+ Reply to Thread
Results 1 to 5 of 5

Thread: Kruskal-Wallis or Poisson regression for parasite count data?

  1. #1
    Points: 13, Level: 1
    Level completed: 25%, Points required for next Level: 37

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Kruskal-Wallis or Poisson regression for parasite count data?




    Trying to figure out how best to carry out this analysis... So essentially I have four different species of Deer mice, infected with a high, medium, or low dose of parasite eggs. After euthanasia, I recovered larvae from different tissues of each mouse to compare differences in numbers recovered among species.

    At first I used Kruskal-Wallis to compare the effect of species on number of larvae recovered, and wasn't sure how to deal with dose levels so I just compared mice within each (more on this later). Since sample size is low and the data aren't particularly normal, I stayed away from ANOVA. But thinking about it in more detail--I AM dealing with count data (number of larvae, bounded at zero of course) as my DV and I'm not entirely sure if K-W is suitable for this?

    Looking into it more, it seems that a lot of people use Poisson regression for counts data. This might also help me get around my dose grouping problem as I can include dose as a co-variate in the glm object in R... right? There are also quite a few zeros in the data at the low dose especially.

    Which would be more suitable here? Do I have any more options? I don't love going into regression since I really lack any sort of formal background in it and my interpretation is based entirely off of self-teaching. Thanks in advance!!

  2. #2
    Omega Contributor
    Points: 38,303, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,993
    Thanks
    397
    Thanked 1,185 Times in 1,146 Posts

    Re: Kruskal-Wallis or Poisson regression for parasite count data?

    You have four types of mice
    some type of level variable?????
    and zero heavy outcome.


    Please describe the sample size and the variables above in more detail. Also, there is a zero inflated Poisson approach as well.

  3. #3
    Points: 13, Level: 1
    Level completed: 25%, Points required for next Level: 37

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Kruskal-Wallis or Poisson regression for parasite count data?

    Hey thanks a bunch for looking! I have four species (let's call them L, M, C, and P), and three doses of parasite eggs (500, 50, and 10).

    6 mice of each species were given 500 (so 24 mice in total received this dose), 6 mice of each species were given 50, and 6 mice of each species were given 10 eggs. So in total, 72 infected mice. 2 mice of each species were held as negative controls/sentinels and not infected.

    Upon euthanasia, parasite larvae were enumerated in brain, muscle, and visceral organs separately. So what I'm interested in is looking at the effect of species on these three separate "pools" of larvae.

    I say lots of zeros, but this mainly only applies to the 10 eggs group (and the brain counts of the 50 eggs group), since the protocol used to recover larvae only yields about 10% of the original dose. Zero inflated Poisson could be useful here, I assume.

  4. #4
    Omega Contributor
    Points: 38,303, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,993
    Thanks
    397
    Thanked 1,185 Times in 1,146 Posts

    Re: Kruskal-Wallis or Poisson regression for parasite count data?

    So the outcome will always be a count which is equal or less than the original dose?

  5. #5
    Points: 13, Level: 1
    Level completed: 25%, Points required for next Level: 37

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Kruskal-Wallis or Poisson regression for parasite count data?


    That's right!

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats