Performing Statistical Analysis on EXCEL

Download Report

Transcript Performing Statistical Analysis on EXCEL

Performing
Statistical
Analyses on
EXCEL
Session #10
CHLA Core Curriculum on Scholarship Skills
Research Methodology Section
Step #1:
Install Statistical
Tools in Your
EXCEL Program
1. Start Excel on your computer
2. Select “Tools” option (upper tool bar)
3. Select “Add Ins” option
1. Click/select “Analysis ToolPak” and “Analysis TookPak-VBA
2. Click/select “OK”
3. The Excel Program should now install these analysis
programs to this copy of Excel
1. Check if the statistical programs have been installed
2. Click on “Tools” option
3. At the bottom of the Tools menu, you should see “Data Analysis”
Step #2:
Organize/Set Up
Data in EXCEL
1.
2.
3.
4.
Enter your data in Excel
5. Use numbers, not text, for
Each subject data per line
variables to be analyzed
Variables in columns
6. Avoid spaces/blank lines
Don’t forget to include code definitions
between data sets
(see Group and Sex variables above) 7. Make sure that data is in
“Number” format
1. Highlight data (left click on beginning
of data, hold left mouse key down,
and drag to end of data) of
interest/inquiry on your spreadsheet
2. Click/select “Format” option on tool
bar
3. Select “Format Cells” on pull-down
menu
1. The “Format Cells” section should have “Number” selected
as the “Category”
2. You can adjust the decimal places and formats as desired
3. Then click on “OK”
Step #3:
Using Data
Analysis Tools
[Using sample or your own Excel spreadsheet]
1. Make a copy of your
database/spreadsheet
(otherwise you could lose all of
your data)
2. Work ONLY on the database copy
(NOT on the original)
3. Highlight ALL of the data
4. Select “Data” in the tool bar
5. From the pull-down menu, select
“Sort”
1. Sort options: in the above example I
want to sort FIRST by Group (column B)
and then by Age (column C)
2. I can sort by ascending or descending
sequence
3. In this example, I did NOT include the
header row (variable names)
4. Once you have selected your desired
options, press “OK”
NOW we can finally run some stats!
1. Select “Tools” on the tool bar
2. Select “Data Analysis” from the
pull down menu
3. Scroll down the “Data Analysis”
menu and select the desired stat
test
4. Then press “OK”
1. In this example, I will examine if there is a difference in Body Mass Index (BMI)
between the 2 groups
2. I selected “t-test: two-sample assuming unequal variances”
3. In the “Input” section, click each range section and indicate the location of the data
for that variable (click on range/white section; then go to spreadsheet and highlight
the desired data – in this case, you can see I highlighted BMI of the Group 2
subjects for the variable #2 range)
4. You have options on where you want EXCEL to list the output/stat results
5. You can either use or ignore the other options
6. Click on “OK” to run this stat test
RESULTS:
• Variable #1 is the BMI of
Control subjects
• Variable #2 is the BMI of HF
subjects
• Note that there is a
significant difference
between the 2 groups (if
one assumes a significance
cut-off of p < 0.05)
Results of the unpaired t-test example.
Limitations for Statistical
Analysis with Excel
• Primarily limited to parametric statistics
• May not be as accurate as standard stat packages
• Have to sort or hand-rearrange data frequently
Chi-Square Analysis
Resources on the Web
• Interactive Chi-Square analysis web site
– http://www.people.ku.edu/~preacher/chisq/chisq.htm
• 2x2 Chi-Square or Fisher’s Exact site
– http://www.graphpad.com/quickcalcs/Contingency1.cf
m
• Program to run Chi-Square on Excel
– http://otc.isu.edu/~hurley/Calculator;%20Chi%20Squa
re.xls