Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Analysis of experiment involving baseline measurements


From   Philip Jones <pjones8@uwo.ca>
To   statalist@hsphsun2.harvard.edu
Subject   st: Analysis of experiment involving baseline measurements
Date   Tue, 28 Jun 2011 05:31:28 -0400

Hello,

I am looking for some guidance from Statalisters with respect to analysis of a
randomized trial. Please forgive me in advance for the length of this message!

The trial's intervention is a teaching technique. Participants are randomly
allocated to the new technique (simulation) or the old (didactic group). The
dependent variable is continuous on a scale from 0-24 (the sum of numeric
responses on a questionnaire), and is measured at baseline ('pre') and twice
more (immediately after the intervention and after six weeks: 'post' and
'sixwk').

I have set up my data file as follows (this sample data looks normal in a
fixed-width font):

    +---------------------------------+
    |      group   pre   post   sixwk |
    |---------------------------------|
 1. |   didactic    14     24      17 |
 2. | simulation    12     23      22 |
 3. |   didactic    18     24      19 |
 4. | simulation    16     23      23 |
 5. |   didactic    10     20      16 |
    |---------------------------------|
 6. | simulation     9     21      24 |
 7. |   didactic    14     22      20 |
 8. | simulation    14     23      16 |
    +---------------------------------+

According to Senn (1), if there were only one measurement post-baseline
measurement, the preferred way of analyzing this data would be to compare the
two groups on the 'post' score while adjusting for the baseline values using
ANCOVA. However, I have two post-baseline measurements, and I'm wondering how I
might be able to take the timing of the measurements into account using a mixed
model, or whether I should bother getting that technical and instead perform two
separate comparisons:

1) 'post' result adjusted for baseline
2) 'sixwk' result adjusted for baseline

These analyses are easy enough:

regress sixwk i.group c.pre

Has anyone got an opinion as to whether or not I should go down the road of a 
mixed ANOVA to analyze these data?

Many thanks in advance for any advice/assistance!

Phil


(1) Senn S, Statistical Issues in Drug Development, 2nd ed.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index