Introduction To Minitab For Windows

advertisement
Introduction To Minitab For Windows
Minitab is a powerful statistical package that provides a wide range of basic and advanced data analysis
techniques. The purpose of this handout is to provide you information about using Minitab with Windows in
the Computer Suites in NUI Galway.
A.
Running Minitab:
Find and double click the Minitab Icon in the Novell Delivered Applications Window: After a few
seconds a split-screen will appear. The top screen is referred to as the "Session" window, and the
bottom screen is the "Data" window.
B.
Entering data into Minitab
Data is entered into Minitab by entering values in a worksheet composed of columns and rows. The
columns are denoted C1, C2, ... Within the data window, at the top of every column is a space where
you should type in a name for the data to be entered in the column. Two primary methods are used to
enter data into the Minitab worksheet.
1.
Opening Data Sets: (This is typically the option you will use)
For most problems you will be assigned in this course, the data values have already been entered
in a data file. All of the data sets relate to this course are on the Q: drive in the
Mathematics/jnewell/MA238 folder.
All you need to do is to "load" the data into the Minitab worksheet. For example, the following
sequence of clicks
File>Open worksheet>
locate the Mathematics/jnewell/MA238 folder and choose the dataset you need.
Note: You cannot use the folder icon to open a data file!
2.
Entering data values directly into the worksheet:
You may enter data values directly into columns in the Data window. For example, place the
cursor in the first cell of C1. Enter a data value, press the "down arrow " key, enter a value,
press the "down arrow" key, etc. until all the data values are in the worksheet. If you need to
correct a value, move the cursor to the incorrect cell and enter the correct value.
C.
Performing Computations
Once the data have been loaded (or entered manually) you can now use Minitab to carry out a variety of
analyses such as calculating summary statistics, creating plots or performing complicated formal
analyses. For example take a look at the drop down menu Stat and look at the various options
available.
1
D.
Printing
You are able to edit your session window. Somewhere within the session window you should type in
your name, some title identifying the work, and any other pertinent information. To print the session
window, double click on the printer icon while in the session window.
Throughout the semester, examples will be provided to demonstrate the entering and analyzing of data.
Using Minitab to Calculate Interval Estimates of a Population Mean
Example1 : Body Fat
Context:
A representative sample of 78 men and women in the age range 18-30 had their percentage body
fat measured by densitometry (i.e. underwater weighing).
Question:
What are the typical values of body fat for men and women respectively? Provide a 95%
Confidence Interval for the true mean % Bodyfat separately for each sex.
Data:
The data is in a Minitab worsheet called ‘BODYFAT.mtw’
which has the percentage body fat (%BF) for males and females in separate columns.
Worksheet For Body Fat Data
Aim:
The key question of interest is to determine the true average % Body fat of men and women in the
population from which the sample was chosen. We are assuming the samples were chosen at random.
A template analysis for the males is presented below which you can adapt to use to analyse the female
data.
Subjective Impression:
In order to get a feel for the data and a subjective answer to the question let us look at a descriptive
statistics and boxplots of the data by means of the commands
Stats > Basic Statistics > Display Descriptive Statistics.
Choose ‘'%BF_Female'as the variable.
Press Graphs > select Boxplot and Press OK.
Press OK.
You should obtain
2
Descriptive Statistics: %BF_Female
Variable
%BF_Female
N
78
N*
0
Variable
%BF_Female
Maximum
28.400
Mean
19.638
SE Mean
0.515
StDev
4.548
Minimum
4.800
Q1
16.100
Median
20.000
Q3
22.900
Boxplot of %BF_Female
30
%BF_Female
25
20
15
10
5
The shape of the boxplot suggest that %Body Fat in females may follow a Normal distribution as the
boxplot is quite symmetric. There is an outlier however representing a female with a particularly low
%Body Fat reading.
The sample mean %Body Fat is calculated as 19.64% with a standard deviation of 4.55. Note that the
median is quite close to the mean which is a further suggestion of symmetry (i.e. normality).
However before any conclusions can be drawn as to what the likely true mean %Body Fat is for females
an objective analysis of the data must be carried out.
Formal Analysis:
To carry out a formal analysis of the question we need to produce a 95% confidence interval for the
population mean %Body Fat. As the sample is larger than 30 we can use a Z based confidence interval.
To do this use the command
Stats > Basic Statistics > 1-Sample Z.
In the Samples in One Column dialog box, choose ‘%BF_Female’
Fill in 4.548 for the Standard deviation.
Press the Options button to verify that the degree of confidence is set at 95.
Press OK.
3
You should now obtain the output:
One-Sample Z: %BF_Female
The assumed standard deviation = 4.548
Variable
%BF_Female
N
78
Mean
19.6385
StDev
4.5483
SE Mean
0.5150
95% CI
(18.6292, 20.6478)
Remember that strictly speaking we should be calculating a t-based interval here as we are substituting the
sample standard deviation as an estimate of the true (an unknown) population standard deviation.
For comparison purposes calculate a t-based confidence
Stats > Basic Statistics > 1-Sample t.
In the Samples in One Column dialog box, choose ‘%BF_Female’
Press OK.
You should now obtain the output:
One-Sample T: %BF_Female
Variable
%BF_Female
N
78
Mean
19.6385
StDev
4.5483
SE Mean
0.5150
95% CI
(18.6130, 20.6639)
Note the similarity in the two intervals.
Conclusion:
The 95% confidence interval calculated suggests that is it quite likely that the true mean %Body Fat in
females % is somewhere between 18.6% and 20.7%.
Additional Questions.
i.
Calculate a 99% confidence interval estimate of the true mean and compare your results. What do
you notice regarding the width of the interval?
ii.
Repeat the analysis for males. Minitab to create a histogram in addition to the boxplot. Have a go at
editing the graph.
4
Download