code style

R/3_p_value.R

@@ -144,20 +144,20 @@ p_value.default <- function(model,
     })
   }
 
-  # output
-  if (!is.null(p)) {
-    params <- insight::get_parameters(model, component = component)
-    if (length(p) == nrow(params) && "Component" %in% colnames(params)) {
-      p <- .data_frame(Parameter = params$Parameter, p = as.vector(p), Component = params$Component)
-    } else {
-      p <- .data_frame(Parameter = names(p), p = as.vector(p))
+  # failure warning
+  if (is.null(p)) {
+    if (isTRUE(verbose)) {
+      insight::format_warning("Could not extract p-values from model object.")
     }
-    return(p)
+    return(NULL)
   }
 
-  # failure warning
-  if (is.null(p) && isTRUE(verbose)) {
-    insight::format_warning("Could not extract p-values from model object.")
+  # output
+  params <- insight::get_parameters(model, component = component)
+  if (length(p) == nrow(params) && "Component" %in% colnames(params)) {
+    .data_frame(Parameter = params$Parameter, p = as.vector(p), Component = params$Component)
+  } else {
+    .data_frame(Parameter = names(p), p = as.vector(p))
   }
 }
 

R/4_standard_error.R

@@ -81,6 +81,10 @@ standard_error.default <- function(model,
   dots <- list(...)
   se <- NULL
 
+  # if a vcov is provided, we calculate standard errors based on that matrix
+  # this is usually the case for HC (robust) standard errors
+  # ------------------------------------------------------------------------
+
   # vcov: matrix
   if (is.matrix(vcov)) {
     se <- sqrt(diag(vcov))
@@ -105,7 +109,9 @@ standard_error.default <- function(model,
     se <- sqrt(diag(.vcov))
   }
 
-  # classical se from summary()
+  # classical SE from summary()
+  # ------------------------------------------------------------------------
+
   if (is.null(se)) {
     se <- .safe({
       if (grepl("Zelig-", class(model)[1], fixed = TRUE)) {
@@ -116,7 +122,10 @@ standard_error.default <- function(model,
     })
   }
 
-  # classical se from get_varcov()
+  # if retrieving SE from summary() failed, we try to calculate SE based
+  # on classical se from get_varcov()
+  # ------------------------------------------------------------------------
+
   if (is.null(se)) {
     se <- .safe({
       varcov <- insight::get_varcov(model, component = component)
@@ -150,27 +159,30 @@ standard_error.default <- function(model,
 
 
 .get_se_from_summary <- function(model, component = NULL) {
-  cs <- suppressWarnings(stats::coef(summary(model)))
+  cs <- .safe(suppressWarnings(stats::coef(summary(model))))
   se <- NULL
 
-  if (is.list(cs) && !is.null(component)) cs <- cs[[component]]
-
-  if (!is.null(cs)) {
-    # do we have a se column?
-    se_col <- which(colnames(cs) == "Std. Error")
-
-    # if not, default to 2
-    if (length(se_col) == 0) se_col <- 2
-
-    se <- as.vector(cs[, se_col])
-
-    if (is.null(names(se))) {
-      coef_names <- rownames(cs)
-      if (length(coef_names) == length(se)) names(se) <- coef_names
+  if (!is.null(se)) {
+    if (is.list(cs) && !is.null(component)) {
+      cs <- cs[[component]]
     }
+    if (!is.null(cs)) {
+      # do we have a se column?
+      se_col <- which(colnames(cs) == "Std. Error")
+      # if not, default to 2
+      if (length(se_col) == 0) {
+        se_col <- 2
+      }
+      se <- as.vector(cs[, se_col])
+      if (is.null(names(se))) {
+        coef_names <- rownames(cs)
+        if (length(coef_names) == length(se)) {
+          names(se) <- coef_names
+        }
+      }
+    }
+    names(se) <- .remove_backticks_from_string(names(se))
   }
-
-  names(se) <- .remove_backticks_from_string(names(se))
   se
 }
 

R/ci_generic.R

@@ -97,10 +97,14 @@
   }
   method <- tolower(method)
 
+  # Fist, we want standard errors for parameters
+  # --------------------------------------------
+
   # if we have adjusted SE, e.g. from kenward-roger, don't recompute
   # standard errors to save time...
   if (is.null(se)) {
     if (!is.null(vcov) || isTRUE(list(...)[["robust"]])) {
+      # robust (HC) standard errors?
       stderror <- standard_error(model,
         component = component,
         vcov = vcov,
@@ -109,6 +113,7 @@
         ...
       )
     } else {
+      # normal standard errors, including small-sample approximations
       stderror <- switch(method,
         kenward = se_kenward(model),
         kr = se_kenward(model),
@@ -133,6 +138,9 @@
     se <- stderror$SE
   }
 
+  # Next, we need degrees of freedom
+  # --------------------------------
+
   # check if we have a valid dof vector
   if (is.null(dof)) {
     # residual df
@@ -146,7 +154,9 @@
     }
   }
 
-  # calculate CIs
+  # Now we can calculate CIs
+  # ------------------------
+
   alpha <- (1 + ci) / 2
   fac <- suppressWarnings(stats::qt(alpha, df = dof))
   out <- cbind(
@@ -171,7 +181,6 @@
 }
 
 
-
 .is_chi2_model <- function(model, dof) {
   statistic <- insight::find_statistic(model)
   (all(dof == 1) && identical(statistic, "chi-squared statistic"))