+ Reply to Thread
Results 1 to 2 of 2

Thread: [Clustered Robust Standard Error] - Custering

  1. #1
    Points: 15, Level: 1
    Level completed: 29%, Points required for next Level: 35

    Thanked 0 Times in 0 Posts

    [Clustered Robust Standard Error] - Custering

    I am investigating the effects of age and exercise in protein expression in mouse motor neurons. I have four groups each with a limited number of measurements associated with it.
    Group (number of mice)

    Young (6)
    Young exercised (3)
    Old (6)
    Old exercised (4)

    Each mouse has measures of proteins from two different regions of the spinal cord- cervical (C) and lumbar (L). Each region has ~30 measures of protein associated with it (minimum=12). I would like to test the inter and intra region difference in protein expression between:
    1)different aged mice, i.e. Young vs. Old
    2)exercised and sedentary mice

    For example, I would like to know whether young L motor neuron has more/less protein than old L, old C, young C, young exercised L, and young exercised C. How should I cluster/group the data? Would a cluster represent L from one mouse and another cluster represent C from the same mouse? If it were done this way there would be 38 clusters in total (2 clusters from each animal) , is that a sufficient cluster number?

    Please help me as I am at my witts-end with this problem and can't seem to find many resources on this statistical technique. I apologise if my question has been too ambiguous, and will do all I can to clarify if needed. Also welcome to any suggestions for other tests if this test is unsuitable for this problem.

    Thank you kindly in advance for your expertise.

  2. #2
    Omega Contributor
    Points: 38,432, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Not Ames, IA
    Thanked 1,186 Times in 1,147 Posts

    Re: [Clustered Robust Standard Error] - Custering

    Cross-sectional multilevel model is the concept you are looking for. This accounts for clusters at different levels and not only takes into consideration errors but also partitioned error from other levels. This later component helps reduce type 1 errors ( incorrectly rejecting null hypothesis).
    Stop cowardice, ban guns!

  3. The Following User Says Thank You to hlsmith For This Useful Post:

    brajo083 (01-16-2015)

+ 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