introduction.html

<h4>In our paper, <a href="https://doi.org/10.3758/s13428-022-01814-7">Rating Norms Should be Calculated from Cumulative Link Mixed Effects Models</a>, we argue that when presenting or sharing item norms of Likert ratings, researchers should account for the ordinal nature of the rating scales, rather than just reporting raw means and SDs. Fitting ordinal models like Cumulative Link Mixed Effects Models (CLMMs) will allow researchers to norm items without assuming the Likert scale is continuous.</h4>
<br></br>
<h4>This <a href="https://shiny.rstudio.com/">Shiny app</a> is an interactive interface for fitting CLMMs to rating data. The app fits CLMMs via the <a href="https://cran.r-project.org/web/packages/ordinal/index.html">ordinal</a> package for R, to estimate the shift in location, of a latent distribution, which is associated with each item and participant in a dataset (i.e., latent means).</h4>
<br></br>
<h4>The app is designed to be suited to the most common setup for a norming experiment, where multiple participants rate multiple items. The models implemented in the app also only estimate differences in the central tendency of the latent distribution, and not in other features of the latent distribution. There may be good reason to include additional fixed or random effects, or to also model changes in the (e.g.) spread of the latent distribution. As a result, if sharing the results from this app, we also recommend sharing the raw, trial-level dataset.</h4>