![]() |
HSTA Lifestyles Project |
CORRELATION The extant to which one variable varies in relation to another
|
First: You may need to SORT DATA
(such as all 16 year old females or lo-fat males or bmi from hi to lo…).
1. Highlight all data by clicking in the left, upper most cell (above the 1 and to the left of the A).
2. Click on “Data” in top tool bar.
3. Then click on “Sort”. 4. Then choose which columns you want ascending or descending. 5. Then click on 'OK', and the columns will sort.
Next: Do CORREL:
1. Click on any blank cell.
2. Type in "=" sign.
3. Above the C column, click on the small 'down' arrow. Look for CORREL and click. 4. For 'Array 1', highlight the first group or column of numbers you want to compare. 5. For 'Array 2', highlight the second group or column of numbers to see a relation to. Note: These arrays must have same number of variables. 6. Click “OK” and the CORREL results will appear.
If CORRELATION is 1, then variables are perfectly related.
As CORRELATION approaches 0 - variables are less related.
If CORRELATION between 0 and 1 - variables are somewhat related.If CORRELATION is –1, then the 2 variables are inversely related.
If CORRELATION between 0 and –1, then variables are somewhat inversely related.
Next: Graph it!
back to Data Analysis