Movatterモバイル変換

r2mlm 0.3.8

Minor Edits

• Adds a warning to the r2mlm_ci() function above low coverage.
• Addresses an issue in r2mlm3_manual() to make input matchdocumentation.

r2mlm 0.3.7

Major Changes

• Add confidence interval functionality for two-level models usingMark Lai’s bootmlm package: Lai, M.H.C. (2021). Bootstrap ConfidenceIntervals for Multilevel Standardized Effect Size. MultivariateBehavioural Research, 56(4), 558-578. This only works with the automaticr2mlm() function and lme4 because bootmlm requires a fitted merModobject from lmer.

Minor Edits

• Edited formatting of r2mlm_manual() and r2mlm_comp_manual()documentation to make the examples easier to read.

r2mlm 0.3.5

Minor Edits

• Added internal function to print citation on attachment.
• Rockchalk fixed, so adding it back as a dependency.

r2mlm 0.3.4

Minor Edits

• Replaces rockchalk dependency with misty v. 0.4.12.

r2mlm 0.3.3

Minor Edits

• Fixes typo in r2mlm3_manual (#59).
• Fixes typo in r2mlm3_manual (#62).

r2mlm 0.3.2

Major Changes

• Output now returns as numeric rather than characters. (#55)

Minor Edits

• Removes broomExtra dependency. (#52, #57)
• Changes how variable types are checked from if() to is().

r2mlm 0.3.1

Major Changes

• Exportedr2mlm_long_manual to be user-facing.

Minor Edits

• Updatedr2mlm_manual andr2mlm_comp_manualdocumentation to reflect changes toteachsat datasetimplemented in version 0.3.0. (#53)

r2mlm 0.3.0

Major Changes

• Adds two manual functions: one for 3-level models (r2mlm3_manual)and one for models with heteroscedasticity, autocorrelation,nonlinearity, and non-centered-within-cluster models(r2mlm_long_manual)
• Bar graph output is now optional. The default behaviour is to outputbar graphs, but if you don’t want graphical output, the argument isbargraph = FALSE. For example,r2mlm(model, bargraph = FALSE). (Issue #46)

Bug Fixes

• To test whether clusters are mean-centered, the code computescluster means for all level-1 variables, sees if the means are roughlyzero (< .0000001), and if yes then it assignsclustermeancentered = TRUE. This update changes the code totest whether theabsolute value of the means are roughly zero,to address the case in which a cluster has a negative non-zero mean(that would otherwise mistakenly be assigned toclustermeancentered = TRUE because the negative number isless than 0.0000001). (Issue #41)
• Fixes an issue where models with non-cwc interaction terms werereturning results as though they were centered-within-cluster. r2mlmreturns non-cwc results, r2mlm_comp breaks. (Issue #42)
• Fixed an error thrown if certain groups only have one unit: “Errorin if (variance_tracker == 0) { : missing value where TRUE/FALSEneeded.” Fixed this (#44).

Minor Edits

• Changed simulated data

r2mlm 0.2.0

Major Changes

• Can now accept data with missing points, handles it with listwisedeletion via broomExtra::augment(model). (#23, #29)
• Related to accepting missing data (#23, #29), this update changesr2mlm_comp() to accept optional data argument. You can nowcallr2mlm_comp(modelA, modelB) orr2mlm_comp(modelA, modelB, data). If data is provided, thefunction will use that data. If data is not provided and models arehierarchically nested, the function will extract data automatically. Ifdata is not provided and models are not hierarchically nested, thefunction will throw an error asking users to input data.

Bug Fixes

• Bug fix: when groups of 1 exist, variance was returning as NA,generating “Error in if (variance_tracker == 0) { : missing value whereTRUE/FALSE needed” (#26)

Minor Edits

• Fixed typo inr2mlm_manual() documentation (#33)
• Updates documentation ofr2mlm() andr2mlm_comp() to note that models run inlme4must be formatted with random effects at the end of the formula.(#30)
• Refactored to increase modularity, adding files: utils.R,prepare_data.R