0
$\begingroup$

  • I am in the early stages of designing a study involving a moose (Alces alces) population.

  • One of the important objects of measure are their moose pellets, which I want to take a battery of measurements on including their dry weight.

  • I recently picked one up off the forest floor and subjectively it felt quite light.

  • Originally I had intended to buy a ~$100.00 weigh scale with 0.01g precision.

  • If the pellets are too light to be detected by the balance, that is a obviously problem.

  • But I am also interested in variation, so it would also be a problem if most of the variation among pellet weights occurs below the precision of the balance.

Has anyone published statistics on the weight and variation in weight of moose pellets?

Improve this question
$\endgroup$
3
  • $\begingroup$ Why individual pellets rather than pellet groups? But 0.01 g is quite small... $\endgroup$
    – Bryan Krause ♦
    Commented Apr 18, 2022 at 18:06
  • $\begingroup$ @BryanKrause If I measure aggregates of pellets then I cannot study their individual variation. Why I care about their individual variation is particular to the study questions. $\endgroup$
    – Galen
    Commented Apr 18, 2022 at 18:12
  • $\begingroup$ @BryanKrause The precision of 0.01 g is quite small, subjectively. I want to know ahead of time, if possible, that I have sufficient precision. $\endgroup$
    – Galen
    Commented Apr 18, 2022 at 18:12

1 Answer 1

Reset to default
4
$\begingroup$

MacCracken and Van Ballenberge, 1987 report that pellet size varies a bit by sex and age; male pellet dry weight (all measurements are dry) averaged 1.9 grams (range 0.9-2.9), female 1.3 g (range 0.6-2.0), yearling 0.8 g (range 0.6-1.7).

Unfortunately they do not report SDs but instead SEs of 0.01 for every group; we can try to reverse engineer the SD but with one significant digit there isn't much precision to work from. SE = SD/sqrt(N), their N were 34/36 for the male/female adult groups, so let's just base it on 35...

0.005 * sqrt(35) = 0.030 g

0.015 * sqrt(35) = 0.089 g

The SD could be somewhere in that general range, but based on the range I suspect it may actually be much higher, and that they used a larger N for calculating the SE (perhaps many pellets per group). The other possibility is that the range is mostly due to outliers. As a rough estimate from the ranges, you could guess the SD is around 1/6th of the range (if normally distributed, 99.7% is within +/- 3 SD), with ranges around 2.0 to 1.4, that would give you a SD around 0.2 to 0.3 grams.

They used a scale with 0.01 g precision; seems sufficient for this work. Dry weight is certainly the standard for this sort of thing, so you may want to look at equipment for drying as well.

MacCracken, J. G., & Van Ballenberge, V. (1987). Age-and sex-related differences in fecal pellet dimensions of moose. The Journal of wildlife management, 360-364.

Improve this answer
$\endgroup$
5
  • 1
    $\begingroup$ (+1) Thank you for (1) finding a citation and (2) clarifying whether it was dry/wet weight. $\endgroup$
    – Galen
    Commented Apr 18, 2022 at 18:21
  • $\begingroup$ Nice use of the definition of standard error of the mean, btw. $\endgroup$
    – Galen
    Commented Apr 18, 2022 at 18:22
  • 1
    $\begingroup$ @AgnesianOperator I think likely my original estimates for SD based on SE are very low; I suspect they put in measurements for many pellets per group and calculated their SE across all of them, so the N=35 is wrong. I think my new estimate based on the ranges will be closer to what you actually observe. $\endgroup$
    – Bryan Krause ♦
    Commented Apr 18, 2022 at 18:28
  • $\begingroup$ Thanks, I appreciate that update to your answer. Regrettably, I can only upvote your answer once. I will read the study to understand their statistics in detail. On the face of it I am thinking of using a random intercept model to allow different piles of scat to be from possibly-different individuals whose age/sex might differ. $\endgroup$
    – Galen
    Commented Apr 18, 2022 at 18:32
  • $\begingroup$ @AgnesianOperator Yes, I would certainly suggest mixed effects models for something like that. I don't think you'll find much use in the stats for this particular paper, as although they are doing some interesting stuff for 1987, statistics has come a long way, particularly in how biologists use statistics. $\endgroup$
    – Bryan Krause ♦
    Commented Apr 18, 2022 at 18:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .