+ Reply to Thread
Results 1 to 3 of 3

Thread: Should I use chi squared to test these observed vs expected ratios?

  1. #1
    Points: 4, Level: 1
    Level completed: 7%, Points required for next Level: 46

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Should I use chi squared to test these observed vs expected ratios?




    I'm doing a study about the height:radius ratios of limpets on exposed and sheltered shores (The ratio is calculated by dividing height by radius). There is a theoretical 'optimum' ratio which is 1.06. I have 2 samples: 219 limpets from an exposed shore and 299 limpets from a sheltered shore. I calculated the height:radius ratio of all the limpets sampled.
    I want to statistically test if my observed ratios of the limpets are significantly different to the expected ratios of 1.06. I want to do this for the sheltered samples and then exposed samples separately to see if they have different results.

    Here are my first 5 data for sheltered limpets...

    observed ratios:
    1.35
    1.64
    1.27
    1.18
    1.63

    expected ratios:
    1.06
    1.06
    1.06
    1.06
    1.06

    Should I use chi squared goodness of fit test? I'm not sure because I thought the variables had to be categorical in order to do a chi sqaured test, but isn't ratio a continuous variable? would it be better to use a Mann Whitney U test? I have to use a non-parametric test because the data are not normally distributed. Any help is appreciated, thanks x
    Last edited by Little.fish; 09-22-2017 at 11:28 AM.

  2. #2
    Points: 1,974, Level: 26
    Level completed: 74%, Points required for next Level: 26

    Location
    New Zealand
    Posts
    227
    Thanks
    3
    Thanked 48 Times in 47 Posts

    Re: Should I use chi squared to test these observed vs expected ratios?

    You are right. Chi square only works with count data. Your data is calculated, not counts. You need to do a one sample test against the value 1.06.
    There are non parametric one sample tests such as the Wilcoxon signed-rank test, but you might find that there is a transformation which will make your data normal. In any event, with such large samples, a one sample t test will give reasonable results.

  3. #3
    TS Contributor
    Points: 12,287, Level: 72
    Level completed: 60%, Points required for next Level: 163
    rogojel's Avatar
    Location
    I work in Europe, live in Hungary
    Posts
    1,471
    Thanks
    160
    Thanked 332 Times in 312 Posts

    Re: Should I use chi squared to test these observed vs expected ratios?


    Twi quick remarks to what katxt was telling you:
    1. in case of a nonparametric test you will get results concerning the nedian, not the mean.
    2. If you transform the data remember to transform the threshold as well (the 1.06)

    regards

+ Reply to 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