News
On June 1, we’re asking you to select a content label when starting a new topic in the Discussions area. Read more to find out why.
Choose Language Hide Translation Bar
Highlighted
Juno
Level III

ANOVA assumption test

Hello All,

 

I am performing ANOVA assumtion test befroe combiningg RCBD data from different enviornments. 

 

I know how to check for normality and unequal variances, but how do I check for indepedence of residuals?

Thank you

0 Kudos
1 ACCEPTED SOLUTION

Accepted Solutions
Highlighted

Re: ANOVA assumption test

The issue is to create a plot of residuals in experimental execution order. So it sounds like you are lacking that order? If so plotting residuals in time sequence will be problematic. If you have the experimental execution order, then just create another column with that order and create a simple run chart in the Analyze -> Quality and Process -> Control Chart - Run Chart platform.

View solution in original post

3 REPLIES 3
Highlighted

Re: ANOVA assumption test

A few thoughts for you...independence of residuals is most influenced by the manner in which the data was collected and the specific randomization (or lack thereof) techniques or other experimental execution methods (like blocking) were implemented and used. If your experiment was run in some nonrandom order for example, then the independence assumption might, I emphasize might, be problematic. A couple thoughts for you:

 

1. Check a plot of residuals in the order in which the data was generated (say from a designed experiment) or collected.

2. Examine the Durbin Watson statistic...this statistic is used to evaluate autocorrelation of error.

0 Kudos
Highlighted
Juno
Level III

Re: ANOVA assumption test

Hello Peter,

 

I can generate a new Column for the "residual of responses". How do I plot these and test if they are scattered evenly?

 

Thank you

0 Kudos
Highlighted

Re: ANOVA assumption test

The issue is to create a plot of residuals in experimental execution order. So it sounds like you are lacking that order? If so plotting residuals in time sequence will be problematic. If you have the experimental execution order, then just create another column with that order and create a simple run chart in the Analyze -> Quality and Process -> Control Chart - Run Chart platform.

View solution in original post