Survey covariate format

[berghsm]

I have a couple of survey covariates that I'm not how to enter in Presence. One is disturbance (none, low, medium, high) and the other is observer (3 different observers). Should I just assign values for disturbance (0,1,2,3)? What's the best way to enter observers? Thank you for the advice.

[jhines]

The observer should be entered as 3 covariates:

<pre>
obs1=1, obs2=0, obs3=0 if observer 1,
obs1=0, obs2=1, obs3=0 if observer 2,
obs1=0, obs2=0, obs3=1 if observer 3.
</pre>
This is a categorical covariate where there is no relationship among the observers.

The disturbance variable could also be entered this way, giving a separate estimate for each category (none,low,med,hi):

<pre>
distNone=1, distLow=0,distMed=0,distHi=0 if disturbance=None,
distNone=0, distLow=1,distMed=0,distHi=0 if disturbance=Low,
distNone=0, distLow=0,distMed=1,distHi=0 if disturbance=Medium,
distNone=0, distLow=0,distMed=0,distHi=1 if disturbance=High
</pre>

Disturbance could also be entered as a 'continuous' covariate:

<pre>
dist=0 if disturbance=none,
dist=1 if disturbance=low,
dist=2 if disturbance=medium,
dist=3 if disturbance=hi
</pre>

This way would be preferable if you'd like to build a model where a parameter (occupancy,detection) is a linear (on logit scale) function of disturbance. For example, detection for no-disturbance sites > detection for low-disturbance sites > detection for med-disturbance sites > detection hi-disturbance sites. You could plot detection versus disturbance on the x-axis and would get an increasing or decreasing line connecting the estimates.

The categorical way would not impose the relationship (none>low>med>hi).

Jim

[berghsm]

Thanks for the quick reply Jim. I was hoping to enter disturbance as numerical values even though it is not measured in that way. Do you think I should standardize like I've done with temperature? Also, have you ever done anything with date? I was thinking of using Julian/Ordinal date. Thanks.

[darryl]

Because of the range of the disturbance value (0-3) you shouldn't need to standardize it.