Continuous Endpoints

A developmental toxicology study looked at 111 individual pregnant mice exposed to 0, 1, 2, 4, or 8 units of a chemical. The tabular output shows summary data for the weight at sacrifice, and indicates that mean weight decreases from 66.30g in the control group to 47.10 in the highest dose group.
Several different summary plots may be chosen from the ToxTools Plot menu; displayed here are a boxplot and scatterplot of the observed weights.

Summary Plots of the Observed Weights by Dose

From the ToxTools Model-fit menu, a power model was chosen using weight as the outcome variable with dose as the predictor variable. The "Advanced" tab in the model dialog, as seen below, can be used to set initial power parameters.

Model Dialog for Power Fit

In ToxTools, both tabular and graphical output are easily attainable for all fitted models. The window below displays tabular output for both summary information on the mouse weight data, as well as results for the Power model fit. The results show a significant non-linear trend. 

Tabular Output for Power Model

The plot below shows the fitted power model with the observed weights.



Fitted Plot for Power Model

In ToxTools, trend tests are available for binary as well as for continuous outcomes such as weight. To test whether the trend in weight is statistically significant, simply select Test from the ToxTools menu, as seen below.

Menu for Trend Tests From Toolbar

The output below shows that for these data, the trend of weight is significant. Just as for the binary case, ToxTools includes pairwise tests for continuous data to complement the trend tests. 

Output for Trend Test

As seen below, for a two-sided test, the weights in dose groups 1, 2, and 4 are not
significantly different from the weights in the control group; the difference in weights between dose group 8 and the
control group is marginally significant.

Output for Pairwise Trend Tests