CH12 ANOVA of repeated measurement data. ppt.ppt

Chapter 12
ANOVA of Repeated Measurement Data
§1 Character of Repeated Measurement Data
Content
Data characteristic
Analysis of two factors and two levels
Analysis of two factors and several levels
Notices
Objective: Inference the effects of treatment, time, and treatment*time on experimental objects
Character of data:
treatment factor: g (≥1 )levels and n experimental objectives in each level, add up to gn experimental objectives.
time factor: m valves at m time in each experimental
objective, add up to gnm valves.
Method: ANOVA
I premeasure-postmeasure design
It is particular case in repeated measurement data and is also called single group premeasure-postmeasure design. g=1, m=2.
table12-1 BHP patient’s diastolic pressure in pretreatment and post-treatment(mmHg)
Table 3-3 measurement result of fat content in lactic acid (%)
Compare:
Differences with paired test:
experimental units in the same pair in
the paired test may be managed at random.
Two experimental units can be observed at the same time. We pare the difference of the group.
The results of premeasure-postmeasure design can’t be observed at the same time
although it can be arranged at pre-post experiment. But in substance it pared with the difference of pre-post experiment. We infer the treatment to be effective on condition that we assume the time don’t affect the results.
paired/matched t-test requires the results of two experimental units in the same pair and difference to be fit for independent. The difference obeys to normal distribution.
Two results of premeasure-postmeasure
design mon not to independent with differences. The first time result is negative correlation to the difference in most cases. Table 12-1 as follows, the correlation of diastolic pressure before treatment with
the difference is .
3. It is inferred the effective of treatment by average difference in paired design. While we can analyze average difference and analyze correlation and regress