Saving Standardized Residuals as a Variable - ANOVA in JASP

RachelKing42

Hello All,

I am wondering whether there is a way of saving the standardized residuals as a variable when running ANOVA in JASP?

Many thanks!

Rachel

Dmartin427

Although I am not the authority on it, here's how I do it in JASP:

1. Run your ANOVA as a Linear Regression, with your DV in the DV box and your groups in the Factors box.
2. Go to the Statistics dropdown, and then look towards the bottom for the Append Residuals to Data box. Check the box, name your raw residuals column whatever you want, and then press enter.
3. Go to Descriptives --> Descriptive Statistics and then run descriptives on your raw residuals column to get a standard deviation.
4. Go back out to your data, and go all the way over to the right of the screen and click the black plus sign to create a computed column. Name it "stdresiduals", set it as scale, and press Create Column.
5. On the left hand side, click on the variable you created for the raw residuals so it populates in the middle box. Then, go to the mathematical operators above that center box and click on the division symbol. That should create a division line below your raw residual variable in the center box.
6. Click on the three dots underneath the division line, and type in the standard deviation number you received from your descriptive analysis. Click on the "Compute Column" box.
7. If it gives you an error where it outlines your standard deviation figure in the center box in red, just click on that red figure and click "Compute Column" again.
8. Congratulations - you have a fully computed standard residuals column!

If you get stumped, let me know and I'll provide a video on how to do it.

-Dan

Dmartin427

Hey Rachel! Here's how I usually do it in JASP:

1. Run your ANOVA as a Linear Regression instead. Just put your DV in the DV box, and then your grouping variable in the Factor box.
2. Go to the Statistics dropdown below, and check the Append Residuals to Data box. In the textbox, name the variable rawresiduals and press enter. That will populate a column on your datasheet with the raw residuals.
3. To change them to standardized residuals, we need to pull the standard deviation of those raw residuals. You can do that by going to Descriptives --> Descriptive Statistics, and then running your rawresiduals variable to generate a standard deviation.
4. Once you've done that, go back out to your datasheet and look all the way to the right of the screen for a black plus sign along your column headers. Click that, and it'll bring up a box for a computed column - name the column "stdresiduals", set it as a Scale variable, and click Create Column. That will bring up a new box near the top of the screen.
5. Within that screen, go to the lefthand side and scroll down to click on your rawresiduals variable. Then, go to the mathematical operators above the center box and click the division symbol. Once the division line populates under the rawresiduals variable, click on the "..." and input the standard deviation output for your raw residuals.
6. Once you've done that, click on "Compute Column". If it gives you an error and makes the standard deviation number you typed in red, just click on the red box again and try clicking "Compute Column" again and it should work.
7. Congratulations, you have a standardized residuals column with fully populated outputs!

Let me know if you have any issues with that and I'll make a video instead.

patc3

Dmartin427 has a nice way to get it done. Steps 5-6 can be condensed into one step by using the function zScores(y) on the right of the screen nstead of manually dividing by the standard deviation of rawresiduals.

Also if you insert column using R code instead of constructor/drag-and-drop, and assuming you called your new residual variable rawresiduals (step 2 of Dmartin427 above), you can input: scale(rawresiduals) or zScores(rawresiduals)

Dmartin427

Thank you, patc3! I'll start using your shortcuts moving forward - I appreciate it!