Update Data Sets & IMM Vignette

I have updated the data sets to include factor codings for some variables and made sure that the dependent variables ranges from -pi to pi.

R/data.R

@@ -10,7 +10,7 @@
 #'   \item{trial}{Trial identifyier}
 #'   \item{setsize}{The setsize of the data in this row}
 #'   \item{RespErr}{The response error, that is the difference between the reponse
-#'   given and the target color.}
+#'   given and the target color in radians.}
 #'   \item{Pos_Lure1, Pos_Lure2, Pos_Lure3, Pos_Lure4, Pos_Lure5}{Position of the lure items relative to the target color.}
 #'
 #' }
@@ -25,7 +25,7 @@
 #' with set size 1 to 8 reported by Oberauer & Lin (2017).
 #'
 #' @format ## `OberauerLin_2017`
-#' A data frame with 4,000 rows and 9 columns:
+#' A data frame with 15,200 rows and 39 columns:
 #' \describe{
 #'   \item{ID}{Integer uniquely identifying different subjects}
 #'   \item{Session}{Session number}

vignettes/IMM.Rmd

@@ -160,17 +160,17 @@ simData$cond <- as.factor(simData$cond)
 To estimate the `IMM` we first need to specify a formula. In this formula, we directly estimate all parameters for each of the four conditions:
 
 ```{r}
-model_formula <- brms::bf(respErr ~ 1,
-                          c ~ 0 + cond,
-                          a ~ 0 + cond,
-                          s ~ 0 + cond,
-                          kappa ~ 0 + cond)
+model_formula <- bmm_formula(c ~ 0 + cond,
+                             a ~ 0 + cond,
+                             s ~ 0 + cond,
+                             kappa ~ 0 + cond)
 ```
 
 Then, we can specify the model that we want to estimate:
 
 ```{r}
-model <- IMMfull(non_target = paste0("color_item",2:setsize),
+model <- IMMfull(respErr = "respErr",
+                 non_target = paste0("color_item",2:setsize),
                  setsize = setsize,
                  spaPos = paste0("dist_item",2:setsize))
 ```
@@ -185,7 +185,8 @@ model = model,
 parallel = TRUE,
 chains = 4,
 iter = 2000,
-backend = "cmdstanr"
+backend = "cmdstanr",
+refresh = 100
 )
 ```
 
@@ -217,10 +218,10 @@ fixedFX[startsWith(rownames(fixedFX),"c_"),]
 fixedFX[startsWith(rownames(fixedFX),"a_"),]
 
 # print posterior means for the s parameter
-exp(fixedFX[grepl("s_",rownames(fixedFX)),])
+exp(fixedFX[startsWith(rownames(fixedFX),"s_"),])
 
 # print posterior means for the s parameter
-exp(fixedFX[grepl("kappa_",rownames(fixedFX)),])
+exp(fixedFX[startsWith(rownames(fixedFX),"kappa_"),])
 ```
 
 These results indicate that all parameters, except for `s` were well recovered. As already noted by @oberauerHierarchicalBayesianMeasurement2017, a good recovery of the generalization gradient `s` requires a lot of data. Thus you might consider opting for the simplified version of the IMM without the `s` parameter, the `IMMabc`.