Updated changelog, vignette, tests for fSquared and minor changes to fSquared function. (#124)

Author: NicholasDanks

CHANGELOG.md

@@ -18,11 +18,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - A test for the bootstrap summary return object
 - Descriptive statistics for item and construct data
 - S3 print method for class "table_output" for printing generic tables
+- A fSquare function to calculating fSquared
+- A test for fSquared function
 
 ### Changed
 - Fixtures for evaluating bootstrap HTMT for versions of R < 3.6.0
 - Changed the R/* file naming to R/estimate_ R/feature_ R/evaluate_ etc.
 - Summary S3 method to return data descriptives in summary object
+- Changed references to include Cohen (2013)
+- Updated vignette to reflect fSquare function
 
 ### Fixed
 - Modified calculation of HTMT to use absolute correlation matrices in order to make HTMT stable

NAMESPACE

@@ -13,6 +13,7 @@ export(confidence_interval)
 export(constructs)
 export(correlation_weights)
 export(estimate_pls)
+export(fSquared)
 export(interaction_2stage)
 export(interaction_ortho)
 export(interaction_scaled)

R/evaluate_effects.R

@@ -37,19 +37,18 @@
 #'
 #' fSquared(mobi_pls, "Image", "Satisfaction")
 #' @export
-# fSquared as per Cohen (2013)
-
 fSquared <- function(seminr_model, iv, dv) {
   with_sm <- seminr_model$smMatrix
   without_sm <- subset(with_sm, !((with_sm[, "source"] == iv) & (with_sm[, "target"] == dv)))
   capture.output(
-    without_pls <- estimate_pls(data = seminr_model$data,
+    without_pls <- estimate_pls(data = seminr_model$rawdata,
                                 measurement_model = seminr_model$mmMatrix,
+                                interactions = seminr_model$interactions,
                                 structural_model = without_sm)
   )
-
   with_r2 <- seminr_model$rSquared["Rsq", dv]
-  without_r2 <- without_pls$rSquared["Rsq", dv]
-
-  (with_r2 - without_r2) / (1 - with_r2)
+  ifelse(any(without_sm[,"target"] == dv),
+         without_r2 <- without_pls$rSquared["Rsq", dv],
+         without_r2 <- 0)
+  return((with_r2 - without_r2) / (1 - with_r2))
 }

man/fSquared.Rd

@@ -57,6 +57,6 @@
   ref: Sarstedt, M., Hair, J. F., Ringle, C. M., Thiele, K. O., & Gudergan, S. P. (2016). Estimation issues with PLS and CBSEM: Where the bias lies! Journal of Business Research, 69(10), 3998–4010
   cite: Sarstedt et al. (2016)
 - code: [short code here]
-  ref: [full academic reference here]
-  cite: [preferred citation format]
+  ref: Cohen, J. (2013). Statistical power analysis for the behavioral sciences. Routledge.
+  cite: Cohen (2013)
 

references.md

@@ -0,0 +1,2 @@
+"","x"
+"1",0.303554364141901

tests/fixtures/V_3_5_X/fsquared1.csv

@@ -0,0 +1,2 @@
+"","x"
+"1",0.303554364141901

tests/fixtures/V_3_6_0/fsquared1.csv

@@ -22,20 +22,26 @@ mobi_sm <- relationships(
 mobi <- mobi
 seminr_model <- estimate_pls(mobi, mobi_mm, interactions = NULL ,mobi_sm)
 
-
 # Load outputs
 rsquared <- seminr_model$rSquared
+fsquared <- fSquared(seminr_model, "Image", "Satisfaction")
 
 ## Output originally created using following lines
 # write.csv(seminr_model$rSquared, file = "tests/fixtures/rsquared1.csv")
+# write.csv(fSquared(seminr_model, "Image", "Satisfaction"), file = "tests/fixtures/V_3_5_X/fsquared1.csv")
+# write.csv(fSquared(seminr_model, "Image", "Satisfaction"), file = "tests/fixtures/V_3_6_0/fsquared1.csv")
+
 
 # Load controls
 rsquared_control <- as.matrix(read.csv(file = paste(test_folder,"rsquared1.csv", sep = ""), row.names = 1))
+fsquared_control <- as.matrix(read.csv(file = paste(test_folder,"fsquared1.csv", sep = ""), row.names = 1))
 
 # Testing
 
 test_that("Seminr estimates the Rsquared correctly", {
-  # Remove BIC for now
   expect_equal(rsquared[1:2,1], rsquared_control[1:2,1], tolerance = 0.00001)
 })
 
+test_that("Seminr estimates the fSquared correctly", {
+  expect_equal(fsquared, fsquared_control[1,1], tolerance = 0.00001)
+})

tests/testthat/test-metric.R

@@ -409,6 +409,12 @@ summary(mobi_pls)
 
 Please note that common-factor scores are indeterminable and therefore construct scores for common factors are not reported (Hair et al.,2011).
 
+3. The fSquare of specific relationships can be determine using the `fSquared()` function as per Cohen (2013). 
+
+```{r}
+fSquared(mobi_pls, "Image", "Satisfaction")
+```
+
 ### Reporting the bootstrapped `boot_seminr_model` 
 
 As with the estimated model, there are multiple ways of reporting the bootstrapped model. The `bootstrap_model()` function returns an object of class `boot_seminr_model`. This can be passed directly to the base R function `summary()`. This can be used in two primary ways:   
@@ -500,6 +506,7 @@ model_summary$descriptives$correlations$constructs
 ## References
 
 * Becker et al. (2018). Estimating Moderating Effects in PLS-SEM and PLSc-SEM: Interaction Term Generation*Data Treatment
+* Cohen, J. (2013). Statistical power analysis for the behavioral sciences. Routledge.
 * Dijkstra, T. K., & Henseler, J. (2015). Consistent Partial Least Squares Path Modeling, MIS Quarterly Vol. 39(X).
 * Dillon, W. R, and M. Goldstein. 1987. Multivariate Analysis: Methods, and Applications. Biometrical Journal 29 (6): 750–756.
 * Fornell, C. and D. F. Larcker (February 1981). Evaluating structural equation models with unobservable variables and measurement error, Journal of Marketing Research, 18, pp. 39-5)