fix splithalf error when two variables

R/standardErrorOfMeasurement.R

@@ -77,7 +77,7 @@ standardErrorOfMeasurement <- function(jaspResults, dataset, options) {
              exitAnalysisIfErrors = TRUE)
 
   if (options[["lord2"]] && options[["lord2NumberOfSplits"]] == "") {
-    .quitAnalysis(gettext("For the Lord's compound method, the test could not be split in equally sized parts with more than 1 item per part. Consider adding or removing items."))
+    .quitAnalysis(gettext("For the Lord generalized method, the test could not be split in equally sized parts with more than 1 item per part. Consider adding or removing items."))
   }
 }
 
@@ -128,7 +128,7 @@ standardErrorOfMeasurement <- function(jaspResults, dataset, options) {
     # at least one method is selected
     for (i in 1:length(selected)) {
       if (is.na(selected[[i]][["name"]])) {
-        .quitAnalysis(gettext("The Lord, Keats, and Lord's compound method are only available for binary data."))
+        .quitAnalysis(gettext("The Lord, Keats, and Lord generalized method are only available for binary data."))
       }
       if (is.null(jaspResults[["semMainContainer"]][[paste0(method[i], "State")]])) {
         out <- eval(parse(text = selected[[i]][["funString"]]))
@@ -485,15 +485,15 @@ standardErrorOfMeasurement <- function(jaspResults, dataset, options) {
     pl <- ggplot2::ggplot(dat) +
       ggplot2::geom_point(ggplot2::aes(x = score, y = tscore), size = 2.5) +
       ggplot2::geom_errorbar(ggplot2::aes(x = score, ymin = lower, ymax = upper), width = 0.5) +
-      ggplot2::labs(x = gettext("Sum Score"), y = gettext("True Score"))
+      ggplot2::labs(x = gettext("Sum Score"), y = gettext("True Score Estimate"))
   } else {
     dat <- as.data.frame(ciData)
     colnames(dat) <- c("score", "lower", "upper")
     dat$tscore <- dat$score
     pl <- ggplot2::ggplot(dat) +
       ggplot2::geom_ribbon(ggplot2::aes(x = score, ymin = lower, ymax = upper), fill = "grey80") +
       ggplot2::geom_line(ggplot2::aes(x = score, y = tscore)) +
-      ggplot2::labs(x = gettext("Sum Score"), y = gettext("True Score"))
+      ggplot2::labs(x = gettext("Sum Score"), y = gettext("True Score Estimate"))
   }
 
   if (!is.na(cutoff)) {
@@ -539,8 +539,19 @@ standardErrorOfMeasurement <- function(jaspResults, dataset, options) {
   out <- .semPrepareOutMatrix(ncol(X), nc, scoresObj)
   fun <- function(partSUMS, ind, cc) {
     K <- ncol(partSUMS)
-    mean_diff <- partSUMS[ind, ] - rowMeans(partSUMS[ind, ]) - matrix(colMeans(partSUMS[ind, ]), length(ind), K, TRUE) + mean(partSUMS[ind, ])
-    ret <- sqrt(d * sum(rowSums(mean_diff^2) / (K - 1)) / length(ind))
+    # mean_diff <- partSUMS[ind, ] - rowMeans(partSUMS[ind, ]) - matrix(colMeans(partSUMS[ind, ]), length(ind), K, TRUE) + mean(partSUMS[ind, ])
+    # ret <- sqrt(d * sum(rowSums(mean_diff^2) / (K - 1)) / length(ind))
+
+    col_means <- colMeans(partSUMS)    # Global marginal means (X_j)
+    grand_mean <- mean(col_means)    # Global grand mean (M)
+    row_means <- rowMeans(partSUMS[ind, ])    # Person means (X_i)
+
+    # Construct Feldt's deviation term
+    mean_diff <- partSUMS[ind, ] -
+      matrix(row_means, length(ind), K, byrow = FALSE) -
+      matrix(col_means - grand_mean, length(ind), K, byrow = TRUE)
+
+    ret <- sqrt( d * sum(mean_diff^2) / ((K - 1) * length(ind)) )
     return(ret)
   }
   out <- .semComputeWithCaseMin(out, S, caseMin, partSUMS, fun)
@@ -557,9 +568,7 @@ standardErrorOfMeasurement <- function(jaspResults, dataset, options) {
   out <- .semPrepareOutMatrix(ncol(X), nc, scoresObj)
   scores <- out[, 1]
 
-  rawDiffK <- d *
-    rowSums((partSUMS - matrix(colMeans(partSUMS), N, K, TRUE) - rowMeans(partSUMS) + mean(partSUMS))^2) /
-    (K - 1)
+  rawDiffK <- d * rowSums((partSUMS - matrix(colMeans(partSUMS), N, K, TRUE) - rowMeans(partSUMS) + mean(partSUMS))^2) / (K - 1)
   betaK <- coef(lm(rawDiffK ~ poly(S, n_poly, raw = TRUE)))
   scrs <- sqrt(betaK[1] + rowSums(matrix(betaK[-1], length(scores), n_poly, TRUE) * poly(scores, n_poly, raw = TRUE)))
   out[, 2] <- scrs
@@ -882,7 +891,7 @@ standardErrorOfMeasurement <- function(jaspResults, dataset, options) {
                         funString = NA,
                         dependencies = NA)),
     lord2 = switch(as.character(nc),
-                   "2" = list(name = "Lord's compound",
+                   "2" = list(name = "Lord generalized",
                               funString = ".semLord2(dataset, as.numeric(options$lord2NumberOfSplits), scrs, options$minimumGroupSize)",
                               dependencies = c("lord2", "lord2NumberOfSplits", "minimumGroupSize")),
                    list(name = NA,

R/unidimensionalReliabilityBayesian.R

@@ -75,13 +75,13 @@ unidimensionalReliabilityBayesian <- function(jaspResults, dataset, options) {
     itemDroppedSelectedItem = unlist(options[c("itemDeletedOmega", "itemDeletedAlpha", "itemDeletedLambda2", "itemDeletedSplithalf")]),
 
     namesEstimators     = list(
-      tables = c("Coefficient \u03C9", "Coefficient \u03B1", "Guttman's \u03BB2", "Split-half coefficient",
+      tables = c("McDonald's \u03C9", "Cronbach's \u03B1", "Guttman's \u03BB2", "Split-half coefficient",
                  "Average interitem correlation", "Mean", "Variance", "SD"),
-      tables_item = c("Coefficient \u03C9", "Coefficient \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
+      tables_item = c("McDonald's \u03C9", "Cronbach's \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
                       gettext("Item-rest correlation"), gettext("Mean"), gettext("Variance"), gettext("SD")),
-      coefficients = c("Coefficient \u03C9", "Coefficient \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
+      coefficients = c("McDonald's \u03C9", "Cronbach's \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
                        gettext("Item-rest correlation")),
-      plots = list(expression("Coefficient"~omega), expression("Cronbach\'s"~alpha), expression("Guttman's"~lambda[2]),
+      plots = list(expression("McDonald\'s"~omega), expression("Cronbach\'s"~alpha), expression("Guttman\'s"~lambda[2]),
                    gettext("Split-half coefficient")),
       plotsNoGreek = c("omega", "alpha", "lambda2", "splithalf")
     )

R/unidimensionalReliabilityFrequentist.R

@@ -2,6 +2,7 @@
 #' @export
 unidimensionalReliabilityFrequentist <- function(jaspResults, dataset, options) {
 
+
   # check for listwise deletion
   datasetOld <- dataset
   dataset <- .handleData(datasetOld, options)
@@ -54,11 +55,11 @@ unidimensionalReliabilityFrequentist <- function(jaspResults, dataset, options)
     itemDroppedSelected = unlist(options[c("itemDeletedOmega", "itemDeletedAlpha", "itemDeletedLambda2", "itemDeletedSplithalf",
                                            "itemRestCorrelation", "itemMean", "itemVar", "itemSd")]),
     namesEstimators     = list(
-      tables = c("Coefficient \u03C9", "Coefficient \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
+      tables = c("McDonald's \u03C9", "Cronbach's \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
                  gettext("Average interitem correlation"), gettext("Mean"), gettext("Variance"), gettext("SD")),
-      tables_item = c("Coefficient \u03C9", "Coefficient \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
+      tables_item = c("McDonald's \u03C9", "Cronbach's \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient"),
                       gettext("Item-rest correlation"), gettext("Mean"), gettext("Variance"), gettext("SD")),
-      coefficientsDeleted = c("Coefficient \u03C9", "Coefficient \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient")))
+      coefficientsDeleted = c("McDonald's \u03C9", "Cronbach's \u03B1", "Guttman's \u03BB2", gettext("Split-half coefficient")))
   )
 
   return(derivedOptions)
@@ -952,8 +953,8 @@ unidimensionalReliabilityFrequentist <- function(jaspResults, dataset, options)
         out[["conf"]][["scaleSplithalf"]] <- quantile(samp, probs = c((1 - ciValue) / 2, 1 - (1 - ciValue) / 2), na.rm = TRUE)
         out[["se"]][["scaleSplithalf"]] <- sd(samp, na.rm = TRUE)
       } else { # interval analytic
-        partSums1 <- rowSums(dtUse[, splits[[1]]])
-        partSums2 <- rowSums(dtUse[, splits[[2]]])
+        partSums1 <- rowSums(dtUse[, splits[[1]], drop = FALSE])
+        partSums2 <- rowSums(dtUse[, splits[[2]], drop = FALSE])
 
         out[["se"]][["scaleSplithalf"]] <- .seSplithalf(partSums1, partSums2, model[["use.cases"]])
         out[["conf"]][["scaleSplithalf"]] <- out[["est"]][["scaleSplithalf"]] + c(-1, 1) * out[["se"]][["scaleSplithalf"]] * qnorm(1 - (1 - ciValue) / 2)
@@ -1168,31 +1169,33 @@ unidimensionalReliabilityFrequentist <- function(jaspResults, dataset, options)
         out[["est"]][["itemDeletedSplithalf"]] <- c(NA, NA)
         out[["lower"]][["itemDeletedSplithalf"]] <- c(NA, NA)
         out[["upper"]][["itemDeletedSplithalf"]] <- c(NA, NA)
-      }
-
-      for (i in seq_len(ncol(dtUse))) {
-        dtCut <- dtUse[, -i, drop = FALSE]
-        nit <- ncol(dtCut)
-        splits <- split(seq_len(nit), 1:2)
-        est <- .splithalfData(dtCut, splits = splits, useCase = model[["use.cases"]])
-        out[["est"]][["itemDeletedSplithalf"]][i] <- est
+      } else {
+        for (i in seq_len(ncol(dtUse))) {
+          dtCut <- dtUse[, -i, drop = FALSE]
+          nit <- ncol(dtCut)
+          splits <- split(seq_len(nit), 1:2)
+          est <- .splithalfData(dtCut, splits = splits, useCase = model[["use.cases"]])
+          out[["est"]][["itemDeletedSplithalf"]][i] <- est
 
-        if (options[["intervalMethod"]] == "analytic") {
+          if (options[["intervalMethod"]] == "analytic") {
 
-          partSums1 <- rowSums(dtCut[, splits[[1]]])
-          partSums2 <- rowSums(dtCut[, splits[[2]]])
+            partSums1 <- rowSums(dtCut[, splits[[1]]])
+            partSums2 <- rowSums(dtCut[, splits[[2]]])
 
-          se <- .seSplithalf(partSums1, partSums2, model[["use.cases"]])
-          conf <- est + c(-1, 1) * se * qnorm(1 - (1 - ciValue) / 2)
-          out[["lower"]][["itemDeletedSplithalf"]][i] <- conf[1]
-          out[["upper"]][["itemDeletedSplithalf"]][i] <- conf[2]
-        } else {
-          itemSamp <- model[["itemDeletedSplithalf"]][["itemSamp"]]
-          conf <- quantile(itemSamp[, i], probs = c((1 - ciValue) / 2, 1 - (1 - ciValue) / 2), na.rm = TRUE)
-          out[["lower"]][["itemDeletedSplithalf"]][i] <- conf[1]
-          out[["upper"]][["itemDeletedSplithalf"]][i] <- conf[2]
+            se <- .seSplithalf(partSums1, partSums2, model[["use.cases"]])
+            conf <- est + c(-1, 1) * se * qnorm(1 - (1 - ciValue) / 2)
+            out[["lower"]][["itemDeletedSplithalf"]][i] <- conf[1]
+            out[["upper"]][["itemDeletedSplithalf"]][i] <- conf[2]
+          } else {
+            itemSamp <- model[["itemDeletedSplithalf"]][["itemSamp"]]
+            conf <- quantile(itemSamp[, i], probs = c((1 - ciValue) / 2, 1 - (1 - ciValue) / 2), na.rm = TRUE)
+            out[["lower"]][["itemDeletedSplithalf"]][i] <- conf[1]
+            out[["upper"]][["itemDeletedSplithalf"]][i] <- conf[2]
+          }
         }
       }
+
+
     }
 
     # item-rest correlation

inst/Description.qml

@@ -16,7 +16,7 @@ Description
 
 	Analysis
 	{
-		title:	qsTr("Unidimensional Reliability")
+		title:	qsTr("Reliability")
 		qml: 	"UnidimensionalReliabilityFrequentist.qml"
 		func: 	"unidimensionalReliabilityFrequentist"
 	}
@@ -57,8 +57,8 @@ Description
 	}
 	Analysis
 	{
-		menu: 	qsTr("Unidimensional Reliability")
-		title: 	qsTr("Bayesian Unidimensional Reliability")
+		menu: 	qsTr("Reliability")
+		title: 	qsTr("Bayesian Reliability")
 		qml: 	"UnidimensionalReliabilityBayesian.qml"
 		func: 	"unidimensionalReliabilityBayesian"
 	}

inst/help/standardErrorOfMeasurement.md

@@ -23,7 +23,7 @@ The Standard Error of Measurement (sem) quantifies the precision of a measuremen
 - Binomial methods: Based on the idea that the item scores follow a binomial distribution
   - Lord: Method that only requires number of correct and incorrect items
   - Keats: Corrects the Lord method for supposed bias and uses a reliability coefficient in the process
-  - Lord's compound: Essentially the Lord method for multiple test parts
+  - Lord generalized: Essentially the Lord method for multiple test parts
     - Number of splits: How many splits to apply, can only be a divisor of the number of items
 
 ### Options

inst/help/unidimensionalReliabilityBayesian.md

@@ -13,8 +13,8 @@ The Bayesian unidimensional reliability analysis allows the user to test the sca
 ### Scale Statistics
 - The CTT-coefficients alpha, lambda 2, and the split-half coefficient are computed from the data covariance matrix. Coefficient omega is computed from the centered data matrix.
 - Credible interval: default is 95%
-- Coefficient omega: The same as McDonald's omega (for unidimensional data, based on the single-factor model). Note the total test variance in the denominator of the reliability equation is the model implied total variance, that is, the summed model implied covariance matrix.
-- Coefficient alpha: The same as Cronbach's alpha (for binary items the coefficient equals KR20)
+- McDonald's omega (for unidimensional data, based on the single-factor model). Note the total test variance in the denominator of the reliability equation is the model implied total variance, that is, the summed model implied covariance matrix.
+- Cronbach's alpha (for binary items the coefficient equals KR20)
 - Guttman's lambda 2
 - Split-half coefficient: Correlates the sum scores of two test-halves. By default the variables are split into odd and even numbered items in order or appearance in the variables window. If another split is desired the variables just need to be reordered.
 - Average interitem correlation

inst/help/unidimensionalReliabilityFrequentist.md

@@ -12,8 +12,8 @@ The frequentist unidimensional reliability analysis allows the user to test the
 
 ### Scale Statistics
 - Confidence interval: default is 95%
-- Coefficient omega: The same as McDonald's omega (for unidimensional data, based on the single-factor model). Note the total test variance in the denominator of the reliability equation is the model implied total variance, that is, the summed model implied covariance matrix.
-- Coefficient alpha: The same as Cronbach's alpha (for binary items the coefficient equals KR20)
+- McDonald's omega (for unidimensional data, based on the single-factor model). Note the total test variance in the denominator of the reliability equation is the model implied total variance, that is, the summed model implied covariance matrix.
+- Cronbach's alpha (for binary items the coefficient equals KR20)
 - Guttman's lambda 2
 - Split-half coefficient: Correlates the sum scores of two test-halves. By default the variables are split into odd and even numbered items in order or appearance in the variables window. If another split is desired the variables just need to be reordered.
 - Average interitem correlation

inst/qml/StandardErrorOfMeasurement.qml

@@ -37,7 +37,7 @@ Form
 			maxLevels: 50
 			minNumericLevels: 2
 			id: variables
-			onCountChanged: 
+			onCountChanged:
 			{
 				var newValues = []
 				for (var i = 2; i <= variables.count; i++)
@@ -81,11 +81,11 @@ Form
 			name:   	"feldt"
 			label:  	qsTr("Feldt")
 			id: feldt
-			DropDown 
+			DropDown
 			{
 				name: "feldtNumberOfSplits"
 				label: qsTr("Number of splits")
-				values: []
+				values: [2, 3, 4, 5, 6, 7, 8, 9, 10]
 				id: feldtNumberOfSplits
 			}
 		}
@@ -94,11 +94,11 @@ Form
 			name:   	"mollenkopfFeldt"
 			label:  	qsTr("Mollenkopf-Feldt")
 			id: mollenkopfFeldt
-			DropDown 
+			DropDown
 			{
 				name: "mollenkopfFeldtNumberOfSplits"
 				label: qsTr("Number of splits")
-				values: []
+				values: [2, 3, 4, 5, 6, 7, 8, 9, 10]
 				id: mollenkopfFeldtNumberOfSplits
 			}
 			IntegerField
@@ -120,7 +120,7 @@ Form
 			name:   	"anova"
 			label:  	qsTr("ANOVA")
 		}
-		
+
 		CheckBox
 		{
 			name:   	"irt"
@@ -148,13 +148,13 @@ Form
 		{
 			enabled: !variables.columnsTypes.includes("ordinal")
 			name:   	"lord2"
-			label:  	qsTr("Lord's compound")
+			label:  	qsTr("Lord generalized")
 			id: lord2
-			DropDown 
+			DropDown
 			{
 				name: "lord2NumberOfSplits"
 				label: qsTr("Number of splits")
-				values: []
+				values: [2, 3, 4, 5, 6, 7, 8, 9, 10]
 				id: lord2NumberOfSplits
 			}
 		}
@@ -164,7 +164,7 @@ Form
 	{
 		title: qsTr("Options")
 
-		CheckBox 
+		CheckBox
 		{
 			name: "sumScoreCiTable"
 			label: qsTr("Sum score table")
@@ -177,7 +177,7 @@ Form
 			}
 		}
 
-		CheckBox 
+		CheckBox
 		{
 			name: "userReliability"
 			label: qsTr("User defined reliability")
@@ -186,7 +186,7 @@ Form
 			DoubleField
 			{
 				name: "reliabilityValue"
-				label: "" 
+				label: ""
 				max: 1
 				defaultValue: .5
 			}
@@ -195,17 +195,17 @@ Form
 		IntegerField
 		{
 			name: "minimumGroupSize"
-			label: qsTr("Minimum number of observations per score group") 
+			label: qsTr("Minimum number of observations per score group")
 			min: 1
 			defaultValue: 20
 		}
 
-		CheckBox 
+		CheckBox
 		{
 			name: "hideTable"
 			label: qsTr("Hide SEM table")
 		}
-		
+
 	}
 
 	Section
@@ -216,7 +216,7 @@ Form
 		{
 			name: "histogramCounts"
 			label: qsTr("Histogram of sum score counts")
-		}	
+		}
 
 		CheckBox
 		{