Significance asterisks

[singmann]

Significance asterisks

[singmann]

I think I did this to be compatible with APA standard. However, I cannot find my APA manual copy right now to check this. But I am pretty sure about it.
If you want the R default just use the anova() method and not the nice() method for an object of class afex_aov (note that the nice() method is called for printing). e.g.:

require(afex)

data(md_12.1)
a <- aov_ez("id", "rt", md_12.1, within = c("angle", "noise"), 
       anova_table=list(correction = "HF", es = "none"))
nice(a)
# Anova Table (Type 3 tests)
# 
# Response: rt
#        Effect    df     MSE         F p.value
# 1       angle 2, 18 3560.00 40.72 ***  <.0001
# 2       noise  1, 9 8460.00 33.77 ***   .0003
# 3 angle:noise 2, 18 1160.00 45.31 ***  <.0001
# ---
# Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘+’ 0.1 ‘ ’ 1
# 
# Sphericity correction method: HF 
# Warning message:
# HF eps > 1 treated as 1 

anova(a, correction = "HF")
# Anova Table (Type 3 tests)
# 
# Response: rt
#             num Df den Df  MSE      F     ges    Pr(>F)    
# angle            2     18 3560 40.719 0.39012 2.087e-07 ***
# noise            1      9 8460 33.766 0.38660  0.000256 ***
# angle:noise      2     18 1160 45.310 0.18827 9.424e-08 ***
# ---
# Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
# Warning message:
# HF eps > 1 treated as 1

I think this solves this issue.

[singmann]

PS: R default is also not to round the p-values in the ways that nice() does. I very much prefer the rounding as well.

[SunByrne]

I prefer the rounded p-values but I'm looking at the APA manual right in front of me and I don't see anything about asterisks on F-values.

[singmann]

I think an example ANOVA table was in the 5th edition handbook. But I cannot find a copy of it around here. Only the new 6th edition.

But note that in the example tables in the new version, the asterisks are usually also with the test statistics. This is also done in afex. I agree that there are no tables that have both test statistics and p-values, but still I am convinced it is better like that.

[SunByrne]

The APA manual specifies that in tables, columns of numbers should be decimal-aligned. Putting the asterisks on makes them not decimal-aligned.

[singmann]

I feel like the alignment question cannot be solve in R (or not without a lot of headache). This needs to be done with whatever software you write your papers. If you want no stars, you can easily ask for no stars.

As I said before, if you want the R default, use the anova() method. If you want what I find better, use the nice() method. If you want something that is in between the R default and the nice() method, then you have to program it on your own. I like the stars with the F-values and I believe this is how it was supposed to be in the previous APA version. You will need a really good argument to change the default nice() output and so far your arguments were not really strong.