Fig. 5

Results from the simulation study where pairs of animals were simulated using two independent correlated random walks (a) no interaction expected, and biased correlated random walks defined as (b) group behaviour, (c) leading behaviour, (d) joint resource use, where interaction is expected. The scatter plots show the number of simulations where p < 0.05 for each of the three interaction scenarios using the case of no interaction as the null distribution. In the top row, the area of the jPPA divided by the total area of the union of the two home ranges is plotted on the x-axis against the area of the OZ divided by the union of the two home ranges on the y-axis. In the bottom row, the area of the jPPA divided by the total area of the union of the two home ranges is plotted on the x-axis against the proportion of time each simulation spent in the interative phase (pInt) on the y-axis. The pInt value is always zero for the Null scenario, and thus this plot is not included. Black points indicate where both the jPPA and naïve OZ measure correctly identify interaction, while blue indicates where the OZ measure did not identify interaction. Red points indicate where jPPA did not identify interaction, but the OZ measure did, while green points indicate where both methods did not identify interaction