NEWS.md
As part of a major overhaul of the code in order to improve maintainability and user-friendliness, there have been a number of significant changes made in this version.
rob_traffic_light() has been changed from 20 to 10. This is because, following a review of published robvis plots, users are not changing the defaults very often and in almost all cases, a smaller default looks better. While existing code will continue to work, the resulting plots will look different unless the psize argument was explicitly defined.
overall, which allows users to specify whether to include an “Overall” column in the traffic light plot, has been added to rob_traffic_light() (default is TRUE). In order to have consistent behaviour across the two main functions, the default of the overall argument inrob_summary() has been changed from FALSE to TRUE [I admit that I’m not sure why I set it to FALSE in the first place]. While existing code will continue to work, the resulting plots will look different unless the overall argument was explicitly defined.
weighted argument for rob_summary() has been changed from TRUE to FALSE, both to ensure better consistency across the two main functions (rob_traffic_light() does not require a “Weight” variable) and to reduce a hurdle to usage among new users. While existing code will continue to work, the resulting plots will look different unless the weighted argument was explicitly defined.
weighted = FALSE as the default for rob_summary().overall argument in the “Generic” now acts slightly differently both to previous versions of the function and to the other templates. Rather than removing an additional column from the dataset if overall = FALSE, which doesn’t make sense if their data doesn’t contain an overall column, it simply highlights the last column (in rob_traffic_light()) or bolds the last heading (in rob_summary()).rob_summary() compatible with tools that have 4 levels of judgement (e.g. the ROBINS-I tool), users must now specify a name for all levels if using the judgement_labels argument of rob_summary(). Existing code that uses the judgement_labels argument will need to be updated.
metafor meta-analyses have been added. rob_forest() adds a traffic-light plot to the right hand side of the standard metafor::forest() output. rob_append_weights() extracts the weights assigned to each study in the meta-analysis and appends them to the risk of bias dataset. A vignette describing these two new functions has been added.rob_save(), has been added, which uses data-driven defaults for figure height and width when saving to a file.robvis now supports “No information” as a judgement has been added.colour = "colourblind"argument has been added.overall argument in rob_summary() now bolds the “Overall” axis label.rob_tools() now returns a message indicating the availability of templates for each function.rob_traffic_light() can now handle a summary table with or without a “Weight column”. rob_summary() still requires a “Weight” column, as it is strongly encouraged that weights be used to create the summary barplot.rob_summary() and rob_traffic_light().rob_traffic_light() have been updated.rob_traffic_light() no longer orders studies alphabetically, instead retaining the order from the uploaded CSV.rob_summary() has been corrected to accurately reflect the order of domains in each tool, rather than presenting them in alphabetical order.rob_traffic_light() are now ordered by ascending/descending (depending on the tool) risk of bias.CODE_OF_CONDUCT and CONTRIBUTING documents have been added to the GitHub repo.rob_summary() and rob_traffic_light() functions. In line with CRAN guidelines, this element was removed as packages should not write to the file system.rob_summary() to allow users to choose whether to create a weighted or unweighted barplot.