Please help, which ANOVA should I use?

#1
Research Design:
Two shock levels: .2ma and .8a.
Two paradigms with 4 levels: NRIA, NR1a, NRCFC, NRC1a
The NRIA and NRCFC groups experienced the .2ma while the NR1a and NRC1a got the .8ma shock.
I want to see if the shock level and paradigm matters for cFos cell quantification. If it helps, I normalized the values of all four groups to Naive controls (collapsed groups as naive values).

I'm thinking that I have 2IVs (shock level and paradigm which are controlled during the experiment) and 1DV (amount of cFos expression in mouse cells)

I was gonna go with factorial anova (2x4) but not sure if this should be repeated measures since at least 2 groups experience one shock level.
 
#3
I don't think it would fall under a factorial design since you don't have every level of one factor combined with every level of the other factor. Perhaps you could consider "The NRIA and NRCFC groups experienced the .2ma while the NR1a and NRC1a got the .8ma shock." as four treatments in a one-way design?
 
#4
I don't think it would fall under a factorial design since you don't have every level of one factor combined with every level of the other factor. Perhaps you could consider "The NRIA and NRCFC groups experienced the .2ma while the NR1a and NRC1a got the .8ma shock." as four treatments in a one-way design?
Thank you for the input! I thought about this, but I'm specifically looking to see if there's an interaction between the shock level and paradigms. I was thinking of a 2x4 design. So 2 (shock level) x 4 ( NR vs 1a groups for both IA and CFC).
 

Dason

Ambassador to the humans
#7
Because your design doesn't have the information necessary to fit an interaction term.

I'm assuming the design is still
The NRIA and NRCFC groups experienced the .2ma while the NR1a and NRC1a got the .8ma shock
The interaction would tell us if the effect of shock level is different for the four treatments (or at least that's one way to think of it). You don't have any treatments receiving both levels so how could you know if the shock level had different effects.

If you actually did use a factorial design (you have observations for every combination of shock and treatment) that would be ideal.

I'm guessing there are logistical reasons you want this design but if you have time to change it I would highly encourage you to do so.
 
#8
Because your design doesn't have the information necessary to fit an interaction term.

I'm assuming the design is still


The interaction would tell us if the effect of shock level is different for the four treatments (or at least that's one way to think of it). You don't have any treatments receiving both levels so how could you know if the shock level had different effects.

If you actually did use a factorial design (you have observations for every combination of shock and treatment) that would be ideal.

I'm guessing there are logistical reasons you want this design but if you have time to change it I would highly encourage you to do so.
I should've clarified so I'm looking at two paradigms, inhibitory avoidance and context fear, IA and CFC, based on that the IA group gets .2ma AND .8mA same as CFC. I have the groups: No react and 1HR for both paradigms so 4 groups. Is that not a 2X4? Shock level (.2 and .8) compared across groups. I really just want to see if there are any differences. I was looking at some other designs and other have used multivariate stats, but that doesn't seem correct to me.