1
$\begingroup$

Reposting this because I fixed the formatting to see if someone can actually provide a solution for this problem. I have a study design in which I am carrying out a 2x2 repeated measures ANOVA in which participants give ratings (DV) for 2 types of stimulus (Stim = 1st IV) in 2 different conditions (Conditions = 2nd IV).

Now, using the same dataframe in SPSS and in R with ezANOVA I get the same conclusions, but different F-ratios. My design is balanced, and I use the same type of sums of squares (Type 3) in both analysis, but still get different F-ratios. Anyone know why?

Here is the code for the ezANOVA in R:

ezANOVA(
  data = NewData, dv = value, wid = ID, 
  within = .(VCondition, Stimulus), type =3)

and here is the output:

$ANOVA
                   Effect DFn DFd           F           p p<.05         
2              VCondition   1  81 0.006498711 0.945754134       
3                Stimulus   1  81 8.398604264 0.005435088     * 
4     VCondition:Stimulus   1  81 6.195138581 0.014578806     * 

In contrast, with SPSS, I get different F-ratios, but the same conclusions, as seen below: SPSS OUTPUT

Here is the syntax used for the SPSS output

DATASET NAME DataSet1 WINDOW=FRONT.
GLM LC_L LC_R HC_L HC_R
  /WSFACTOR=Stim 2 Polynomial Condition 2 Polynomial
  /METHOD=SSTYPE(3)
  /PLOT=PROFILE(Cue*Stim) TYPE=LINE ERRORBAR=NO MEANREFERENCE=NO YAXIS=AUTO
  /PRINT=DESCRIPTIVE ETASQ OPOWER
  /PLOT=RESIDUALS
  /CRITERIA=ALPHA(.05)
  /WSDESIGN=Stim Cue Stim*Cue.

Does anyone have any idea what may be happening here? Why do I have different f-ratios when using the exact same test, exact same dataset, only difference being one is in SPSS and the other in ezANOVA in r?

$\endgroup$
2
  • $\begingroup$ See here for the same type of question stats.stackexchange.com/questions/565517/… Differences can occur if SPSS and R are not using the same intercept. $\endgroup$ Commented Mar 19, 2022 at 14:35
  • $\begingroup$ Something weird in the SPSS table is that you have three times a different residual. $\endgroup$ Commented Mar 19, 2022 at 15:02

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.