NEWS.md
add_between() and add_within() don’t convert non-character levels to factors any morermulti() function for multivariate distributions that aren’t all normal (experimental)add_random()
long argument for sim_df()
rmulti() and associated helper functions convert_r() and fh_bounds().rmulti() function for multivariate distributions that aren’t all normal (experimental)add_random() now names random factor items with the full random factor name (e.g., “class1”, not “c1”)add_random() allows you to set specific factor item names (see vignette)sim_design() now names anonymous within and between factors like W and B or W1, W2, W3, …, and B1, B2, … instead of A, B, C, …add_contrast() and associated contr_code_*** functionsadd_random() and associated mixed design building functionsget_params() doesn’t need between, within, id, and dv set for date created by sim_design()
plot_design() can display a subset of factorssim_design() fixed a bug in when setting n with an unnamed vector and within-subjects factorssim_design() when setting n with an unnamed vector and within-subjects factors (wouldn’t run before).add_between() and add_within() to make new columns factors with the same ordering as the specificationadd_between() .prob argument works as expected now (and has tests)contr_code_deviation() to contr_code_anova()
add_contrast() functioncontr_code_
plot_design() can display a subset of factorssim_design() are now named W and B or W1, W2, W3, …, B1, B2, … instead of A, B, C, … (and fixed relevant tests and vignette code)get_params() so it doesn’t need between, within, id, and dv set for date created by sim_design()
rnorm_pre() when simulating a vector with correlations to more than 2 pre-existing vectors.sim_design() should no longer mangle level values in long format if they have underscoressim_design() should play better with different separator. FOr example, if you set faux_options(sep = ".") and have within-subject factors A and B with levels A_1/A_2 and B_1/B_2, your wide data will have columns A_1.B_1, A_1.B_2, A_2.B_1, A_2.B_2sim_design() where parameters specified as a named vector couldn’t be in a different order unless both between and within factors were specified (e.g., mu = c(A2 = 2, A1 = 1) resulted in a mu of 2 for A1 and 1 for A2).sim_joint_dist() function to simulate the joint distribution of categoriessim_df() no longer breaks if there are NAs in the DV columnssim_df() now has an option to include missing data, it simulates the joint distribution of missingness for each between-subject cellsim_df() and messy()) can choose columns with unquoted names now (e.g., messy(mtcars, .5, mpg))messy() now takes a vector of proportions so you can simulate different amounts of missing data per selected columnsample_from_pop() is now vectorisedget_params() doesn’t require within and between set for data made with faux (that has a “design” attribute)get_params() where the var column was alphabetised, but the corresponding columns for the correlation table were in factor ordernested_list() updated to match scienceverse version and handle edge cases betterrnorm_multi() can get column names from mu, sd, or r namesseed arguments reinstated as deprecated and produce a warningseed arguments (at the request of CRAN)seed argument to rnorm_multi()
nested_list function for printing nested lists in Rmdcodebook function and vignettenorm2beta functiontrunc2norm now works if min or max are omitted.rep argument to sim_design() and sim_data(). If rep > 1, returns a nested data frame with rep simulated datasets.get_params()
make_id() functionfaux_options(plot = TRUE)
dv = list(colname = "Name for Plots"))sim_design() can take intercept-only designsrnorm_multi() can take vars = 1 for intercept-only designsjson_design() to output or save design specs in JSON formatmessy() (thanks Emily)long2wide() (handle designs with no between or no within factors)sim_df() returns subject IDs and takes data in long formatcheck_sim_stats() to get_params(), which now returns the designsim_design()
sim_mixed_cc() to simulate null cross-classified mixed effect designs by subject, item and error SDssim_design(), sim_df(), sim_mixed_cc() and sim_mixed_df() take a seed argument now for reproducible datasetscheck_design() and sim_design()
check_design() have a more consistent format
within and between are named lists; factors and labels are no longer separately namedsim_design() (failed when within or between factor number was 0)NEWS.md file to track changes to the package.sim_design() to simulate data for mixed ANOVA designs.