Reporting interaction effects
I am a bit confused. We were told in class that if there were more than four means then we were to report them in a table. However, in the blog it says to not report the marginal means in a table. For the interaction effect, we have FOUR means, etc.
-Do we report the descriptive statistics for the Interaction in a table (it has the same information as the marginal means). this would mean that there would be two tables for section 2.
-Are we supposed to report the descriptive statistics at all?
-Do we simply report the descriptive statistics in our result write up. include (M = , SD = , n = ), p = , 95% of difference [, ] for each group ( please pardon the lack of italics)?
-Do we report the descriptive statistics for the Interaction in a table (it has the same information as the marginal means). this would mean that there would be two tables for section 2.
-Are we supposed to report the descriptive statistics at all?
-Do we simply report the descriptive statistics in our result write up. include (M = , SD = , n = ), p = , 95% of difference [, ] for each group ( please pardon the lack of italics)?
posted by Rebecca Kulzer @ 11:20 PM
4 Comments:
At 9:42 AM,
Grace Liu said…
You are correct on having more than 3 lines of statistics, you should report them in a table.
Whenever you report main effects and interaction effects, you need to include the descriptive statistics, otherwise, you don't know the direction.
As for how many table you need to use, you use one table for each analysis as Dr. Clements has answered in her earlier response. So, how many analysis/ses is/are there in question 2 in your homework.
If you have reported your descriptive statistics in a table, reporting descriptive statistics in text will consider as redundancy.
Does everything make sense to you? Hope this helps.
At 8:19 PM,
Mari said…
As noted in chapter 4 of the Publication Manual:
If you are comparing 3 or fewer means, put this in the text.
Put 4-20 in a table.
More than 20, in a figure.
The model that you actually analyzed (which is NOT the marginal means) is what goes in the table.
You need not report the marginal means when you are looking at main effects, but you must LOOK at them to know the direction of effect.
At 9:36 PM,
Nikki Frederick said…
I thought we were to ignore the "marginal means" table?
At 6:37 PM,
Mari said…
There is no marginal means table. The marginal means are included in the SPSS table of descriptive statistics, but are not to be included in your table of means and standard deviations because you analyzed the model that includes the separate cells for the complete crossing of the factors.
However, just as we looked at the marginal means to determine whether the ratings of Major managers were higher or lower than Minor managers (and even figured out where those marginal means would land in our graph), you still need to look at them in interpreting main effects.
Post a Comment
<< Home