Mini Project – Par Inc.

New Golf Ball

Model Report
Table of Contents
1 Project Objective ............................................................................... 3
2 Assumptions ..................................................................................... 4
3 Step by step approach ........................................................................ 4
3.1 Exploratory Data Analysis .............................................................. 4
3.1.1 Number of Rows and Columns:.................................................. 4
3.1.2 Features and their Types: ......................................................... 4
3.1.3 Check for Missing Values: ......................................................... 4
3.1.4 Dataset Summary: .................................................................. 4
3.2 Descriptive Statistics ..................................................................... 5
3.2.1 Measures of Central Tendency:.................................................. 5
3.2.2 Measures of Dispersion: ........................................................... 6
3.3 Data Visualization – Histogram and Boxplot: ..................................... 7
3.3.1 Key Observations: ................................................................... 7
3.4 Hypothesis formation .................................................................... 7
3.5 Hypothesis Testing Method............................................................. 8
3.6 Confidence Interval ....................................................................... 9
3.7 Need of Larger Sample Size: ........................................................ 10
4 Conclusion and Recommendation ....................................................... 12
5 Appendix – Source Code ................................................................... 13
5.1 Exploratory Data Analysis ............................................................ 13
5.2 Descriptive Statistics ................................................................... 14
1 Project Objective

Par Inc., is a major manufacturer of golf equipment. Management believes

that Par’s market share could be increased with the introduction of a cut-
resistant, longer-lasting golf ball. Therefore, the research group at Par has
been investigating a new golf ball coating designed to resist cuts and provide
a more durable ball. The tests with the coating have been promising. One of
the researchers voiced concern about the effect of the new coating on driving
distances. Par would like the new cut-resistant ball to offer driving distances
comparable to those of the current-model golf ball. To compare the driving
distances for the two balls, 40 balls of both the new and current models were
subjected to distance tests. The testing was performed with a mechanical
hitting machine so that any difference between the mean distances for the
two models could be attributed to a difference in the design.

The results of the tests, with distances measured to the nearest yard, are
contained in the data set “Golf”. Prepare a Managerial Report

1. Formulate and present the rationale for a hypothesis test that par could
use to compare the driving distances of the current and new golf balls
2. Analyze the data to provide the hypothesis testing conclusion. What is the
p-value for your test? What is your recommendation for Par Inc.?
3. Provide descriptive statistical summaries of the data for each model
4. What is the 95% confidence interval for the population mean of each
model, and what is the 95% confidence interval for the difference
between the means of the two population?
5. Do you see a need for larger sample sizes and more testing with the golf
balls? Discuss.

The Golf Dataset:

Sr. Sr. Sr.

Current New Current New Current New
No. No. No.
1 264 277 15 260 260 29 278 268
2 261 269 16 283 281 30 275 262
3 267 263 17 255 250 31 281 283
4 272 266 18 272 263 32 274 250
5 258 262 19 266 278 33 273 253
6 283 251 20 268 264 34 263 260
7 258 262 21 270 272 35 275 270
8 266 289 22 287 259 36 267 263
9 259 286 23 289 264 37 279 261
10 270 264 24 280 280 38 274 255
11 263 274 25 272 274 39 276 263
12 264 266 26 275 281 40 262 279
13 284 262 27 265 276
14 263 271 28 260 269

3| Page
2 Assumptions
The sample size of the data set is 40 (> 30) from each model of golf ball.
Central Limit Theorem states that irrespective of the shape of the original
population, the sampling distribution of the mean will approach a normal
distribution as the size of the sample increases and becomes large (>30).
We also assume that the sample estimate will be reflective of the reality.

3 Step by step approach

We shall follow step by step approach to arrive to the conclusion as follows:

1. Exploratory Data Analysis

2. Descriptive Statistics
3. Data Visualization
4. Hypothesis formation
5. Selection of appropriate Hypothesis Testing method
6. 95% Confidence Intervals
7. Need of Larger Sample Size
8. Conclusion and Recommendation.

3.1 Exploratory Data Analysis

Please refer Appendix Section 5.1 for related code.

3.1.1 Number of Rows and Columns:

• The number of rows in the dataset is 40
• The number of columns (Features) in the dataset is 2

3.1.2 Features and their Types:

• Both the features, i.e. Current and New are continuous variables.

Feature Code Type Continuous/

Categorical
Current Integer Continuous
New Integer Continuous

3.1.3 Check for Missing Values:

• The data was checked for Missing Values using R function
colsums(is.na), and found no missing values.

3.1.4 Dataset Summary:

• Current: The minimum value is 255 yards and the maximum value is
289 yards. The average value is 270.3 yards
• New: The minimum value is 250 yards and the maximum value is 289
yards. The average value is 267.5 yards

4| Page
Feature Minimum Maximum Average Value
Code Value Value
Current 255 yards 289 yards 270.3 yards
New 250 yards 289 yards 267.5 yards

3.2 Descriptive Statistics

Please refer Appendix Section 5.2 for related code.

In this step, the features are explored in detail. The goal is to describe or
summarize data in ways that are meaningful and useful for insights
generation. It provides simple summaries about the sample and the
measures. Together with simple graphics analysis, it forms the basis of
virtually every quantitative analysis of data.

Given both the features – ‘Current’ and ‘New’ are continuous in nature; the
following measures are relevant to understand the central tendency and
spread of the variable.

Measures of Measures of Dispersion Visualization Method

Central Tendency
Mean Range Histogram
Median 1st Quartile Boxplot
Mode 3rd Quartile
Minimum Inter Quartile Range
(IQR)
Maximum Variance
Standard Deviation

3.2.1 Measures of Central Tendency:

Measures of Central
Current New
Tendency
Mean 270.3 267.5
Median 270.0 265.0
Mode 272.0 263.0
Minimum 255.0 250.0
Maximum 289.0 289.0

5| Page
3.2.2 Measures of Dispersion:
Measures of
Current New
Dispersion
Range 34.0 39.0
1st Quartile 263.0 262.0
3rd Quartile 275.2 274.5
Inter Quartile Range
12.2 12.5
(IQR)
Variance 76.6 97.9
Standard Deviation 8.8 9.9

The data for ‘Range’, ‘1st Quartile’ & ‘3rd Quartile’ have been obtained
using the R function ‘summary’ (The output is provided in the section 9.1
Measures of Central Tendency).

The data for ‘Inter Quartile Range (IQR)’ has been computed and is the
difference in value between 3rd Quartile (75 percentile) & 1 st Quartile (25
percentile).

The variances have been obtained using R function ‘var’:

The standard deviations have been obtained using R function ‘sd’:

6| Page
3.3 Data Visualization – Histogram and Boxplot:

3.3.1 Key Observations:

• The average distance covered by ‘New’ golf ball (Mean = 267.5) is
lower as compared to ‘Current’ golf ball (Mean = 270.3)
• ‘New’ golf ball observed to have relatively higher variation in the data
distribution in comparison to ‘Current’ golf ball
• No outliers were observed in the data range for Current & New data
sets.

3.4 Hypothesis formation

For the hypothesis formulation, we have to define the Null Hypothesis &
Alternative Hypothesis.

• Null Hypothesis
o It is a hypothesis that says there is no statistical significance
between the two variables. The null hypothesis is formulated such
that the rejection of the null hypothesis proves the alternative
hypothesis is true
• Alternative Hypothesis
7| Page
o It is one that states there is a statistically significant relationship
between two variables. The alternative hypothesis is the hypothesis
used in hypothesis testing that is contrary to the null hypothesis

In Par Inc. case, the management would like to produce the new golf balls once
it is comparable to the current golf balls. A sample of 40 balls of both the current
and new models were tested with a mechanical hitting machine so that any
difference between the mean distances for the two models could be attributed to
a difference in the two models. Therefore, a hypothesis test that Par Inc. could
use to compare the driving distance of the current and new golf balls. The Null
Hypothesis & Alternative Hypothesis is formulated as follow:

Null Hypothesis (H0): µ1 - µ2 = 0 (i.e. they are the same)

Alternative Hypothesis (Ha): µ1 - µ2 ≠ 0 (i.e. they are not the same)

Where,

• µ1: Mean driving distance of current model golf ball

• µ2: Mean driving distance of new golf ball

By formulation of above hypotheses, we assume that the current and new golf
balls show no significant difference to each other.

3.5 Hypothesis Testing Method

Based on the details shared for the Par Inc. project, we can assume the
following:

• One machine
• Two populations
• No other influences considered
• Independently chosen

It seems to be an independent sample case. The two-tailed test will be

applicable for the project.

Let us calculate the p-value using the R function ‘t.test’.

t.test(Current,New)

##
## Welch Two Sample t-test
##
## data: Current and New
## t = 1.3284, df = 76.852, p-value = 0.188
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -1.384937 6.934937
## sample estimates:
## mean of x mean of y
## 270.275 267.500
8| Page
Since it is a two-tailed test, the p-value = 0.188 ÷ 2 = 0.094.

The p-value for the two-tailed test is 0.094, which is greater than level of
significance α (0.05). Therefore, the Null Hypothesis (H0) will not be rejected.
The conclusion is that this data does not provide statistical evidence that the
new golf balls have either a lower mean driving distance or a higher mean
driving distance. This implies that Par Inc. should take the new golf balls in
production as the p-value indicate that there is no significant difference between
estimated population mean of current as well as new golf balls.

3.6 Confidence Interval

Analysis of Current Model:

t.test(Current)

##
## One Sample t-test
##
## data: Current
## t = 195.29, df = 39, p-value < 2.2e-16
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 267.4757 273.0743
## sample estimates:
## mean of x
## 270.275

Inference: The 95% confidence interval of population mean for Current model
is between 267.4757 & 273.0743. This implies that, with 95% confidence, we
can say that the sample mean driving distance of current balls will be within this
range.

Analysis of New Model:

t.test(New)

##
## One Sample t-test
##
## data: New
## t = 170.94, df = 39, p-value < 2.2e-16
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 264.3348 270.6652
## sample estimates:
## mean of x
## 267.5

9| Page
Inference: The 95% confidence interval of population mean for New model is
between 264.3348 & 270.6652. This implies that, with 95% confidence, we can
say that the sample mean driving distance of New balls will be within this range.

95% confidence interval for the difference between the means of the
two population:

t.test(Current,New)

##
## Welch Two Sample t-test
##
## data: Current and New
## t = 1.3284, df = 76.852, p-value = 0.188
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -1.384937 6.934937
## sample estimates:
## mean of x mean of y
## 270.275 267.500

Inference: The 95% confidence interval of difference in population mean

between both the models is between –1.38 yards on the lower end 6.93
yards on the upper end. This implies that, with 95% confidence, we can say
that the difference in sample mean driving distance of both the models will be
within the above range. For the 40 balls we have taken in this sample, the
difference in mean driving distance is -2.775 yards (267.5 – 270.275) which falls
in the range.

3.7 Need of Larger Sample Size:

Steps to follow:

• Get the difference between two sample means (2.775 as calculated

earlier)
• Calculate pooled Standard Deviation using following formula:

• Execute the power T Test, with current parameters, and decide if larger
size is needed.
• Calculate the samples number (in case Power of Test is insignificant)

Pooled Standard Deviation:

#
# Calculation to see need of Larger Sample Size:
# Power of the test
delta=mean(Current)-mean(New)

10 | P a g e
pooledSD <- (((40-1)*(8.75^2)+(40-1)*(9.9^2))/(40+40-2))^0.5
delta

## [1] 2.775

pooledSD

## [1] 9.342711

Power T Test:

power.t.test(n=40, delta = 2.775, sd=9.342,

sig.level = 0.05,type = "two.sample",
alternative = "two.sided" )

##
## Two-sample t test power calculation
##
## n = 40
## delta = 2.775
## sd = 9.342
## sig.level = 0.05
## power = 0.258536
## alternative = two.sided
##
## NOTE: n is number in *each* group

Inference: The Power of test is 0.258 or 25.8%, which means there are only
25% chances that the null hypothesis will not be rejected when it is false.
Hence, we should revisit the number of samples to increase the power of test.

Recalculate Sample size using Power T Test:

Consider Power of test 95%, and significance level 0.188 (The P value
calculated) and execute the Power T test once again.

#
# Sample size required
#
power.t.test(power=0.95, delta = 2.775,
sd=9.342,sig.level = 0.188,type = "two.sample",
alternative = "two.sided" )

##
## Two-sample t test power calculation
##
## n = 199.2145
## delta = 2.775
## sd = 9.342
## sig.level = 0.188
## power = 0.95
## alternative = two.sided
##
## NOTE: n is number in *each* group

11 | P a g e
Inference: We can see that we need sample size of 200 (rounded up) to get
95% power of Test.

4 Conclusion and Recommendation

• From the Preliminary Data Analysis, we confirm that the mean driving
distance of New Ball is less than Old Ball. (267.5 yards Vs 270.3 yards).
• When the Data Set explored further with the help of descriptive statistics
and Visualization, we learnt that
o The New Golf ball has relatively higher variation.
o No outliers observed in both the samples.
o Both the samples have nearly Normal distribution; however New
Design is slightly more skewed towards right.
• The Hypothesis Testing concludes that this data does not provide
statistical evidence that the new golf balls have either a lower or higher
mean driving distance.
• This implies that Par Inc. should take the new golf balls in production as
the p-value indicate that there is no significant difference between
estimated population mean of current as well as new golf balls.
• Confidence Interval:
o The 95% confidence interval of population mean for Current model
is between 267.4757 & 273.0743.
o The 95% confidence interval of population mean for New model is
between 264.3348 & 270.6652.
• Power of Test:
o Although the 2 Tail Hypothesis Test recommends to launch the new
ball design into Production, the Power of Test is only 25%.
o In order to have 95% Power of Test, it is recommended to have
the number of samples as 200 for both the designs, and then
conclude the Hypothesis test recommendations.

12 | P a g e
5 Appendix – Source Code
#=======================================================================
#
# New Golf Ball Design - Should be Launched or not
#
#=======================================================================
# Environment Set up and Data Import
#=======================================================================
# Install Packages
#=======================================================================
#
# Setup Working Directory
setwd("D:/00 Great Lakes/BACP.Aug17 FBS")
getwd()

## [1] "D:/00 Great Lakes/BACP.Aug17 FBS"

#
# Read Input File
golf_data=read.csv("golf.csv")
attach(golf_data)
#

5.1 Exploratory Data Analysis

# Find out Total Number of Rows and Columns

dim(golf_data)

## [1] 40 2

# Find out Names of the Columns (Features)

names(golf_data)

## [1] "Current" "New"

# Find out Class of each Feature, along with internal structure

str(golf_data)

## 'data.frame': 40 obs. of 2 variables:

## $ Current: int 264 261 267 272 258 283 258 266 259 270 ...
## $ New : int 277 269 263 266 262 251 262 289 286 264 ...

#
# Check top 6 and bottom 6 Rows of the Dataset
head(golf_data)

tail(golf_data)

#
#Check for Missing Values
colSums(is.na(golf_data))

## Current New
## 0 0

13 | P a g e
5.2 Descriptive Statistics
#
# Provide Summary of a Dataset.
summary(golf_data)

## Current New
## Min. :255.0 Min. :250.0
## 1st Qu.:263.0 1st Qu.:262.0
## Median :270.0 Median :265.0
## Mean :270.3 Mean :267.5
## 3rd Qu.:275.2 3rd Qu.:274.5
## Max. :289.0 Max. :289.0

#
# Check all values of a Feature with it's frequencies.
table(Current)

## Current
## 255 258 259 260 261 262 263 264 265 266 267 268 270 272 273 274 275 276
## 1 2 1 2 1 1 3 2 1 2 2 1 2 3 1 2 3 1
## 278 279 280 281 283 284 287 289
## 1 1 1 1 2 1 1 1

table(New)

## New
## 250 251 253 255 259 260 261 262 263 264 266 268 269 270 271 272 274 276
## 2 1 1 1 1 2 1 4 4 3 2 1 2 1 1 1 2 1
## 277 278 279 280 281 283 286 289
## 1 1 1 1 2 1 1 1

#
#=======================================================================
# Feature Exploration
#=======================================================================
# Data Visualization using Graphs:
# Histogram for Continuous Variables
# Box Plot to see the Outliers in Continuous Variables
#=======================================================================
#
dev.off() # To Reset the earlier partition command.

## null device
## 1

par(mfrow=c(2,2))
hist(Current,main='Current Ball',xlab = "Driving Distance",ylab =
"Frequency",col = "turquoise")
hist(New,main='New Ball',xlab = "Driving Distance",
ylab = "Frequency",col = "turquoise1")
#
boxplot(Current,main='Current Ball',xlab = "Driving Distance",
ylab = "Frequency",col = "turquoise",horizontal = TRUE)
boxplot(New,main='New Ball',xlab = "Driving Distance",
ylab = "Frequency",col = "turquoise1",horizontal = TRUE)
14 | P a g e
#
#=======================================================================
# Mean, Standard Deviation and Variance
#=======================================================================
# Normal mean(),sd(),var() Functions are used.
# round() function is used with parameter digits,
# to display the results up to two decimal places
#=======================================================================
#
Col_Head <- c("Mean","Standard Deviation","Variance","Remark")
Current_Stats <- c(round(mean(Current),digits = 2),
round(sd(Current),digits = 2),
round(var(Current),digits = 2),
"Current Ball ")
New_Stats <- c(round(mean(New),digits = 2),
round(sd(New),digits = 2),
round(var(New),digits = 2),
"New Ball Stats")
#
Combined_Stats <- rbind(Col_Head,Current_Stats,New_Stats)
Combined_Stats

## [,1] [,2] [,3] [,4]

## Col_Head "Mean" "Standard Deviation" "Variance" "Remark"
## Current_Stats "270.27" "8.75" "76.61" "Current Ball "
## New_Stats "267.5" "9.9" "97.95" "New Ball Stats"

#
# Welch Two Sample t test
# Also provides 95% Confidence Interval for the difference in means.
t.test(Current,New)

##
## Welch Two Sample t-test
##
## data: Current and New
## t = 1.3284, df = 76.852, p-value = 0.188
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -1.384937 6.934937
## sample estimates:
## mean of x mean of y
## 270.275 267.500

# One Sample t test to get 95% Confidence Interval

t.test(Current)

##
## One Sample t-test
##
## data: Current
## t = 195.29, df = 39, p-value < 2.2e-16
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 267.4757 273.0743
15 | P a g e
## sample estimates:
## mean of x
## 270.275

t.test(New)

##
## One Sample t-test
##
## data: New
## t = 170.94, df = 39, p-value < 2.2e-16
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 264.3348 270.6652
## sample estimates:
## mean of x
## 267.5

#
# Calculation to see need of Larger Sample Size:
# Power of the test
delta=mean(Current)-mean(New)
pooledSD <- (((40-1)*(8.75^2)+(40-1)*(9.9^2))/(40+40-2))^0.5
delta

## [1] 2.775

pooledSD

## [1] 9.342711

power.t.test(n=40, delta = 2.775, sd=9.342,

sig.level = 0.05,type = "two.sample",
alternative = "two.sided" )

##
## Two-sample t test power calculation
##
## n = 40
## delta = 2.775
## sd = 9.342
## sig.level = 0.05
## power = 0.258536
## alternative = two.sided
##
## NOTE: n is number in *each* group

#
# Sample size required
#
power.t.test(power=0.95, delta = 2.775,
sd=9.342,sig.level = 0.188,type = "two.sample",
alternative = "two.sided" )

##
## Two-sample t test power calculation
16 | P a g e
##
## n = 199.2145
## delta = 2.775
## sd = 9.342
## sig.level = 0.188
## power = 0.95
## alternative = two.sided
##
## NOTE: n is number in *each* group

#
#=======================================================================
#
# T H E - E N D
#
#=======================================================================

17 | P a g e