Analysis of variance (ANOVA)

Use Analysis of Variance (ANOVA) to determine sources of variation in statistical models. ANOVA employs several methods to estimate variance in the data. It then tests to what extent the results agree. PLA 3.0 supports two ANOVA models, that is, pure-error ANOVA and residual-error ANOVA.

Navigation: Analysis > Analytical model > Advanced settings > ANOVA model

Pure-error ANOVA

Pure-error ANOVA decomposes total error into pure error and treatment error, and treatment error into preparation, non-parallelism, regression, and lack-of-fit error:
Pure-error ANOVA decomposition
Figure 1. Pure-error ANOVA decomposition

Residual-error ANOVA

Residual-error ANOVA decomposes total error into residual error and model error, residual error into pure error and lack-of-fit error, and model error into preparation, non-parallelism, and regression error:
Residual-error ANOVA decomposition
Figure 2. Residual-error ANOVA decomposition

ANOVA block effects

Use block effects to make use of observation data structures and reduce variance of your data. Block effects allow you to explain additional variances like:

  • Rows and columns of the plate (edge and replication effects)

  • Plates of multi-plate assays (plate effects)

  • Assay randomization factors (randomized block design, Latin square design)