MATH325 Lab 5
UsingMinitab to conduct ANOVA procedures
The steps required for completing the deliverables for this
assignment (screen shots that correspond to these instructions can be found
immediately following them).Complete the questions below and paste the answers
from Minitab below each question (type your answers to the questions where
noted). Therefore, your response to the
lab will be this ONE document submitted to the Dropbox.
Context (remember that statistics are far more than numbers
or values – you need to know the context to perform a good analysis!).
Study: A nurse practitioner is studying the effect of blood
sugar (glucose) control, which involves collecting the average daily AC and QHS
(fasting) blood sugar levels of the patients to determine if there is
relationship between these and the patients’ Hemoglobin A1C level. She hypothesizes that good blood sugar
control will result in ideal Hemoglobin A1C levels and inadequate control of
the patients’ blood sugar will result in high Hemoglobin A1C levels.
She also tracks other factors that may contribute to the
patients’ control of their blood sugar such as carbohydrate intake, age,
frequency of glucose checks, and insurance coverage of diabetic supplies.
Hemoglobin: Ideal Hemoglobin A1C levels for patients are 6
or 7, a value of 8 or 9 merits concern, values 10 and up are considered
severely uncontrolled, while values less than 6 are rare in diabetic
patients. 4 and 5 can be found normally
in patients that are not diabetic.
Blood Sugar: Glucose levels under 70 are considered low,
between 70 and 110 is considered normal, 111 to 170 is considered moderately
high, and values above 170 are considered high. There is some debate on the cut
points, however, these are the values used to categorize glucose levels in this
study.
Glucose _Range: This is a categorical variable describing
the group into which the patient’s glucose level places them: low, normal,
moderately high, and high.
Glucose_Group: This is a numeric variable containing the
same information as the Glucose_Range, however, the numeric value assigned to
each group can be used in analysis that requires a ratio or interval level of
measurement. Low is assigned a 1, Normal is assigned a 2, Moderately High is
assigned a 3, and High is assigned a 4.
Carbohydrates: Diabetic patients try to consume 14 servings
of carbohydrates daily where each serving is approximately 15 grams. This study tracks the average grams of
carbohydrates consumed on a daily basis by these patients.
Age_Range: This is a Categorical Variable where each patient
is classified by age: under 10, 11 – 16, 17 – 25, 26 – 40, 41 – 60, and over
60.
Age_Group: This is a numeric variable that contains the same
information as the Age_Range, however, the numeric value assigned to each group
can be used in analysis that requires a ratio or interval level of
measurement. Each patient is classified
by age: under 10 is assigned a 0, 11 – 16 is assigned a 1, 17 – 25 is assigned
a 2, 26 – 40 is assigned a 3, 41 – 60 is assigned a 4, and over 60 is assigned
a 5.
Insurance: This is a categorical variable that describes if
the patient’s insurance covers diabetic supplies. Yes/No.
Insurance_Group: This is a numeric variable that describes
the same information as the Insurance variable; however, the numeric value
assigned to each group can be used in analysis that requires a ratio or
interval level of measurement. Yes is assigned a 1 and No is assigned a 0.
Frequency: This
numeric variable describes how many daily checks of their glucose level are
typically performed on a given day for each diabetic patient.
View the Minitab tutorial on ANOVA. The ANOVA tutorial can
be found by going to the Help menu in Minitab, selecting ANOVA and then
selecting One-Way ANOVA.Read through Uses, Data and How To in the Tutorial
window.
Note: The data files
referenced in the tutorial are available in DocSharingfile (Minitab_Sample
DataSets_HelpMenu). I suggest you print
out the steps needed to perform the deliverables for the lab and as these items/steps
come up in the tutorials, also use the HealthCareData.mpj data set to work
along at that point.
For a specific example, choose Stat, ANOVA and then
One-Way. In the dialog box that pops
ups, select Help.
With the One-Way ANOVA procedure, you will be able to:
· Validate the
assumption of variance equality
· Obtain the
ANOVA table and results
· Visually
inspect the group means
· Perform custom
contrasts, tailored to your specific hypotheses
· Compare each
mean to every other mean, assuming variance equality or not
· Perform two
types of robust analysis of variance
One of the first steps in the performance of an analysis of
variance (ANOVA) is to validate the assumptions necessary for use of the test.
Then we perform the analysis.
Examine descriptive statistics for each of the independent
variable’s groups and the Levene statistic to assess if the variances are
equal. The F statistic used in the ANOVA
test can be robust to unequal variances if the samples sizes are approximately
equal. However, our first step is always to test the equality of the variances.
It is also advisable to perform a plot of the means to get a
visual indication of where you may expect to find similarities and differences
among the groups.
We will use Minitab and request the descriptive statistics,
the Levene statistic and a plot of the Means as well as calculate the
F-Statistic with its significance.
After the initial analysis, we will then perform an analysis
using contrasts to target specific groups indicated by the data:
Is there truly a difference between the low and normal
groups?
Is there a difference between the moderately high and high
groups?
Is there a difference between the normal and moderately high
groups?
To obtain a One-Way ANOVA using Minitab
1. Open the
HealthCareData.mpj file using Minitab.
2. First, we
want to test the equality of the variances.
From the menus, select Stat, ANOVA, Test for Equal Variances
Choose Hemoglobin for the Response and Glucose_Group.
3. Click OK to
examine the output and perform an initial analysis of what you see.
4. To obtain the
One-Way ANOVA, select Stat, ANOVA, One-Way.
Choose Hemoglobin for the Response and Glucose_Group for the Factor.
5. Click Graphs
and then select Boxplots of data.
6. Click OK, OK
and examine the output. Now we begin the
contextual analysis.
7. Think about
it: Were the assumptions for ANOVA
met? Can we proceed even if they
aren’t? Under what circumstances? What did the data and tests show us about the
two variables and the groups involved?
Did we need the Levene test?
8. Deliverable:
Save this document and submit it as Week_6_i-Lab_YourNameHere.docx to the
Dropbox.
