Variation Levene test

To calculate if there is a significant difference in variation between two data-sets. This test is robust when the data is not normally distributed. It calculates the Levene variation test between this column and the comparing column.

Interpretation

• The smaller the p value is the more likely there is a significant difference in variation between the data-sets.
• Develve uses the commonly accepted value of p < 0.05 for significance.
• If the data-sets is normally distributed use the Variation F-test
• For this test is no minimum sample size (power) calculation

Options

• Bigger/Smaller (one sided test) or not equal test (two sided)

Colors of the cells

• Green
  No significant difference
• Yellow
  Significant difference

Formula Calculating W value

With the W value and the degrees of freedom can the program calculate the p value.

Legend

Mean of all data
Mean group
total number of samples
number of samples in group
number of groups

Example

Select Variation test. To use the Levene test first select "non normal distributed" when the box is not selected the F-test is calculated. Then select Diff variation. In this example C and D are significantly different in variance compared to A.

Data file

Multiple Levene test

To calculate if there is a significant difference in variation between two or more data-sets. This test is also robust when the data is not normally distributed. It calculates the Levene variation test between the selected columns.

Interpretation

• The smaller the p value is the more likely there is a significant difference of one ore more of the selected data-sets.
• Develve uses the commonly accepted value of p < 0.05 for significance.
• For this test is no minimum sample size (power) calculation

Colors of the cells

• Green
  No significant difference
• Yellow
  Significant difference

formula

With the W value and the degrees of freedom the program can calculate the p value out of the F distribution.

Legend

Mean of all data
Mean group
total number of samples
number of samples in group
number of groups

Example

To use the Multiple Levene test "Menu: Tools=> Multiple Levene test". Then select the data-sets to compare. In this example there no significantly different in variance in the selected columns.

Data file

External links

• http://en.wikipedia.org/wiki/Levene_test
• http://www.itl.nist.gov/div898/handbook/eda/section3/eda35a.htm