How to best display data?

#1
Dear all,

I've got three different patient groups and I am comparing the means of 12 different parameters (blood count, cholesterol, etc.). I want to know if there are significant differences for means in each group. So I used a two-way ANOVA. Nevertheless, Graphpad also offers a mixed model (which is also a ANOVA-approach?) which seems to fit better. But is that correct? If yes, how to best display it in a graph as the variables have different units. Is a table the best option here?

Thank you in advance and greetings from germany,
Levres.
 

katxt

Active Member
#4
Because the parameters should be independent
I'm not sure what this means.
A two way ANOVA has two independent variables at various levels and one dependent response variable. For example DV could be cholesterol, and patient group and sex your two IV's. This two way ANOVA would compare the DV cholesterol between groups, and between sexes in the same analysis. Imagine a two way table. The rows and columns are patient group and sex. All the data inside the table must be measuring the same thing - in this case cholesterol. As I read your post, you two way table had 12 different sorts of response variables in it.
A one way ANOVA has one IV and one DV. Say DV, blood count, depending on IV, patient group. Each of the 12 variables will have its own one way ANOVA. The comment about adjusted significance levels is very important.
(The comment about the MANOVA is just a possible extra, not needed at this stage.)
 
Last edited: