Merge pull request #19 from IndrajeetPatil/remove_groupedstats

remove groupedstats; closes #18
IndrajeetPatil · Jan 24, 2020 · de42720 · de42720 · lintr-bot · Jan 24, 2020
2 parents fa0e23a + c380735
commit de42720
Show file tree

Hide file tree

Showing 34 changed files with 1,602 additions and 582 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -34,14 +34,14 @@ Imports:
     crayon (>= 1.3.4),
     dplyr (>= 0.8.3),
     ez (>= 4.4-0),
-    groupedstats (>= 0.1.1),
     metaBMA (>= 0.6.2),
     metafor (>= 2.1-0),
     metaplus (>= 0.7-11),
     psych (>= 1.9.12.31),
     purrr (>= 0.3.3),
     rcompanion (>= 2.3.21),
     rlang (>= 0.4.2),
+    sjstats (>= 0.17.7),
     stats,
     tibble (>= 2.1.3),
     tidyr (>= 1.0.0),

diff --git a/NAMESPACE b/NAMESPACE
@@ -24,6 +24,7 @@ export(expr_anova_robust)
 export(expr_contingency_tab)
 export(expr_corr_test)
 export(expr_mann_nonparametric)
+export(expr_meta_bayes)
 export(expr_meta_parametric)
 export(expr_meta_robust)
 export(expr_onesample_proptest)
@@ -33,10 +34,7 @@ export(expr_t_onesample)
 export(expr_t_parametric)
 export(expr_t_robust)
 export(expr_template)
-export(lm_effsize_ci)
 export(long_to_wide_converter)
-export(set_cwd)
-export(signif_column)
 export(specify_decimal_p)
 export(tibble)
 importFrom(BayesFactor,anovaBF)
@@ -87,11 +85,6 @@ importFrom(dplyr,tibble)
 importFrom(dplyr,ungroup)
 importFrom(dplyr,vars)
 importFrom(ez,ezANOVA)
-importFrom(groupedstats,lm_effsize_ci)
-importFrom(groupedstats,lm_effsize_standardizer)
-importFrom(groupedstats,set_cwd)
-importFrom(groupedstats,signif_column)
-importFrom(groupedstats,specify_decimal_p)
 importFrom(metaBMA,meta_random)
 importFrom(metaBMA,prior)
 importFrom(metafor,rma)
@@ -121,6 +114,9 @@ importFrom(rlang,expr)
 importFrom(rlang,is_formula)
 importFrom(rlang,new_formula)
 importFrom(rlang,quo_is_null)
+importFrom(sjstats,eta_sq)
+importFrom(sjstats,omega_sq)
+importFrom(stats,aov)
 importFrom(stats,chisq.test)
 importFrom(stats,cor)
 importFrom(stats,cor.test)

diff --git a/NEWS.md b/NEWS.md
@@ -1,11 +1,15 @@
 # statsExpressions 0.2.1.9000
 
+NEW FEATURES
+
+  - New functions to carry out meta-analysis: `expr_meta_bayes`.
+
 # statsExpressions 0.2.1
 
 NEW FEATURES
 
   - New functions to carry out meta-analysis: `expr_meta_parametric`,
-    `expr_meta_robust`, `expr_meta_bayes`.
+    `expr_meta_robust`, `bf_meta`.
 
 # statsExpressions 0.2.0
 

diff --git a/R/helpers_anova_expressions.R b/R/helpers_anova_expressions.R
@@ -30,9 +30,8 @@
 #'
 #' @importFrom dplyr select rename matches
 #' @importFrom rlang !! enquo eval_tidy expr ensym
-#' @importFrom stats lm oneway.test na.omit
+#' @importFrom stats aov oneway.test na.omit
 #' @importFrom ez ezANOVA
-#' @importFrom groupedstats lm_effsize_standardizer
 #'
 #' @examples
 #' \donttest{
@@ -236,7 +235,7 @@ expr_anova_parametric <- function(data,
 
     # creating a standardized dataframe with effect size and its CIs
     effsize_object <-
-      stats::lm(
+      stats::aov(
         formula = rlang::new_formula({{ y }}, {{ x }}),
         data = data,
         na.action = na.omit
@@ -245,12 +244,12 @@ expr_anova_parametric <- function(data,
 
   # creating a standardized dataframe with effect size and its CIs
   effsize_df <-
-    groupedstats::lm_effsize_standardizer(
-      object = effsize_object,
+    aov_effsize(
+      model = effsize_object,
       effsize = effsize,
       partial = partial,
-      conf.level = conf.level,
-      nboot = nboot
+      ci = conf.level,
+      iterations = nboot
     )
 
   # preparing subtitle

diff --git a/R/helpers_bf_tests.R b/R/helpers_bf_tests.R
@@ -57,7 +57,7 @@ bf_extractor <- function(bf.object, ...) {
 #' set.seed(123)
 #'
 #' # dataframe containing results
-#' bf_results <-
+#' bf.df <-
 #'   statsExpressions::bf_extractor(BayesFactor::correlationBF(
 #'     x = iris$Sepal.Length,
 #'     y = iris$Petal.Length
@@ -66,15 +66,15 @@ bf_extractor <- function(bf.object, ...) {
 #'
 #' # creating caption (for null)
 #' statsExpressions::bf_expr(
-#'   bf.df = bf_results,
+#'   bf.df = bf.df,
 #'   output = "null",
 #'   k = 3,
 #'   caption = "Note: Iris dataset"
 #' )
 #'
 #' # creating caption (for alternative)
 #' statsExpressions::bf_expr(
-#'   bf.df = bf_results,
+#'   bf.df = bf.df,
 #'   output = "alternative"
 #' )
 #' }
@@ -190,7 +190,7 @@ bf_corr_test <- function(data,
   # ========================= subtitle preparation ==========================
 
   # extracting results from Bayesian test and creating a dataframe
-  bf_results <-
+  bf.df <-
     bf_extractor(
       BayesFactor::correlationBF(
         x = data %>% dplyr::pull({{ x }}),
@@ -206,7 +206,7 @@ bf_corr_test <- function(data,
   if (output != "results") {
     bf_message <-
       bf_expr(
-        bf.df = bf_results,
+        bf.df = bf.df,
         output = output,
         k = k,
         caption = caption
@@ -218,7 +218,7 @@ bf_corr_test <- function(data,
   # return the text results or the dataframe with results
   return(switch(
     EXPR = output,
-    "results" = bf_results,
+    "results" = bf.df,
     bf_message
   ))
 }
@@ -358,7 +358,7 @@ bf_contingency_tab <- function(data,
       )
 
     # extracting results from Bayesian test and creating a dataframe
-    bf_results <-
+    bf.df <-
       bf_extractor(
         BayesFactor::contingencyTableBF(
           x = table(
@@ -427,7 +427,7 @@ bf_contingency_tab <- function(data,
     pr_y_h1 <- BayesFactor::logMeanExpLogs(tmp_pr_h1)
 
     # computing Bayes Factor and formatting the results
-    bf_results <-
+    bf.df <-
       dplyr::tibble(bf10 = exp(pr_y_h1 - pr_y_h0)) %>%
       bf_formatter(.) %>%
       dplyr::mutate(.data = ., prior.concentration = prior.concentration)
@@ -438,11 +438,11 @@ bf_contingency_tab <- function(data,
   # changing aspects of the caption based on what output is needed
   if (output %in% c("null", "caption", "H0", "h0")) {
     hypothesis.text <- "In favor of null: "
-    bf.value <- bf_results$log_e_bf01[[1]]
+    bf.value <- bf.df$log_e_bf01[[1]]
     bf.subscript <- "01"
   } else {
     hypothesis.text <- "In favor of alternative: "
-    bf.value <- -bf_results$log_e_bf01[[1]]
+    bf.value <- -bf.df$log_e_bf01[[1]]
     bf.subscript <- "10"
   }
 
@@ -472,7 +472,7 @@ bf_contingency_tab <- function(data,
           bf.subscript = bf.subscript,
           bf = specify_decimal_p(x = bf.value, k = k),
           sampling.plan = sampling_plan_text,
-          a = specify_decimal_p(x = bf_results$prior.concentration[[1]], k = k)
+          a = specify_decimal_p(x = bf.df$prior.concentration[[1]], k = k)
         )
       )
   }
@@ -500,7 +500,7 @@ bf_contingency_tab <- function(data,
           top.text = caption,
           bf.subscript = bf.subscript,
           bf = specify_decimal_p(x = bf.value, k = k),
-          a = specify_decimal_p(x = bf_results$prior.concentration[[1]], k = k)
+          a = specify_decimal_p(x = bf.df$prior.concentration[[1]], k = k)
         )
       )
   }
@@ -510,7 +510,7 @@ bf_contingency_tab <- function(data,
   # return the text results or the dataframe with results
   return(switch(
     EXPR = output,
-    "results" = bf_results,
+    "results" = bf.df,
     bf_message
   ))
 }
@@ -675,7 +675,7 @@ bf_ttest <- function(data,
   }
 
   # extracting the Bayes factors
-  bf_results <-
+  bf.df <-
     bf_extractor(bf.object = bf_object) %>%
     dplyr::mutate(.data = ., bf.prior = bf.prior)
 
@@ -685,7 +685,7 @@ bf_ttest <- function(data,
   if (output != "results") {
     bf_message <-
       bf_expr(
-        bf.df = bf_results,
+        bf.df = bf.df,
         output = output,
         k = k,
         caption = caption
@@ -695,7 +695,7 @@ bf_ttest <- function(data,
   # return the text results or the dataframe with results
   return(switch(
     EXPR = output,
-    "results" = bf_results,
+    "results" = bf.df,
     bf_message
   ))
 }
@@ -790,7 +790,7 @@ bf_oneway_anova <- function(data,
       dplyr::mutate(.data = ., rowid = as.factor(rowid))
 
     # extracting results from Bayesian test (`y ~ x + id`) and creating a dataframe
-    bf_results <-
+    bf.df <-
       bf_extractor(BayesFactor::anovaBF(
         formula = rlang::new_formula(
           {{ rlang::enexpr(y) }}, rlang::expr(!!rlang::enexpr(x) + rowid)
@@ -812,7 +812,7 @@ bf_oneway_anova <- function(data,
     data %<>% tidyr::drop_na(.)
 
     # extracting results from Bayesian test and creating a dataframe
-    bf_results <-
+    bf.df <-
       bf_extractor(
         BayesFactor::anovaBF(
           formula = rlang::new_formula({{ y }}, {{ x }}),
@@ -831,7 +831,7 @@ bf_oneway_anova <- function(data,
   if (output != "results") {
     bf_message <-
       bf_expr(
-        bf.df = bf_results,
+        bf.df = bf.df,
         output = output,
         k = k,
         caption = caption
@@ -841,7 +841,7 @@ bf_oneway_anova <- function(data,
   # return the text results or the dataframe with results
   return(switch(
     EXPR = output,
-    "results" = bf_results,
+    "results" = bf.df,
     bf_message
   ))
 }
@@ -855,6 +855,7 @@ bf_oneway_anova <- function(data,
 #' @inherit metaBMA::meta_random return Description
 #'
 #' @inheritParams expr_meta_parametric
+#' @inheritParams bf_expr
 #' @inheritParams metaBMA::meta_random
 #' @inheritDotParams metaBMA::meta_random -y -SE
 #'
@@ -894,7 +895,9 @@ bf_oneway_anova <- function(data,
 #'   data = df,
 #'   k = 3,
 #'   iter = 1500,
-#'   messages = TRUE
+#'   messages = TRUE,
+#'   # customizing analysis with additional arguments
+#'   control = list(max_treedepth = 15)
 #' )
 #' }
 #'
@@ -905,6 +908,7 @@ bf_meta <- function(data,
                     d = prior("norm", c(mean = 0, sd = 0.3)),
                     tau = prior("invgamma", c(shape = 1, scale = 0.15)),
                     k = 2,
+                    output = "null",
                     caption = NULL,
                     messages = TRUE,
                     ...) {
@@ -935,18 +939,29 @@ bf_meta <- function(data,
     dplyr::filter(.data = ., term == "d")
 
   # dataframe with bayes factors
-  bf_results <-
+  bf.df <-
     dplyr::tibble(bf10 = meta_res$BF["random_H1", "random_H0"]) %>%
     bf_formatter(.)
 
+  # changing aspects of the caption based on what output is needed
+  if (output %in% c("null", "caption", "H0", "h0")) {
+    hypothesis.text <- "In favor of null: "
+    bf.value <- bf.df$log_e_bf01[[1]]
+    bf.subscript <- "01"
+  } else {
+    hypothesis.text <- "In favor of alternative: "
+    bf.value <- bf.df$log_e_bf10[[1]]
+    bf.subscript <- "10"
+  }
+
   # prepare the Bayes factor message
   bf_text <-
     substitute(
       atop(displaystyle(top.text),
         expr = paste(
-          "In favor of null: ",
+          hypothesis.text,
           "log"["e"],
-          "(BF"["01"],
+          "(BF"[bf.subscript],
           ") = ",
           bf,
           ", ",
@@ -962,8 +977,10 @@ bf_meta <- function(data,
         )
       ),
       env = list(
+        hypothesis.text = hypothesis.text,
         top.text = caption,
-        bf = specify_decimal_p(x = bf_results$log_e_bf01[[1]], k = k),
+        bf.subscript = bf.subscript,
+        bf = specify_decimal_p(x = bf.value, k = k),
         d.pmean = specify_decimal_p(x = df_estimates$mean[[1]], k = k),
         d.pmean.LB = specify_decimal_p(x = df_estimates$hpd95_lower[[1]], k = k),
         d.pmean.UB = specify_decimal_p(x = df_estimates$hpd95_upper[[1]], k = k)