R1.13 - results: is prof as continuous new

[jvcasillas]

In the discussion the authors highlight the fact that proficiency was treated as a continuous variable, and the reader would benefit from knowing whether this is actually an innovation of the present study or if, instead, other studies have done it before (now it seems that it is an innovation).

Action: respond to reviewer

[kparrish92]

Proficiency as a continuous variable is not an innovation of our study, and has been done in several studies to date (See X, Y, and Z). The reason that proficiency may not have been a continuous variable in other previous studies is the choice of analysis, such as an analysis of variance (ANOVA). This analysis requires a continuous outcome variable to be predicted by factors (categorical variables). We use multiple (Bayesian) regression, which allows for both categorical and continuous predictor variables. In our case, we were interested in including proficiency as a predictor in order to hold in constant relative to empathy. In other words, we included both proficiency and empathy to ensure that the effect of empathy is not confounded with proficiency.

A note for @jvcasillas: Should we add any references here? There seem to be quite a few studies that use proficiency as a continuous variable, but I am not sure which ones I can justify picking over others for the sake of the reviewer.

[jvcasillas]

A note for @jvcasillas: Should we add any references here? There seem to be quite a few studies that use proficiency as a continuous variable, but I am not sure which ones I can justify picking over others for the sake of the reviewer.

Yeah, I think adding a couple (just in the review letter) is fine. We can cite the SSLA study from the lab, and any of yours that use LexTALE.

[jvcasillas]

I'm going to close this after editing your paragraph. Thanks.

[jvcasillas]

I simplified this. What you wrote is fine, but probably more than needed.

Addressed via https://github.com/RAP-group/empathy_intonation_perc/pull/69

The use of proficiency as a continuous variable is not an innovation of our study. The reason continuous variables are discretized in other research likely results from the choice of analysis, e.g., analysis of variance (ANOVA), which requires a continuous outcome variable be predicted by factors (categorical variables). While we do believe that the common practice of discretization of continuous variables is less than ideal, it is not our intention to speak to this topic in the manuscript. The description regarding how the variable is coded is included (a) for the sake of completeness and reproducibility, and (b) to highlight our departure from Brandl et al, which our study replicates conceptually.