17,380 rated or chosen profiles from 1,750 respondents, in up to 5 tasks of 2 profiles varying on 8 attributes.
| Description | Paired conjoint on recidivism-forecasting algorithms: 1,750 US online respondents (MTurk March 2018; Lucid March and April 2019) compared 5 pairs of algorithms varying 8 features (developer, accuracy, data size, human involvement, factor weights, data location, number of factors, error), choosing one and rating each 1-7 on willingness to use. |
|---|---|
| Reference | Kennedy, R. P., Waggoner, P. D., & Ward, M. M. (2022). Trust in public policy algorithms. The Journal of Politics, 84(2), 1132-1148. |
| DOI | 10.1086/716283 |
| Licence | CC0 1.0 |
| Source data | https://doi.org/10.7910/DVN/8JFNSN |
These notes describe the source data, which the IRW reproduces as released. All notes are listed on the Data Notes page.
| Respondents | 1,750 |
|---|---|
| Rows (one per respondent, task and profile) | 17,380 |
| Tasks per respondent (most) | 5 |
| Profiles per task | 2 |
| Attributes | 8 |
| Tasks with no profile chosen | 0 |
| Topic | technology ai privacy, crime justice policing |
|---|---|
| A profile represents | product service |
| Country | US |
| Language shown | en |
| Language of stored levels | en |
| Randomization restrictions | unknown |
| Task order | recorded in the deposit |
| Profile position | recorded in the deposit |
Outcomes:
| choice | Which of the two algorithms (Choice 1, Choice 2) do you choose? (paraphrase) [choice, forced] |
|---|---|
| rating | Rating of each algorithm profile on whether the respondent would use it (paraphrase) [rating, 1 (Definitely Would Not Use) to 7 (Definitely Would Use)] |
attr_accuracyattr_data_locationattr_developerattr_factor_weightsattr_human_involvementattr_margin_of_errorattr_n_casesattr_n_factorsattrpos_accuracyattrpos_data_locationattrpos_developerattrpos_factor_weightsattrpos_human_involvementattrpos_margin_of_errorattrpos_n_casesattrpos_n_factorschoicecov_agecov_educationcov_gendercov_party_idcov_party_leancov_party_strengthcov_sampleidprofileratingtask
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("kennedy_2022_policy_algorithms", source = "conj")
# Python
pip install irw
import irw
df = irw.fetch("kennedy_2022_policy_algorithms", source="conj")
@article{Kennedy_2022, title={Trust in Public Policy Algorithms}, volume={84}, ISSN={1468-2508}, url={http://dx.doi.org/10.1086/716283}, DOI={10.1086/716283}, number={2}, journal={The Journal of Politics}, publisher={University of Chicago Press}, author={Kennedy, Ryan P. and Waggoner, Philip D. and Ward, Matthew M.}, year={2022}, month=Apr, pages={1132–1148} }
| 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 |