|
|
I don't know so much Statistics and I don't write a very good English, so
I would like to apologize in advance for the mistakes I commit with the
language and with Statistics.
I am trying to analyze the data of my doctoral study (Dentistry) and I
would like to know if you, who know statistics, could help me with the
analysis:
I have 3 animals (dogs), each one had their mouth divided into 4
quadrants, each quadrant, with 3 teeth in each, received one of four
different treatments (3 treatments and 1 control, one treatment per
quadrant, all treatments at the same time in a mouth). They call this
split-mouth design, similar do split-plot. My variable, I thought, it was
ordinal, but it seems to be a binned scale variable (0= 0 counts of
bacteria cells, 1= less than 100,000; 2- almost equal to 100,000; 3=
between 100,000 and 1,000,000; 4 - almost equal to 1,000,000; 5= above
1,000,000; the microbiological method I used only gave me the scores, not
the exact counts, the exact values I don't have) with repeated measures (a
within-subject measure several time points in the same teeth t0=0
hour, t1=1 hour, t2= 24 hours, t3=7 days, t4= 30 days, t5= 90 days). I
don't know if in my case the treatment is nested within the factor animal
(3 levels).
I had another variables also, but with continuous data. Those, I analyzed
by means of a repeated measures ANOVA. Was my procedure right for the
continuous data? The split-mouth design was the same, so were the
treatments.
How about the other variable, the scores? What kind of variable it really
is and what is the best way to handle with these data?
Thanks in advance and regards to all,
Sandra
|
Locked
