How To Use Minitab 1 Basics

HOW TO USE MINITAB:
INTRODUCTION
AND BASICS
Noelle M. Richard
08/27/14
CONTENTS
1.
2.
3.
4.
5.
6.
7.
8.
* Click on the links to

jump to that page in
the presentation. *
Minitab Environment
Uploading Data to Minitab/Saving Projects
Graphical Summary (First Look)
Normality Test (Is your data normally distributed?)
Descriptive Statistics (Mean, Variance, Range, etc.)
Confidence Intervals and Plots (Estimates of the mean)
ANOVA tables (Analysis of Variance)
Hypothesis Tests (Are two sample means significantly
different? Or are they equivalent?)
9.
10.
11.
One-Way ANOVA (Comparing 3 or more sample means)

Tukey Comparisons (Comparing 12+ sample means)
Standardizing Data
WHAT
MINITAB?
* For more Minitab

tutorials, click the
links below. *
A program used for statistical analysis
IS
As of June 2014, version 16
Can be used for:
Basic Statistics
Graphs/Summary Statistics
Confidence Intervals
Hypothesis Tests
ANOVA Tables
Regression Analysis and Modeling

Simple Regression
General Linear Models
Return to Contents
WHAT
IS
MINITAB?
Quality Control
Control Charts
Capability Analysis
Gauge R & R Studies
Design of Experiments
Randomized Complete Block Designs
Factorial Designs
Fractional Factorial Designs
Split-Plot Designs
Response Surface
Unfamiliar with these terms? No worries! They will be

explained later on.
Return to Contents
MINITAB ENVIRONMENT
Project Window
Session Window
Where you enter in your data
Where your results/output will be displayed
Toolbar/Dropdown Menus
Return to Contents
MINITAB ENVIRONMENT
Return to Contents
PROJECT WINDOW
Minitab analyses data

in columns
If data arranged in rows,
transpose in excel first
Transposing in Minitab
may mess up your data
First Row (C1, C2, C3, etc.) Column Labels

D = Date T = Text Blank = Numeric
Second Row Variable Names
Return to Contents
GETTING DATA
MINITAB
Type in yourself
1.
2.
INTO
Copy/Paste from Excel

Highlight/copy cells in Excel
Highlight exact # of cells in Minitab Project Window
(MUST be exact dimensions)
Paste
3.
Import
FileOpen Worksheet
Select the Excel file and Open
This will import all worksheets from the file.
Close out any you dont want.
Return to Contents
IMPORTING DATA FROM EXCEL
Return to Contents
SAVING PROJECT
Saving your Project will:
Save all data entered into Project Window(s)

All output in the Session Window
Can also save

individual
worksheets to
open in other
projects
10
Return to Contents
GRAPHICAL SUMMARY
Data in the following examples: There are

rectangle pieces with die on each. The die
have different diameter sizes. We are
measuring the width of a porous material
surrounding each die. Pieces were etched
using different methods.
Great first look at your data
Add the variable(s) you want to look at in Variables box
Can also split data up using By variables
Will produce separate graphs of Variable for each By Variable
11
Return to Contents
GRAPHICAL SUMMARY/NORMALITY
Output: Summary Statistics, Histogram, Test for Normality, etc.
Normality Test:
P-Value 0.05
Data not normally distributed
Why is this important? Many statistical

techniques assume normality of your data
Anderson-Darling Test only good for larger
samples (n > 20)
If n < 20, see next slide.
12
Return to Contents
NORMALITY TEST USING PROBABILITY PLOT

Use if sample
size < 20
Output
If the data are normally distributed, most

red data points should lie on the center
blue line. Almost all points should be
within the outer blue lines.
13
The plot to the left shows data that is NOT

normally distributed.
Return to Contents
DESCRIPTIVE STATISTICS
Shows various summary

statistics like mean,
variance, range, etc.
Can select what
statistics you want to
see
14
Return to Contents
DESCRIPTIVE STATISTICS
Output stored in Session Window
Can Highlight and Copy all text
15
Return to Contents
CONFIDENCE INTERVALS/INTERVAL PLOTS
95% confidence intervals (C.I.) for mean
Graphical Summary produces C.I. for mean, median

and standard deviation
Your data is a sample from a large population

Use C.I. to estimate the true mean for the population
95% = confidence level
Can change confidence level to what you want

Higher level means wider interval. Usually 95% is sufficient
Smaller sample sizes and larger std. deviation produce wider intervals
Can draw interval plot to visualize the intervals

16
** http://en.wikipedia.org/wiki/Confidence_interval#Meaning_and_interpretation
Return to Contents
CONFIDENCE INTERVALS/INTERVAL PLOTS

Can find
intervals (for
the mean) of
specific
group(s) in
your data
Output:
17
Return to Contents
ANOVA TABLES
Analysis of Variance
How much of your variation is within groups?

How much of your variation is between groups?
Are the means of 2 or more samples statistically
equivalent?
ANOVA tables can answer these questions. They

are used in various other areas of statistics too
Seeing how well a regression model fits data

Finding Repeatability and Reproducibility in Gage R & R Studies
One-Way ANOVA tables are used to compare the means of
more than 2 samples. In order to proceed with One-Way

ANOVA, we need to understand Hypothesis Tests.
18
Return to Contents
HYPOTHESIS TESTS
Answer questions like: Are population means equivalent? Is one

greater than the other? Do populations have the same std.
deviation?
Each test contains two hypotheses:
1. Null Hypothesis: This is the default hypothesis. Usually, it
assumes there is no relationship between 2 phenomena
ex.) 2 population means are equivalent.
2. Alternative Hypothesis: Another hypothesis that could be true
ex.) mean 1 is larger than mean 2
The goal is to either:

1. Not Reject the null hypothesis or
2. Reject the null hypothesis, in favor of the alternative
19
Return to Contents
HYPOTHESIS TESTS
After the hypotheses have been stated, the test quantifies the
evidence against the null hypothesis. This evidence is reflected in
a value known as the test statistic.
From this test statistic, we get a p-value, which is used determine
whether to reject the null hypothesis or not.
Large p-value large amount of evidence in favor of the null hypothesis
Small p-value small amount of evidence in favor of the null hypothesis
Rule of Thumb:
p-value 0.05 do not reject the null hypothesis

p-value < 0.5 reject the null hypothesis in favor of the alternative
Click here for more information about p-values, if interested.
20
Return to Contents
ONE-WAY ANOVA
As mentioned before, it is used to compare the means of more than

2 samples.
How does it work?
Its a Hypothesis Test!
Null Hypothesis: the samples were drawn from populations with the same
mean values.
Alternative Hypothesis: the samples were drawn from pops with different
mean values.
Note: In this case, if you reject the null hypothesis, you ONLY know that at
least one of the samples is different than the others. But you do not know
which one it is.
Can use confidence intervals to estimate the difference between population
means. Intervals that do not contain 0 indicate the means are significantly
different.
21
Return to Contents
ONE-WAY ANOVA / HYPOTHESIS TESTS
Can use Minitab Assistant (v.16 only)
Select what type of test

you want to do.
Here, we will proceed with
One-Way ANOVA in the
3rd column.
Remember, this is for

comparing means of more
than 2 samples.
If you are comparing only two

samples, use the 2-Sample t test.
Data from your samples should be
in separate columns.
22
Return to Contents
ONE-WAY ANOVA / HYPOTHESIS TESTS

* This process in
Minitab only works for
12 samples or less. *
Use Tukey method if
you have more than
12
Choose how your data is set up
Several windows will

now pop up with your
results, as well as
some interpretation.
23
Return to Contents
ASSISTANT OUTPUT
Also provides confidence intervals for

the means
24
Return to Contents
TUKEY COMPARISONS
Are means of samples/groups significantly different?
Use Tukeys Method
Pair-wise comparisons of means

Calculates confidence intervals for Mean 1 Mean 2, for all pairs.
If the confidence interval contains 0, no significant difference b/w the
two means.
Then, using these intervals, can group means. Means in same group
are not significantly different.
Whats special about Tukeys method? Considers all comparisons
simultaneously. Confidence levels of the individual intervals are
adjusted in order to reduce error.
Use if you have more than 12 groups whose means you want to
compare.
25
Return to Contents
MINITAB- TUKEY
Output
Groupings based
on Tukey
26
Return to Contents
STANDARDIZING DATA
Sometimes, standardizing data is useful.
Can be easier to work with the data

Standardizing: For each data point, subtract off the mean, divide
by standard deviation
Store your
results in the
next empty
column in
your dataset
27
Return to Contents
REFERENCES
Khan, R. M. (2013). Problem solving and data analysis using

minitab: A clear and easy guide to six sigma methodology
(1st ed.). West Sussex, United Kingdom: Wiley.
http://en.wikipedia.org/wiki/One-way_analysis_of_variance
http://en.wikipedia.org/wiki/Confidence_interval
http://en.wikipedia.org/wiki/Statistical_hypothesis_testing
http://www.itl.nist.gov/div898/handbook/pri/pri.htm
28
Return to Contents

How To Use Minitab 1 Basics

Uploaded by

Copyright:

Available Formats

How To Use Minitab 1 Basics

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

How To Use Minitab 1 Basics

Uploaded by

Copyright:

Available Formats

HOW TO USE MINITAB:

* Click on the links to

One-Way ANOVA (Comparing 3 or more sample means)

* For more Minitab

A program used for statistical analysis

As of June 2014, version 16

Can be used for:

Regression Analysis and Modeling

Unfamiliar with these terms? No worries! They will be

Where you enter in your data

Where your results/output will be displayed

Minitab analyses data

First Row (C1, C2, C3, etc.) Column Labels

Copy/Paste from Excel

IMPORTING DATA FROM EXCEL

Saving your Project will:

Save all data entered into Project Window(s)

Can also save

Data in the following examples: There are

Great first look at your data

Add the variable(s) you want to look at in Variables box

Can also split data up using By variables

Will produce separate graphs of Variable for each By Variable

Data not normally distributed

Why is this important? Many statistical

NORMALITY TEST USING PROBABILITY PLOT

If the data are normally distributed, most

The plot to the left shows data that is NOT

Shows various summary

CONFIDENCE INTERVALS/INTERVAL PLOTS

95% confidence intervals (C.I.) for mean

Graphical Summary produces C.I. for mean, median

Your data is a sample from a large population

Can change confidence level to what you want

Can draw interval plot to visualize the intervals

CONFIDENCE INTERVALS/INTERVAL PLOTS

How much of your variation is within groups?

ANOVA tables can answer these questions. They

Seeing how well a regression model fits data

One-Way ANOVA tables are used to compare the means of

more than 2 samples. In order to proceed with One-Way

Answer questions like: Are population means equivalent? Is one

The goal is to either:

p-value 0.05 do not reject the null hypothesis

Click here for more information about p-values, if interested.

As mentioned before, it is used to compare the means of more than

How does it work?

Its a Hypothesis Test!

ONE-WAY ANOVA / HYPOTHESIS TESTS

Can use Minitab Assistant (v.16 only)

Select what type of test

Remember, this is for

If you are comparing only two

ONE-WAY ANOVA / HYPOTHESIS TESTS

Choose how your data is set up

Several windows will

Also provides confidence intervals for

Are means of samples/groups significantly different?