SPSS Modeler Menu-Driven Commands for Regression Trees

*Outcome: Fall Math Performance
*Build the regression tree with 1 level using the first cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV1R" as “partition”
Modeling->CHAID-Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”> Choose “Basics” and specify the tree depth "1"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 1 level using the second cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV2R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "1"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 1 level using the third cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV3R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "1"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 1 level using the fourth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV4R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "1"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 1 level using the fifth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV5R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "1"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Compute the mean absolute error across the five data sets separately for the training dataset and the testing dataset for the 1 level regression tree. 

*Build the regression tree with 2 level using the first cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV1R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 2 level using the second cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV2R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 2 level using the third cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV3R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 2 level using the fourth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV4R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 2 level using the fifth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV5R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Compute the mean absolute error across the five data sets separately for the training dataset and the testing dataset for the 2 level regression tree. 

*Build the regression tree with 3 level using the first cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV1R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 3 level using the second cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV2R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 3 level using the third cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV3R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 3 level using the fourth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV4R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 3 level using the fifth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV5R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Compute the mean absolute error across the five data sets separately for the training dataset and the testing dataset for the 3 level regression tree. 

*Build the regression tree with 4 level using the first cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV1R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "4"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 4 level using the second cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV2R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "4"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 4 level using the third cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV3R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "4"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 4 level using the fourth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV4R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "4"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 4 level using the fifth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV5R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "4"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Compute the mean absolute error across the five data sets separately for the training dataset and the testing dataset for the 4 level regression tree. 

*Build the regression tree with 5 level using the first cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV1R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "5"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 5 level using the second cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV2R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "5"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 5 level using the third cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV3R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "5"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 5 level using the fourth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV4R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "5"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 5 level using the fifth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV5R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "5"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Compute the mean absolute error across the five data sets separately for the training dataset and the testing dataset for the 5 level regression tree. 

*Build the regression tree with 6 level using the first cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV1R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "6"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 6 level using the second cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV2R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "6"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 6 level using the third cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV3R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "6"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 6 level using the fourth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV4R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "6"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Build the regression tree with 6 level using the fifth cross-validation data set
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” and assign the cross-validation indicator "CV5R" as “partition”
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "6"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
Output->Analysis->Choose “Find predicted/predictor fields using model output field metadata” and “separate by partition” to obtain the mean absolute error in the training dataset and the test dataset. 
Graph->choose “cumulative plot”, “include baseline”, and “include best line” to get the evaluation plot. 

*Compute the mean absolute error across the five data sets separately for the training dataset and the testing dataset for the 6 level regression tree. 

*Compare the mean absolute errors from the 1-6 level regression trees and choose the level of regression tree with the smallest mean absolute error for the testing set. 

*Build the reported regression tree using the whole fall math performance dataset with the identified level (i.e., 2 level for fall math performance based on the original data)
Sources->Statistics File->Open the data file "X1MathSMFinal"
Field Ops->Type->Choose the outcome variable "X1Math" as “target” 
Modeling-> CHAID->Choose the target variable "X1Math" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"

*The above procedures were repeated for each of the outcome variables below: Spring Math Performance, Math Gain, Fall Reading Performance, Spring Reading Performance, and Reading Gain. 
* Only the commands for the final tree models for these outcome variables are reported below. 
*Outcome: Spring Math Performance
Sources->Statistics File->Open the data file "X2MathSMFinal"
Field Ops->Type->Choose the outcome variable "X2Math" as “target” 
Modeling-> CHAID->Choose the target variable "X2Math" and the predictors "X1Math" "X2Region" "X2Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"

*Outcome: Math Gain
Sources->Statistics File->Open the data file "X2MathSMFinal"
Field Ops->Type->Choose the outcome variable "MathGain" as “target” 
Modeling-> CHAID->Choose the target variable "MathGain" and the predictors "X2Region" "X2Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"

*Outcome: Fall Reading Performance
Sources->Statistics File->Open the data file "X1ReadingSMFinal"
Field Ops->Type->Choose the outcome variable "X1Reading" as “target” 
Modeling-> CHAID->Choose the target variable "X1Reading" and the predictors "X1Region" "X1Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"

*Outcome: Spring Reading Performance
Sources->Statistics File->Open the data file "X2ReadingSMFinal"
Field Ops->Type->Choose the outcome variable "X2Reading" as “target” 
Modeling-> CHAID->Choose the target variable "X2Reading" and the predictors "X1Reading" "X2Region" "X2Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "2"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"

*Outcome: Math Gain
Sources->Statistics File->Open the data file "X2MathSMFinal"
Field Ops->Type->Choose the outcome variable "ReadingGain" as “target” 
Modeling-> CHAID->Choose the target variable "ReadingGain" and the predictors "X2Region" "X2Locale" "Gender" "Race" "SES" "FRPL"
->Choose “Build Options”-> Choose “Basics” and specify the tree depth "3"
->Choose “stopping rules” and “use percentage” enter “0.83” % for the minimum records in parent branch and “0.42” % for the minimum records in child branch
->Choose "Run"
