In studies in which one of the variables recorded is the minimum neighbor distance of a number of focal animals, how do people handle cases where a neighbor cannot be identified? Excluding them from the analysis does not seem a reasonable strategy because you would then missing useful information.
Position infrared motion camera ($100US) along suspected home range routes.
Moultrie Series A works well. Then a proper trap can be used if necessary.
You move your quadrats or sampling to where the objects are. If objects aren't present, you can't measure them!
Of course if a sample site is verified to be devoid of any subjects over time that is your result at that specific sample site. You have neither resident or neighbors. If the subject of the study is not the site, but an animal/animals, you change your quadrant size, its location at the site, or leave the area entirely and find a completely new study location.
However, the question as written appears to be one of neighbor "identification" rather than one of lack of neighbor existence. This is also implied in the second part of the original question where it is stated, "Excluding them from the analysis". You cannot exclude something that is not there. Again this implies something was observed-- some sort of evidence of a neighbor or different neighbors is evident-- a path, scat, browsed vegetation, tracks, prey remains.
Furthermore, what is the reason for the study? Sometimes you do not want to find a neighbor, e.g., too small a habitat to accommodate more than one individual of a species. Again, the way the original question is written I assume one subject individual has already been identified in the study area-- because without a resident you cannot even have a neighbor. He is wondering how to address unidentified neighbors according to the question.
Thanks a lot for your answer and I apologize if my question was not correctly formulated. The data I'm trying to analyze come from a field study of lizard behavior. Briefly, I run a number of transects and every time I encountered a lizard (designated the focal lizard) I noted its behavior, what other lizards were found in the surroundings, and the distance between the focal lizard and the one nearest to it. In a large percentage of cases, I could not detect any lizards near the focal one. The focus of the study is on signaling behavior, so cases in which a lizard performed a display without any obvious receiver nearby are of particular interest. Of course, I realize that in theory there always is a nearest neighbor, even if it is several hundred meters from the focal lizard and unlikely to interact with it, but I declared no neighbor present when I could not see/hear any other lizards within a radius of a few meters (variable depending on the presence of rocks, vegetation, ...). I could analyze the results comparing the number of cases in which a neighbor was or was not present (count data), but given that I measured the distance to the nearest neighbor when one was identified, I also wanted to attempt an analysis using distances. The problem is, what distance do I assign to those cases in which a neighbor was not identified? I could give them an arbitrarily large, nominal distance (e.g. 10 m), but I was wondering what others had done in similar cases.
Thanks again,
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning