August 29, 2021 at 10:45 pm #119184OskMalParticipant
I have quite a niche bunch of questions about the use and standardization of Presence/Absence (P/A) & Presence-only (P-O) data for a project that i working on.
I doing a litteraturer review of studies that has sample fish around a perticular kind of reef, and compiling their findings into a datasheet with presence information for each species. *See table* for example.
The crux is that not all of these studies have complete records of Presence/Absence data available, only mentions of specific instances when a fish has been found (aka presence only information). So i end up with a datasheet with a MIX of data types; some columns containing Precense/Absence data, and others Precense-only data.
Important side note is that i´m not looking at the explicit geographical locations of these observations (which rules out using methods for spatial modeling to combine the data types, as i understand can otherwise be an option.) I am only interested in the relative frequency of presence of a fish within separate location classes based on proximity to reefs (e.g. “very close”, “close “, “far away”).
So, my first questions are: Is mixing these two data-types appropriate? If not, is there a way for me to “treat” my data so that a fusion of the two data -types would be possible? Or should i go with the easier options of either convert my P/A data into P-O data, or simply treat them separately?
Secondly : Not all studies i´m looking at has data from all the distance classes i´m interested in (regardless of whether or not it is P/A data, P-O data. See “study B” and “C” respectively in *table below*). How best would i calculate the relative frequency of fish presence so that it´s comparable between location classes? Would the following equation work?
“Nr. of Presence records * (Tot. Nr. of studies in datasheet / Nr. studies with data)”
I apologize for the long post. Thanks in advance!
All the best!
// Desperate student
Study A (P/A) A A B (P/A) B B C (P-O) C C Distance class Very Close Close Far Away Very Close Close Far Away Very Close Close Far Away Fish 1 A (Absent) P P A P P Fish 2 P (Present) P A P A P
August 31, 2021 at 10:14 pm #119189Josie-Ann Le BlancParticipant
So I’ve read the information that you’ve put up and hear are the answers I have based on your questions:
1. Combining the data should be fine giving the nature of the project. Since the data is only based on one type of reef and your project is only interested in the frequency of proximity, it should be fine to combine present only with present/absent. If you do want to combine your P/A into P-O data that would be easier, smarter and safer too. Either would work, however, if you want to be safe I would do the latter.
2. The formula suggested by you looks fine as well. Since not all your studies have data from all distance classes, it is better to just work with what you have. So yes the equation should be fine.
Best of luck to you, hope this helps!
You must be logged in to reply to this topic.