Collinearity / analysis quality issue

[egouldo]

• [x] @parkerth to identify which analyses to drop
• [x] @egouldo BT forest plot, with dropped analyses that have substantial collinearity

[egouldo]

for posterity, Updates from THP (June 10, 2024):

Hi Elliot, This morning I did the following for the new meta-analyses (forest plot and heterogeneity estimates) excluding the blue tit analyses that included the two highly collinear predictors

1. A section in the 'Additional explanation' box in the Methods that describes the other post hoc subset meta-analyses
2. A line in the caption (mention this post hoc analysis) of the table where the new heterogeneity estimate will go (tbl-effects-heterogeneity)
3. A new subsection in the results explaining these post hoc results right after the results of the other post hoc analyses. In this final section, I added an Issue and assigned it to you, because I cite the new forest plot, but I don't actually know the citation for this new plot, so included the following placeholder: (@fig-forest-plot-?????) -Tim

[egouldo]

Update text caption for tbl-effects-params to mention collinearity subset, ensure that table includes collinear subset.

Update tbl-effects-heterogeneity caption to mention collinearity subset, and also ensure table includes the collinear subset.

[egouldo]

• [x] tbl-model-summary-stats-ratings-cont
• [x] model-summary-stats-ratings-cont
• [x] tbl-deviation-rating-estimates
• [x] interrater
  • [x] use filter_vars_main_analysis object to filter
  • [x] Ensure downstream code still executes
• [x] fig-cat-peer-rating
• [x] fig-sorensen-plots
• [x] fig-mixed-effect-marginal-means-plot

[egouldo]

• [x] update code in SM3 to reflect addition of collinearity analysis: