Between-Days Variability

The model in the regression analysis was further refined to control for the high between-days experimental variability, seen for example in Figure 2. Simply aggregating the data over experimental days in the presence of this extra source of variability could mask or be confounded with the significance of the true effects of the treatments. Therefore it was necessary to add separate predictors for the experimental days into the model. A refined version of the model is as follows:

$$\begin{array}{ll} SpecificActivity_i\,=&\beta_0\,+\,\beta_1Se... ...NU)_i\,+\,\sum^{12}_{j=1}\beta_jDay_j\,+\,\epsilon_i\end{array}$$

Figure 2: Example of high between-days variability