I am writing a paper on an experiment involving the measuring of lung volumes (via spirometry) of participants grouped together based on their skill in apnea or skin diving. The lung volumes I want to measure are as follows:
- tidal volume (TV)
- inspiratory reserve volume (IRV)
- forced vital capacity (FVC)
- forced expiratory volume in 1 second (FEV1)
- residual volume (RV)
What implications may arise from getting the significance of the lung volumes' correlation via MANOVA? Or would it be sufficient enough to run multiple independent ANOVAs per lung volume?
I do not understand the specific groups involved, but I am guessing, from what I see, that you intend some contrast between groups with the five lung volumes as "dependent" variables. In that case, unless you have some interest in considering a superordinate construct composed of a weighting of the five variables, but rather are interested in the between groups effect of each of the five variables, I would not use MANOVA regardless of the intercorrelations of these five variables. MANOVA would be simply addressing a question that is not yours. You should accomplish some sort of per-hypothesis alpha correction in respect to testing five hypotheses. There are options here, but the most important thing is to note the "probability pyramiding" issue and take appropriate steps. Carl Huberty and I treated all of this in some detail. If you want to take a look, I have the full-text here on RG.
Article Multivariate Analysis Versus Multiple Univariate Analyses
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA