1 observation for control group & 3 obs for experimental group

#1
Hello,

My background in statistics is quite limited, so I'm searching for help in order to choose which test I should use for my research.

I am comparing some formant values between two groups: native speakers (control) and L2 learners (experimental group). The participants of the experimental group have been tested three times, whereas I have only one observation for the participants of the control group. My aim is to see whether formant values of the experimental group become closer to the formant values of the control group over time.
Should I use a mixed ANOVA for both between and within-subjects? :confused:

Many thanks in advance for your help. :)

Best,
Pernelle
 
#2
Honestly I have to say that if you have so small groups i believe you run into serious troubles trying to make statistical analyses on them. But maybe others here can telle me if I'm wrong? Have you considered doing quantitative intervews or observations instead?
 
Last edited:
#3
Many thanks for your reaction.
Well, actually you're right that my groups are quite small because it is a pilot study, but I guess I was not clear --> I have 10 participants in the control group and 10 participants in the experimental group: for each participant of the control group I have ONE set of observations, and for each participant of the experimental group I have three set of observations (corresponding with three different measuring moments).
Meanwhile I have read about "mixed models", and I think I should go in this direction...?
 

Karabiner

TS Contributor
#4
My aim is to see whether formant values of the experimental group become closer to the formant values of the control group over time.
So you have only to compare the third measure of the experimental group
with the control group (t-test for independent groups, perhaps).

With kind regards

K.
 
#5
Thank you for your reaction Karabiner.
I am interested in both the end result and the evolution, so I would like to compare all three measurement moments with the control group. But if I only do three separated t-test (control group with experimental group moment1 + control with experimental moment2 + control with experimental moment3), I guess I will lose information about the evolution?