Turn on thread page Beta
    • Thread Starter
    Offline

    15
    ReputationRep:
    can someone explain what these two statistical tests are for and how they work?

    By the way what does 'non parametric' mean?
    Offline

    15
    ReputationRep:
    http://lmgtfy.com/?q=mann-whitney+U+test

    http://lmgtfy.com/?q=spearman+rank+test

    http://lmgtfy.com/?q=non-parametric+statistical+test

    :facepalm:
    • Thread Starter
    Offline

    15
    ReputationRep:
    I have tried to find out for myself by reading my textbook and googling it before. I just don;t understand
    • Community Assistant
    Offline

    20
    ReputationRep:
    Community Assistant
    I can't really explain Mann-Whiteny U as I've not covered it but Spearman's Rank is simple to explain.

    Basically if you have two data sets (for example, distance downstream of a river, and the speed of the river at these points) then you can use spearman's rank to determine how strong the correlation between the data sets is.

    this, for example, is an example of a strong correlation
    http://www.itl.nist.gov/div898/handb...f/scatplo2.gif
    Notice how all the points are pretty close to the trend line there?
    If you did the spearman's rank test on this data then you'd probably get a result of about 0.8 or 0.9.
    +1 is a perfect positive linear correlation. The points start near the bottom left and go up to the top right. You're never going to get this from fieldwork though.

    0 is defined as "no linear correlation". That does not mean that if you plotted the points you would notice a trend - if there was a trend it would not be a straight line, it would be some kind of curve, if anything.

    -1 is a perfect negative correlation, the points go from the top left to the bottom right, like this:
    http://www.itl.nist.gov/div898/handb...f/scatplo3.gif

    As for how it works, I assume you mean how to calculate it. You have to rank the each data set from largest to smallest values. Where 1 is the largest. Then for each coresponding value in the raw data (i.e. in the same row if you've tabulated the results), you find the difference in the ranks of the data. Then you square this value (so that it is always positive) then you find the sum of the squares.

    Then you multiply the value by 6, then divide that answer by (n^3-n) where n is the number of rows you had in your table of raw data.
    Then finally you take that answer from 1.

    As for the non parametric thing I'm not sure, although a parametric equation in maths is where rather than having one equation in terms of x and y, you have two equations where x and y are linked indirectly by a parameter, t.

    Hope that helps a bit.
    Offline

    0
    ReputationRep:
    (Original post by W.H.T)
    can someone explain what these two statistical tests are for and how they work?

    By the way what does 'non parametric' mean?

    Mann Whitney is a test for difference.
    for example the differneces in sphere of influences.

    first of all you always do a null hypothesis to show no difference.

    ok to describe it :
    • it's a stats test for difference
      telling us the difference is statistically significant
      it results in a number which we can the compare to the critical value, which is a chart you can find on internet or in a book etc.


    you can explain it by:
    • the distances are put in order seperatly for each area
      but are then ranked together. (does that make sense?)
      if 2 numbers are the same you give them the middle number, eg, 1 and 2 have the same value so they both become 1.5

      you then sum the ranks in each column and calculate the U value for each column by using the formula.

      you then take the lowest of the two U values and compare to the critical values at 95% significance level table.

      if this value is lower than the critical value you must reject your hyp.


    this may be really complicated but i just interpreted my notes for you.
    i hope it helps. message me if you need help!
    x
 
 
 
Reply
Submit reply
Turn on thread page Beta
Updated: February 1, 2010

University open days

  1. University of Cambridge
    Christ's College Undergraduate
    Wed, 26 Sep '18
  2. Norwich University of the Arts
    Undergraduate Open Days Undergraduate
    Fri, 28 Sep '18
  3. Edge Hill University
    Faculty of Health and Social Care Undergraduate
    Sat, 29 Sep '18
Poll
Which accompaniment is best?

The Student Room, Get Revising and Marked by Teachers are trading names of The Student Room Group Ltd.

Register Number: 04666380 (England and Wales), VAT No. 806 8067 22 Registered Office: International House, Queens Road, Brighton, BN1 3XE

Write a reply...
Reply
Hide
Reputation gems: You get these gems as you gain rep from other members for making good contributions and giving helpful advice.