44,118 Portuguese–English response pairs from 14,706 survey-specific respondent records across four workplace surveys. One third of the pairs addressed mental health or well-being. Our analysis kept duplicate-aware retrieval, COMETKiwi quality estimates, language identification, and affective-lexicon coverage separate because they answer different questions. The retrieval encoder found the paired target or an exact normalized duplicate at rank one for 71.80% of English queries and 73.37% of Portuguese queries overall. Response length changed the picture. Rank-one recovery ranged from 8.39–12.96% for one- or two-word answers to 98.45–99.08% for answers longer than 20 words. COMETKiwi scores did not rise steadily with length, and about one quarter of responses in either language contained no emotion term covered by the lexicons. A failed match therefore does not, by itself, identify a poor translation. For brief and repetitive survey text, reference-free audits should report how distinguishable the response is and how much of it the diagnostic resources cover.</abstract>
<identifier type="citekey">correa-etal-2026-short</identifier>
<location>
<url>https://aclanthology.org/2026.multipsyche-1.2/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>15</start>
<end>25</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T When Short Responses Are Hard to Distinguish in Translation Audits
%A Correa, João Victor M.
%A Shiozawa, Pedro
%A Ferreira, Victor Hierro Moraes
%A Marinho, Emanoel Cristhian Leão
%Y Wang, Zimu
%Y Jiang, Yiwen
%Y Shen, Yaling
%Y Zhao, Xiangyu
%Y Fong, Stephanie
%Y Sun, Xin
%Y Zhou, Yi
%Y Ji, Shaoxiong
%Y Desimone, Robert
%Y Nelson, Barnaby
%Y Dwyer, Dominic
%Y Ge, Zongyuan
%S Proceedings of the 1st Workshop on Multimodal, Multilingual, and Multicultural Mental Health and Psychotherapy (MultiPsyche)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, Zhuhai, China
%@ 979-8-89176-448-4
%F correa-etal-2026-short
%X Reference-free translation audits have little to work with when a survey answer is only one or two words long. A translation may be reasonable and still be nearly impossible to reconnect to its source among thousands of similar answers. We examined this problem in 44,118 Portuguese–English response pairs from 14,706 survey-specific respondent records across four workplace surveys. One third of the pairs addressed mental health or well-being. Our analysis kept duplicate-aware retrieval, COMETKiwi quality estimates, language identification, and affective-lexicon coverage separate because they answer different questions. The retrieval encoder found the paired target or an exact normalized duplicate at rank one for 71.80% of English queries and 73.37% of Portuguese queries overall. Response length changed the picture. Rank-one recovery ranged from 8.39–12.96% for one- or two-word answers to 98.45–99.08% for answers longer than 20 words. COMETKiwi scores did not rise steadily with length, and about one quarter of responses in either language contained no emotion term covered by the lexicons. A failed match therefore does not, by itself, identify a poor translation. For brief and repetitive survey text, reference-free audits should report how distinguishable the response is and how much of it the diagnostic resources cover.
%U https://aclanthology.org/2026.multipsyche-1.2/
%P 15-25
Markdown (Informal)
[When Short Responses Are Hard to Distinguish in Translation Audits](https://aclanthology.org/2026.multipsyche-1.2/) (Correa et al., MultiPsyche 2026)
ACL
- João Victor M. Correa, Pedro Shiozawa, Victor Hierro Moraes Ferreira, and Emanoel Cristhian Leão Marinho. 2026. When Short Responses Are Hard to Distinguish in Translation Audits. In Proceedings of the 1st Workshop on Multimodal, Multilingual, and Multicultural Mental Health and Psychotherapy (MultiPsyche), pages 15–25, Hengqin, Zhuhai, China. Association for Computational Linguistics.