20,544 rated or chosen profiles from 2,078 respondents, in up to 5 tasks of 2 profiles varying on 2 attributes.
| Description | Taiwanese respondents (Study 1, n=2,078 with an answer) compared 5 pairs of ballot designs drawn from 7 (Chinese only, or Chinese with Paiwan and/or Vietnamese, stacked or in separate columns) and picked which ballot of each pair was more fair, more democratic and more efficient. |
|---|---|
| Reference | Liu, A. H., Selsky, S., Xu, M., & Yew, J. (2026). When bilingual ballot designs promote or undermine inclusivity: Evidence from three studies. Public Opinion Quarterly. |
| DOI | 10.1093/poq/nfag048 |
| Licence | CC0 1.0 |
| Source data | https://doi.org/10.7910/DVN/YULUZ1 |
These notes describe the source data, which the IRW reproduces as released. All notes are listed on the Data Notes page.
| Respondents | 2,078 |
|---|---|
| Rows (one per respondent, task and profile) | 20,544 |
| Tasks per respondent (most) | 5 |
| Profiles per task | 2 |
| Attributes | 2 |
| Topic | democracy institutions |
|---|---|
| A profile represents | message |
| Country | TW |
| Language shown | zh, pwn, vi |
| Language of stored levels | en |
| Randomization restrictions | yes (stated by the source): profiles are 7 fixed ballots: arrangement (Stacked/Separated) only for multilingual ballots; 'Chinese only' always '(not shown)' arrangement; pairs are two different ballots (all 42 ordered pairs occur) |
| Task order | recorded in the deposit |
| Profile position | recorded in the deposit |
Because the attribute levels were not randomized independently, an estimator has to account for the restriction; a plain difference in means across levels can mislead.
Outcomes:
| choice_democratic | Which of the two ballots is more democratic? (paraphrase) [choice, forced] |
|---|---|
| choice_efficient | Which of the two ballots is more efficient? (paraphrase) [choice, forced] |
| choice_fair | Which of the two ballots is more fair? (paraphrase) [choice, forced] |
attr_arrangementattr_languagechoice_democraticchoice_efficientchoice_faircov_gendercov_regionidprofiletask
The layout (one row per respondent, task and profile; choice/rating outcomes; attr_ columns holding each attribute level as displayed) is defined in the IRW conjoint standard. What the outcomes asked is under Design above.
Or load it directly in R or Python (both need a free Redivis account; the first call opens a browser to sign in):
# R
install.packages("remotes")
remotes::install_github("itemresponsewarehouse/Rpkg")
library(irw)
df <- irw_fetch("liu_2026_taiwan_ballots", source = "conj")
# Python
pip install irw
import irw
df = irw.fetch("liu_2026_taiwan_ballots", source="conj")
@article{Liu_2026, title={When Bilingual Ballot Designs Promote or Undermine Inclusivity: Evidence from Three Studies}, volume={90}, ISSN={1537-5331}, url={http://dx.doi.org/10.1093/poq/nfag048}, DOI={10.1093/poq/nfag048}, number={SI}, journal={Public Opinion Quarterly}, publisher={Oxford University Press (OUP)}, author={Liu, Amy H and Selsky, Sam and Xu, Meiying and Yew, Joel}, year={2026}, pages={1173–1219} }
| IRW version | v584 |
|---|---|
| Redivis dataset | irw_conjoint v4.2 |
| Redivis dataset DOI | 10.57761/qkye-q530 |
| Manifest pin for this IRW version | v4.2 |
| Metadata source | irw_meta v44.0 |