MINI PROJECT – GOLF

BY VISHNU VINOD V.K

ASSIGNMENT

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
SOLLUTION

SET WORKING DIRECTORY AND LOADING DATA SET

 Setting working directory

setwd("D:/GL/miniproject")

 Loading dataset
golf <- read.csv("Golf.csv")

 Sample size: 40
 No.of samples : 2

EXPLORATORY DATA ANALYSIS

 Checking structure of dataset

str(golf)

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

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

 Checking total number of rows and columns

dim(golf)

[1] 40 2

 Checking the names of the columns

names(golf)

[1] "New" "Current"

 Five point summary and standard deviation on both the samples
 New Current
Min. :250.0 Min. :255.0
1st Qu.:262.0 1st Qu.:263.0
Median :265.0 Median :270.0
Mean :267.5 Mean :270.3
3rd Qu.:274.5 3rd Qu.:275.2
Max. :289.0 Max. :289.0

Summary of the given data shows mean and median are very close
the data is normally distributed.

 Checking standard deviation for current

sd(golf$Current)

[1] 8.752985

 Checking standard deviation for new

sd(golf$New)

[1] 9.896904

Also 5-point summary and standard deviations for both columns says
that there is no significant change in the driving distance of balls with
and without coating.

 Variance for current

var(golf$Current)

[1] 76.61474

 Variance for new

var(golf$New)

[1] 97.94872
HISTOGRAM AND BOXPLOT
From histogram we can see that both variable are nearly normally
distributed

Boxplot shows there are no outliers.

OBSERVATIONS
  Sample size:40

 Number of samples: 2

 Unpaired variables.

 DOF = 40+40-2 = 78

 There are no outliers in given data, neither missing values.

 Both the samples seem to be normally distributed.

 Mean and median values are not much different.

 The Current driving distance data looks more normally

distributed,whereas the driving distances data for New balls
looks right skewed.

 There is dip in the performance of Current and New balls driving

force as mean, median, min, max values differ.

HYPOTHESIS FORMULATION AND TESTING

  The level of significance (Alpha) = 0.05

 The sample size N = 40 which is sufficiently large for a Z stat

Test.

 But since the population standard deviation (Sigma) is unknown,

we have to use a T stat Test.

 Since the sample is different for both Sampling tests, we have

N+N-2 degrees of freedom = 78

 Since the sole purpose of the test is to check whether there is

any effect on driving distances due to the new coating, we could
prefer a Two Tailed T Test.

Null Hypothesis:

H0: µold - µnew = 0 (New coating does not have effect on driving
distances)
Alternate Hypothesis:

H1: µold – µnew #0 (New coating does have significant effect on

driving distances)
d̅ = Mean difference
µd = hypothesized difference (usually 0)
sd = Standard deviation of the difference

Welch Two Sample t-test

data: golf$Current and golf$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

Since it is a two-tailed test, the p-value = 0.1879 ÷ 2 = 0.094 (approx.)

The calculated p-value is greater than level of significance α (0.05)
Therefore, the Null Hypothesis (H0) will not be rejected.

TWO TAILED INDEPENDENT ONE SAMPLE T TEST FOR

CURRENT MEAN
 One Sample t-test

data: golf$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

TWO TAILED INDEPENDENT ONE SAMPLE T TEST FOR

NEW MEAN

One Sample t-test

data: golf$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

T-TEST CONCLUSION
TWO TAILED TWO SAMPLE INDEPENDENT T TEST

In this scenario, the p value is 0.094 which is greater than the 0.05.
Hence, we failed to reject the Null Hypothesis.

Thus, accepting the Null Hypothesis that there is no significant change in driving
distances due to the new coating.

95% confidence interval for difference in mean is [-1.384937 TO 6.934937]

TWO TAILED ONE SAMPLE T TEST

95% confidence interval for Current balls driving distance mean is

[267.4757 TO 273.0743]

95% confidence interval for New balls driving distance mean is

[264.3348 TO 270.6652]

 The difference in mean in the case of new balls can also be attributed to
the higher variance compared to `Current` balls.
 The variance of `New` balls driving distances is 97.95 is 28% more than
the variance of the driving distances of `Current` balls 76.61.
 We are unsure of the sampling error present in the data.
 Statistically there is no effect of new coating on driving distances. Though
it is suggested to check the effect on the weights and other characteristics
like size and shape of the new balls.
 Also, the given sample is from only one golf course, It is advisable that test
should perform on different

TYPE 1 AND TYPE 2 ERRORS

Type I Error alpha(α): Probability of rejecting null hypothesis when it is true, the
probability of a Type I error in hypothesis testing is predetermined by the
significance level.
Type II error (β) : Probability of falling to reject the null when it is false. Type II
error calculation **depends on the population mean which is unknown

POWER OF THE TEST AND SAMPLE SIZE

If alternative hypothesis µNew - µcurrent = µd = 5 yard as per our

assumption.

Null Hypothesis µNew - µcurrent = µd = 0

First we need to calculate the probability of Type I error which is

predetermined by significance level. If the significance level is 0.05,
then Type I error is 0.05 i.e. 5% probability we make Type I error -
rejecting null hypothesis when it is true.

Type II error calculation depends on a particular value of µ. In this

case lets assume difference between population µ is 5 yard. Lets also
assume that the significance level for the test is 0.05. Then the
calculation is as below:

This is a two tailed test.

We fail to reject the null hypothesis (commit a type II error) if we get a

Tstatistic less than 1.685954 for the sample size of 40

> abs(qt(0.05,38))
[1] 1.685954

SD for difference is 13.74397

Difference in mean is -2.775
 Two-sample t test power calculation

n = 40
delta = 2.775
sd = 13.74397
sig.level = 0.05
power = 0.14274
alternative = two.sided

NOTE: n is number in *each* group

Basically, the power of the test is the probability that we make the right
decision when the null is not correct (i.e. we correctly reject it)

SAMPLE SIZE TO MAKE PROBABLITIES OF TYPE I AND

TYPE II ERROR

Let us assume that, we need Type I error and Type II error equal to
0.05
Assuming sample standard deviation is equal to population standard
deviation, we can calculate sample size needed as below:

Null hypothesis' mean difference µ0 is 0.

Alternative hypothesis' mean difference µ1is 5.
Sample Standard Deviation is 13.74397.
alpha value (α) is 0.05
Beta value (β) is 0.05 i.e. power of the test is 0.95 = 95 %

Two-sample t test power calculation

n = 197.3383
delta = 5
sd = 13.74397
 sig.level = 0.05
power = 0.95
alternative = two.sided

NOTE: n is number in *each* group

Hence, In order to retain the power, we need to round the value to

next whole number. Therefore, we may conclude that we need a
sample size of 198 to get the Type I and Type II Errors equal.

CONCLUSION

From the given data, it may be concluded that, statistically there is no

significance change in driving distance due to new coating on golf
balls. However, our recommendation is that the test be carried out
with a larger sample size covering number of golf courses (at least a
five different) to improve the accuracy of the test results and negating
any effect of one type of ground. Also, the results need to interpreted
and future actions be planned with the understanding of other
characteristics like size, shape, weight etc