+ Reply to Thread
Results 1 to 6 of 6

Thread: Power calculation vs Power of the study

  1. #1
    Points: 2,858, Level: 32
    Level completed: 72%, Points required for next Level: 42

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Power calculation vs Power of the study




    Hi everybody,

    first of all, thanks for accepting me in your forum and sorry for my bad English.

    I have a problem. I have sent a research article to be published in a medical journal.
    My sample size is of 76 individuals and the control group 157. I have analyzed the results by using the suitable tests (Chi-squared or Fisher when necessary).
    I have obtained a very significant p values with very good Odds ratios.

    The reviewers have considered this paper as very interesting but one of them ask me the following:
    "Since the sample size is small, power calculation can be performed and the power of the study should be presented"

    1. Is power calculation and power of the study the same thing?
    I have used the PS Power and Sample Size Calculations, Version 2.1.30 software (http://www.mc.vanderbilt.edu/prevmed/ps/index.htm), and I have calculated the power calculation for each analysed parameter. In spite of the small sample size, the power calculation of each one of the comparisons is really acceptable (70-95%)

    2. How can I calculate the power of the study?
    Have I to do a mean of all the calculated powers of each analysed parameters?.

    Please, This subject is urgent, because I have one week to answer the reviewer.

    Thank you very much in advance for your consideration.

    Sincerely yours,

    Nice2U

  2. #2
    TS Contributor
    Points: 17,636, Level: 84
    Level completed: 58%, Points required for next Level: 214
    JohnM's Avatar
    Posts
    1,948
    Thanks
    0
    Thanked 6 Times in 5 Posts
    1. Power calculation and power of the study are the same thing.

    2. All you need to do is determine the power of each statistical test in your study, based on the sample size and the effect size you want to detect, and then present them in some fashion in your paper. No need to "average" all of the powers.

  3. #3
    Points: 2,858, Level: 32
    Level completed: 72%, Points required for next Level: 42

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Power calculation vs Power of the study

    Thank you very much JohnM.

    When you say "All you need to do is determine the power of each statistical test in your study, based on the sample size and the effect size you want to detect", I don´t understand you, because each analysed parameter gives me a different OR.
    I have used Chi-squared test for almost all the comparisons and Fisher`s exact test when the number of affected cases or controls were equal or lower than 10.

    For Instance: cases (n=76), controls (n=157)

    genotype 1 cases 43 controls 116
    Chi-square test gives an OR of 0.46
    The power calculation (alfa 0.05, probability of exposure in controls 0.7389) is 74.88%

    genotype 2 cases 56 controls 141
    Chi-square test gives an OR of 0.32
    The power calculation (alfa 0.05, probability of exposure in controls 0.8981) is 85.82%

    and the same with other genotypes, haplotypes, alleles, etc. I have done several comparisons and each one gives a different OR and a different probability of exposure in controls,

    Please, help me with this matter because I am inexpert

  4. #4
    TS Contributor
    Points: 17,636, Level: 84
    Level completed: 58%, Points required for next Level: 214
    JohnM's Avatar
    Posts
    1,948
    Thanks
    0
    Thanked 6 Times in 5 Posts
    OK - it looks like you already did the power calculations - did you include them in the article?

  5. #5
    Points: 2,858, Level: 32
    Level completed: 72%, Points required for next Level: 42

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts
    No, I don't know how I must do it.
    One thing is the power calculation of one comparison (genotipe 1 for example) and the other is the power of the study considering that I have performed the comparison on 2 alleles, 6 genotypes, 4 haplotypes and 9 diplotypes).
    Each one gives a different power calculation because of the different OR and probability of exposure in controls.
    Thus, I have 2+6+4+9 = 21 different power calculations. I'm desolated.

    Help please

  6. #6
    TS Contributor
    Points: 17,636, Level: 84
    Level completed: 58%, Points required for next Level: 214
    JohnM's Avatar
    Posts
    1,948
    Thanks
    0
    Thanked 6 Times in 5 Posts

    I think you're making a bigger deal out of this than is necessary - just report the 21 different power calculations......

+ Reply to Thread

           




Similar Threads

  1. Power Calculation- case cotrol study
    By medsonia in forum Biostatistics
    Replies: 1
    Last Post: 08-23-2010, 12:56 PM
  2. G*Power Apriori Power Calculation
    By fraenzchen1 in forum Biostatistics
    Replies: 0
    Last Post: 02-16-2010, 06:04 PM
  3. Power Calculation - Despair!
    By stinker in forum General Discussion
    Replies: 1
    Last Post: 08-10-2009, 09:57 AM
  4. power calculation in inter-rater reliability study?
    By crawford in forum Psychology Statistics
    Replies: 1
    Last Post: 04-05-2009, 09:50 PM
  5. power calculation
    By docmaa in forum Statistics
    Replies: 2
    Last Post: 09-03-2008, 04:33 AM

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