nmm (version 0.9)

dataM: Example dataset

Description

Data "MathPlacement" taken from Stat2Data package.

Usage

data(dataM)

Format

A data frame containing:

Student

Identification number for each student

Gender

0=Female, 1=Male

PSATM

PSAT score in Math

SATM

SAT score in Math

ACTM

ACT Score in Math

Rank

Adjusted rank in HS class

Size

Number of students in HS class

GPAadj

Adjusted GPA

PlcmtScore

Score on math placement exam

Recommends

Recommended course: R0 R01 R1 R12 R2 R3 R4 R6 R8

Course

Actual course taken

Grade

Course grade

RecTaken

1=recommended course, 0=otherwise

TooHigh

1=took course above recommended, 0=otherwise

TooLow

1=took course below recommended, 0=otherwise

CourseSuccess

1=B or better grade, 0=grade below B

DR_Course

according to recommendations, which level of course was taken: alow - lower, bnormal - recommended, chigh - higher

Details

Code for data modifications can be found in the example section.

Examples

# NOT RUN {
data(dataM)
library(magrittr)
library(dplyr)
if (requireNamespace("recipes", quietly = TRUE)&requireNamespace("Stat2Data", quietly = TRUE)) {
data("MathPlacement", package="Stat2Data")
head(MathPlacement)
library(recipes)
# As some of the data is missing, k-nearest neighbors (knn) imputation is 
# used to fill the gaps. This is done with recipes package and function 
# step_knnimpute.
dataM <- recipe(~ ., data = MathPlacement) %>%
step_knnimpute(everything()) %>% prep() %>% juice()
# Afterwards we create a categorical variable that will show whether a 
# student took a course which was too high, too low, the recommended one or
# something else happened:
dataM %<>% mutate(Student = 1:n(), DR_Course = case_when(
TooHigh == 1 ~ "chigh",
TooLow == 1 ~ "alow",
RecTaken == 1 ~ "bnormal",
TRUE ~"dother"
)) 
# We remove observations with ambiguous course status:
dataM %<>% filter(DR_Course!="dother")
dataM %>% select(DR_Course) %>% table %>% t 
}
# }