When design matters more than sample size
In many applied research projects—from precision medicine to ecology or industry—we encounter experimental designs where repeated measurements are taken from each individual. For example:
A patient whose biomarker evolution is measured before, during, and after a treatment.
- A fish from which samples are taken from different organs.
- An industrial machine whose parameters are recorded at various stages of its work cycle.
These cases, when multiple response variables are studied (sometimes thousands, tens of thousands, or more), are known as multivariate repeated measures designs. They are very powerful because they allow for a reduction in individual variability and the derivation of more robust conclusions. But they also hide a challenge: if not analyzed correctly, the results can be misleading.
The problem: incorrectly coded individuals
In such a design, each individual belongs to only one group (e.g., treatment or control). This means the individual is nested within the group. If the analysis treats them as if they were a crossed factor, the statistical calculations are distorted and false positives appear.
This error is more common than it seems: small differences in how the method is applied can lead to opposite conclusions in the same study.
The solution: ASCA+
Advantages of ASCA+ over other approaches (such as LiMM-PCA):
Greater statistical power
It better detects real effects
Much more efficient
Power curves are generated in minutes
Robust
It works well even with unbalanced data or data that does not follow a normal distribution
Why it matters to you
- Your results reflect real effects and not statistical artifacts.
- Individual variability does not mask or inflate the conclusions.
- The analysis is reliable and reproducible, even under complex experimental conditions.
From theory to software
At Datharsis, we have incorporated this logic directly into our tools::
- The software automatically recognizes whether a factor should be treated as nested or crossed.
- It selects the appropriate statistic based on the experimental design.
- It implements ASCA+ in an optimized way, reducing times and avoiding common errors.
In this way, we transform a complex statistical problem into a simple, transparent, and reliable analysis for our clients.
Do you want to ensure that your studies with repeated measures are analyzed rigorously and without statistical errors?
At Datharsis, we help you design and analyze your experiments with the best practices in computational statistics, combining univariate and multivariate strategies
