The influence of monitoring duration on power to detect tren

questions concerning analysis/theory using program PRESENCE

The influence of monitoring duration on power to detect tren

Postby leroy » Wed Mar 20, 2019 2:52 am

Hi all,

Just wondering if I should expect the duration of monitoring (i.e., number of years) to influence power to detect a certain trend (say -30 % occupancy)? Should power increase as the duration of monitoring increases?

Thanks,

Leroy.
leroy
 
Posts: 10
Joined: Thu Aug 04, 2016 4:04 am

Re: The influence of monitoring duration on power to detect

Postby jhines » Thu Mar 21, 2019 4:35 pm

Yes, increasing the duration of monitoring should increase power. Increasing the number of surveys per season should also increase power, since it increases the precision of p and psi.
jhines
 
Posts: 599
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: The influence of monitoring duration on power to detect

Postby jhines » Thu Mar 21, 2019 4:37 pm

You can test it out by running Genpres to simulate data and analyze various scenarios.
jhines
 
Posts: 599
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: The influence of monitoring duration on power to detect

Postby darryl » Thu Mar 21, 2019 4:51 pm

Increasing the number of surveys per season will help, depending on how you do that. At some stage doing more surveys per unit per season becomes inefficient in terms of occupancy estimation because the probability of a false absence becomes small, so you're better to increase the number of units surveyed per season. BUT, you do want to have enough surveys per unit per season for things to work well, where 'enough' depends on the level of occupancy and detection. As Jim says, tools like GenPRES can be useful for you to assess different options.

One further comment though is whether you really expect there to be a long-term trend (eg a decline at a constant rate) in the first place, or whether any change might be more variable. Also, if there really is a change of 30% per year, by the time you have enough data to detect it statistically with high power, it may already be too late...
darryl
 
Posts: 495
Joined: Thu Jun 12, 2003 3:04 pm
Location: Dunedin, New Zealand


Return to analysis help

Who is online

Users browsing this forum: No registered users and 12 guests

cron