search

DistanceDevelopment / spatial-workshops

Distance sampling workshop content
http://distancesampling.org/workshops/
2 stars 0 forks source link

Non-integer group size #20

Closed dill closed 8 years ago

dill commented 8 years ago

Jason: I think we discussed this before but I've forgotten the answer (sorry).

Some values of group size in the distance data for the sperm whale surveys are non-integer:

> unique(distdata$size)
 [1] 2.000000 1.000000 4.000000 3.000000 5.000000 2.200000 2.250000 1.500000
 [9] 1.333333 4.500000 3.500000 2.666667 2.500000 4.600000 1.800000 1.600000
[17] 3.666667 7.800000 1.250000 2.333333

Was this some feature of the surveying? Something to cover in the data description lecture, for sure.

jjrob commented 8 years ago

Yes, this was a feature of the surveying. NOAA SEFSC protocol is to make multiple estimates of the group size, then average them. I can be irritating if methodology must assume integers. This might be something to mention at the workshop. IIRC, NEFSC also may make multiple estimates but puts down their "best" estimate--an integer. I don't know how they determine the "best" estimate, though.

jjrob commented 8 years ago

If you want, I can go over this in the introduction to the data. If so, assign this back to me but do not close it.

lenthomas commented 8 years ago

This is really common in distance sampling; methods do not assume integers. Cheers, Len

dill commented 8 years ago

@lenthomas Yup, no problem stats-wise, just trying to remember the explanation to we can tell participants.

@jjrob if you can go through that in the data intro that would be great. Thanks!

jjrob commented 8 years ago

Yes I will mention this. I have waited to close this issue until I decided whether to discuss this in slides or in a live exploration of the data. I will do the latter.