Data Source

Description: https://raw.githubusercontent.com/vincentarelbundock/Rdatasets/master/doc/wooldridge/meap93.html

CSV: https://raw.githubusercontent.com/vincentarelbundock/Rdatasets/master/csv/wooldridge/meap93.csv

Description of Data

MEAP stands for Michigan Education Assessment Program.

This dataset has 408 observations and 17 variables.

lnchprg. perc of studs in sch lnch prog ***
enroll. school enrollment
staff. staff per 1000 students
expend. expend. per stud, $ ***
salary. avg. teacher salary, $
benefits. avg. teacher benefits, $
droprate. school dropout rate, perc
gradrate. school graduation rate, perc
math10. perc studs passing MEAP math ***
sci11. perc studs passing MEAP science ***
totcomp. salary + benefits
ltotcomp. log(totcomp)
lexpend. log of expend
lenroll. log(enroll)
lstaff. log(staff)
bensal. benefits/salary
lsalary. log(salary)

Objective of Analysis

The object of this analysis is to explore relationship between:

academic performance and SES
academic performance and expenditure per student
SES and expenditure per student

408 observations
lnchprg - percent of student in school lunch program
math10 - percent of students passing the MEAP math test
sci11 - percent of students passing the MEAP science test
expend - school expenditure per student
academic perforamnce is going to be measured by taking the average of math10 and sci11
lnchprg is proxy for SES
school SES profile is going to be created based on lnchprg (proxy for SES)

Background

Source: http://facultyprofiles.ucr.edu/gsoe_dept/faculty/Robert_Ream/Ream_Ryan_Racial%20Inequality_2013.pdf

Income is the most common way of measuring socioeconomic status (SES).
Schools often do not have complete income information.
Schools with lower SES students tend to receive additional revenue from state and federal governments.
Measures like free or reduced-price lunch only serve as a proxy for lower SES.

Load data in CSV file to a data frame

setwd("C:/Users/stina/Documents/R programming Bridge Workshop/Week 3 Assignment")
# load data onto dataframe
data.frame_meap_all <- read.csv("Meap93Data.csv")

Subset data to include only columns we will be investigating

Subset: Take a look at lnchprg, expend, math10, and sci111

data.frame_meap <- data.frame_meap_all[, c(2,5, 10,11)]

Rename columns

names(data.frame_meap)[1] <- "lunch_program_rate"
names(data.frame_meap)[2] <- "expend_per_student"
names(data.frame_meap)[3] <- "math_pass_rate"
names(data.frame_meap)[4] <- "science_pass_rate"

Check for any “NA” data

NOTE: The result below tells us that there are no NA values

# any checks if there is at least one value that is true
# is.na returns a vectore of logicals. If a value is NA, then it returns TRUE
any(is.na(data.frame_meap[,]))

## [1] FALSE

Check for any values that are ZERO OR NEAGATIVE

any(data.frame_meap$lunch_program_rate <= 0)

## [1] FALSE

any(data.frame_meap$expend_per_student <= 0)

## [1] FALSE

any(data.frame_meap$math_pass_rate <= 0)

## [1] FALSE

any(data.frame_meap$science_pass_rate <= 0)

## [1] FALSE

Check if all the columns are numeric

is.numeric(data.frame_meap$lunch_program_rate)

## [1] TRUE

is.numeric(data.frame_meap$expend_per_student)

## [1] TRUE

is.numeric(data.frame_meap$math_pass_rate)

## [1] TRUE

is.numeric(data.frame_meap$science_pass_rate)

## [1] TRUE

Create new column called “meap_score”

This combines both the math_pass_rate and science_pass_rate.

data.frame_meap$meap_score <- (data.frame_meap$math_pass_rate + data.frame_meap$science_pass_rate)/2

Create new column “school_SES_profile”

The “lunch_program_rate” serves as a proxy for socioeconomic status (SES) profile of the school.

An average lunch program participation corresponds to an average-SES profile school. A high lunch program participation corresponds to a low-SES profile school. A low lunch program participation corresponds to a high-SES profile school.

The average value of lunch_program_rate across the sample is around 25.

The column “school_SES_profile” is going to be a factor that is going to have values: “high_SES”, “average_SES”, and “low_SES”.

high_SES: lunch_program_rate < 18
average_SES: 18 <= lunch_program_rate <= 30
low_SES: lunch_program_rate > 30

#data.frame_meap$school_SES_profile <- "unassigned"
data.frame_meap$school_SES_profile[data.frame_meap$lunch_program_rate < 18] <- "high_SES"
data.frame_meap$school_SES_profile[data.frame_meap$lunch_program_rate > 30] <- "low_SES"
data.frame_meap$school_SES_profile[data.frame_meap$lunch_program >=18 & 
                                     data.frame_meap$lunch_program <= 30] <- "average_SES"

Convert “school_SES_profile” from character to factor

class(data.frame_meap$school_SES_profile)

## [1] "character"

data.frame_meap$school_SES_profile <- as.factor(data.frame_meap$school_SES_profile)
class(data.frame_meap$school_SES_profile)

## [1] "factor"

Create Expenditure Level

data.frame_meap$expend_level[data.frame_meap$expend_per_student <= 3500] <- "Expend Level 1"
data.frame_meap$expend_level[data.frame_meap$expend_per_student > 3500 &
                               data.frame_meap$expend_per_student <= 4500] <- "Expend Level 2"
data.frame_meap$expend_level[data.frame_meap$expend_per_student > 4500 &
                               data.frame_meap$expend_per_student <= 5500] <- "Expend Level 3"
data.frame_meap$expend_level[data.frame_meap$expend_per_student > 5500 &
                               data.frame_meap$expend_per_student <= 6500] <- "Expend Level 4"
data.frame_meap$expend_level[data.frame_meap$expend_per_student > 6500] <- "Expend Level 5"

Convert “expend_level” from character to factor

class(data.frame_meap$expend_level)

## [1] "character"

data.frame_meap$expend_level <- as.factor(data.frame_meap$expend_level)
class(data.frame_meap$expend_level)

## [1] "factor"

Summary of Data

summary(data.frame_meap)

##  lunch_program_rate expend_per_student math_pass_rate  science_pass_rate
##  Min.   : 1.40      Min.   :3332       Min.   : 1.90   Min.   : 7.20    
##  1st Qu.:14.62      1st Qu.:3821       1st Qu.:16.62   1st Qu.:41.30    
##  Median :23.85      Median :4145       Median :23.40   Median :49.10    
##  Mean   :25.20      Mean   :4377       Mean   :24.11   Mean   :49.18    
##  3rd Qu.:33.83      3rd Qu.:4659       3rd Qu.:30.05   3rd Qu.:57.15    
##  Max.   :79.50      Max.   :7419       Max.   :66.70   Max.   :85.70    
##    meap_score      school_SES_profile         expend_level
##  Min.   : 4.95   average_SES:135      Expend Level 1: 10  
##  1st Qu.:30.75   high_SES   :136      Expend Level 2:273  
##  Median :36.25   low_SES    :137      Expend Level 3: 83  
##  Mean   :36.64                        Expend Level 4: 27  
##  3rd Qu.:42.31                        Expend Level 5: 15  
##  Max.   :65.80

Summary of Data by School SES Profile

# summary by grouping with school_SES_profile
library(purrr)
data.frame_meap %>% split(.$school_SES_profile) %>% map(summary)

## $average_SES
##  lunch_program_rate expend_per_student math_pass_rate  science_pass_rate
##  Min.   :18.00      Min.   :3425       Min.   : 4.30   Min.   :22.10    
##  1st Qu.:20.15      1st Qu.:3770       1st Qu.:16.70   1st Qu.:41.15    
##  Median :23.80      Median :4003       Median :22.50   Median :47.60    
##  Mean   :23.78      Mean   :4209       Mean   :23.77   Mean   :48.79    
##  3rd Qu.:27.30      3rd Qu.:4381       3rd Qu.:27.75   3rd Qu.:56.20    
##  Max.   :29.90      Max.   :7419       Max.   :60.60   Max.   :80.00    
##    meap_score      school_SES_profile         expend_level
##  Min.   :17.65   average_SES:135      Expend Level 1:  5  
##  1st Qu.:30.48   high_SES   :  0      Expend Level 2:100  
##  Median :35.30   low_SES    :  0      Expend Level 3: 23  
##  Mean   :36.28                        Expend Level 4:  4  
##  3rd Qu.:41.38                        Expend Level 5:  3  
##  Max.   :62.05                                            
## 
## $high_SES
##  lunch_program_rate expend_per_student math_pass_rate  science_pass_rate
##  Min.   : 1.400     Min.   :3332       Min.   : 2.40   Min.   :30.10    
##  1st Qu.: 8.475     1st Qu.:4075       1st Qu.:21.45   1st Qu.:48.25    
##  Median :11.050     Median :4355       Median :26.80   Median :53.75    
##  Mean   :11.071     Mean   :4634       Mean   :28.27   Mean   :54.14    
##  3rd Qu.:14.475     3rd Qu.:5030       3rd Qu.:33.23   3rd Qu.:60.60    
##  Max.   :17.800     Max.   :7034       Max.   :62.40   Max.   :74.60    
##    meap_score      school_SES_profile         expend_level
##  Min.   :24.55   average_SES:  0      Expend Level 1: 2   
##  1st Qu.:35.89   high_SES   :136      Expend Level 2:78   
##  Median :40.60   low_SES    :  0      Expend Level 3:34   
##  Mean   :41.21                        Expend Level 4:13   
##  3rd Qu.:45.26                        Expend Level 5: 9   
##  Max.   :64.55                                            
## 
## $low_SES
##  lunch_program_rate expend_per_student math_pass_rate  science_pass_rate
##  Min.   :30.40      Min.   :3396       Min.   : 1.90   Min.   : 7.20    
##  1st Qu.:33.80      1st Qu.:3751       1st Qu.:14.00   1st Qu.:36.40    
##  Median :38.60      Median :4042       Median :18.80   Median :44.60    
##  Mean   :40.63      Mean   :4286       Mean   :20.31   Mean   :44.65    
##  3rd Qu.:45.10      3rd Qu.:4594       3rd Qu.:26.10   3rd Qu.:53.60    
##  Max.   :79.50      Max.   :6641       Max.   :66.70   Max.   :85.70    
##    meap_score      school_SES_profile         expend_level
##  Min.   : 4.95   average_SES:  0      Expend Level 1: 3   
##  1st Qu.:26.65   high_SES   :  0      Expend Level 2:95   
##  Median :32.30   low_SES    :137      Expend Level 3:26   
##  Mean   :32.48                        Expend Level 4:10   
##  3rd Qu.:38.20                        Expend Level 5: 3   
##  Max.   :65.80

Histogram

Lunch Program Participation

x <- data.frame_meap$lunch_program_rate
hist(x, breaks=50, col="blue", xlab="Lunch Program", main="Percent of Students in Lunch Program")

Histogram

MEAP Score

x <- data.frame_meap$meap_score
hist(x, breaks=50, col="blue", xlab="MEAP Score", main="School Academic Perfomance")

Histogram

Expenditure Per Student

x <- data.frame_meap$expend_per_student
hist(x, breaks=50, col="blue", xlab="Expenditure Per Student", main="Expenditure Per Student")

Expenditure Per Student and School SES Profile

library(ggplot2)
ggplot(data=data.frame_meap, aes(x=expend_per_student)) + 
  geom_histogram(binwidth=1000, aes(fill=school_SES_profile), color="Gray") +
  labs(x="Expenditure Per Student", y="Frequency")

Scatter Plot

MEAP Score and Lunch Program Participation

A LOWESS (Locally Weighted Scatterplot Smoothing) line is added on the scatter plot.
This is a smooth line through a timeplot or scatter plot to help you to see relationship between variables and foresee trends.

plot(data.frame_meap$lunch_program_rate, data.frame_meap$meap_score, main = "School Academic Performance (MEAP Score)", 
     xlab = "Percent of Student in Lunch Program", ylab = "MEAP Score", pch=19)

#Create a LOWESS (Locally Weighted Scatterplot Smoothing)line
lines(lowess(data.frame_meap$lunch_program_rate, data.frame_meap$meap_score), 
      col="blue") # lowess line (x,y)

Scatter Plot

Meap Score and Lunch Program Participation with Expenditure Levels

Expend Level 1: expend_per_student <= 3500
Expend Level 2: 3500 < expend_per_student <= 4500
Expend Level 3: 4500 < expend_per_student <= 5500
Expend Level 4: 5500 < expend_per_student <= 6500
Expend Level 5: expend_per_student > 6500

library(ggplot2)

ggplot(data=data.frame_meap, aes(x=lunch_program_rate, y=meap_score,
                        colour=expend_level)) + 
  geom_point(size=3, alpha=0.7) +
  labs(y="Meap Score", x="Percent of Student in Lunch Program", color="Expenditure Levels")

Box Plot

MEAP Score and School SES Profile

# specify order of school SES profile (factor) on the box plot
data.frame_meap$school_SES_profile <- factor(data.frame_meap$school_SES_profile, levels=c("low_SES", "average_SES", "high_SES"))

# create box plot  
boxplot(meap_score~school_SES_profile, data=data.frame_meap, main="School Academic Performance (MEAP Score)",
        xlab = "School SES Profile", ylab="MEAP Score")

Scatter Plot

Expenditure Per Student and MEAP Score

plot(data.frame_meap$expend_per_student, data.frame_meap$meap_score, main = "School Academic Performance (MEAP Score)", 
     xlab = "Expenditure per Student", ylab = "MEAP Score", pch=19)

#Create a LOWESS (Locally Weighted Scatterplot Smoothing) line
lines(lowess(data.frame_meap$expend_per_student, data.frame_meap$meap_score), 
      col="blue") # lowess line (x,y)

Box Plot

Expenditure Per Student by School SES Profile

# specify order of school SES profile (factor) on the box plot
data.frame_meap$school_SES_profile <- factor(data.frame_meap$school_SES_profile, levels=c("low_SES", "average_SES", "high_SES"))

# create box plot  
boxplot(expend_per_student~school_SES_profile, data=data.frame_meap, main="Expenditure Per Student",
        xlab = "School SES Profile", ylab="Expenditure Per Student")

Scatter Plot

Expenditure Per Student MEAP Score, School SES Profile

library(ggplot2)

ggplot(data=data.frame_meap, aes(x=expend_per_student, y=meap_score,
                        colour=school_SES_profile)) + 
  geom_point(size=3, alpha=0.7) +
  labs(y="Meap Score", x="Expenditure Per Student", color="School SES Profile")

Conclusion

meap_score and lunch_program_rate

There appears to be an inverse relationship between meap_score and lunch_program_rate.
As the lunch_program_rate increases, meap_score tends to decrease.
This relationship suggests that higher SES schools tend to do better academically.

meap_score and expenditure_per_student

There appears to be a direct relationship between meap_score and expenditure_per_student.
As the expenditure_per_student increases, meap_score tends to increase as well.
This relationship suggets that schools that spend more per student tend to do better academically.

school_SES_profile and expenditure_per_student

The mean expenditure per student for the high SES schools is higher compared to both low and average SES schools.

R Final Project - Week 3

Sherranette Tinapunan

January 18, 2018

Data Source

Description of Data

MEAP stands for Michigan Education Assessment Program.

Objective of Analysis

Background

Load data in CSV file to a data frame

Subset data to include only columns we will be investigating

Rename columns

Check for any “NA” data

Check for any values that are ZERO OR NEAGATIVE

Check if all the columns are numeric

Create new column called “meap_score”

Create new column “school_SES_profile”

Convert “school_SES_profile” from character to factor

Create Expenditure Level

Convert “expend_level” from character to factor

Summary of Data

Summary of Data by School SES Profile

Histogram

Lunch Program Participation

Histogram

MEAP Score

Histogram

Expenditure Per Student

Expenditure Per Student and School SES Profile

Scatter Plot

MEAP Score and Lunch Program Participation

Scatter Plot

Meap Score and Lunch Program Participation with Expenditure Levels

Box Plot

MEAP Score and School SES Profile

Scatter Plot

Expenditure Per Student and MEAP Score

Box Plot

Expenditure Per Student by School SES Profile

Scatter Plot

Expenditure Per Student MEAP Score, School SES Profile

Conclusion

meap_score and lunch_program_rate

meap_score and expenditure_per_student

school_SES_profile and expenditure_per_student