Scribd Logo
Upload

Welcome to Scribd!

  • 311 views

    Rank Correlation

    Uploaded by

    Rishab Jain 2027203
    Rank correlation is a nonparametric measure of the statistical relationship between two variables that involves ranking the data rather than using the actual values. It is calculated by finding the difference between paired ranks, squaring the differences, summing the squared differences, and using the result in a formula. Some key advantages are that it is easy to understand, can be used when only ranks and not actual data are available, and applies when the data is qualitative in nature. Limitations include difficulty with calculations when there are more than 30 data points and inability to handle grouped data.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Rank Correlation

    Uploaded by

    Rishab Jain 2027203
    311 views19 pages
    Rank correlation is a nonparametric measure of the statistical relationship between two variables that involves ranking the data rather than using the actual values. It is calculated by finding the difference between paired ranks, squaring the differences, summing the squared differences, and using the result in a formula. Some key advantages are that it is easy to understand, can be used when only ranks and not actual data are available, and applies when the data is qualitative in nature. Limitations include difficulty with calculations when there are more than 30 data points and inability to handle grouped data.

    Original Description:

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Rank correlation is a nonparametric measure of the statistical relationship between two variables that involves ranking the data rather than using the actual values. It is calculated by finding the difference between paired ranks, squaring the differences, summing the squared differences, and using the result in a formula. Some key advantages are that it is easy to understand, can be used when only ranks and not actual data are available, and applies when the data is qualitative in nature. Limitations include difficulty with calculations when there are more than 30 data points and inability to handle grouped data.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    311 views19 pages

    Rank Correlation

    Uploaded by

    Rishab Jain 2027203
    Rank correlation is a nonparametric measure of the statistical relationship between two variables that involves ranking the data rather than using the actual values. It is calculated by finding the difference between paired ranks, squaring the differences, summing the squared differences, and using the result in a formula. Some key advantages are that it is easy to understand, can be used when only ranks and not actual data are available, and applies when the data is qualitative in nature. Limitations include difficulty with calculations when there are more than 30 data points and inability to handle grouped data.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 19

    Rank Correlation

    BBA 104 1
    Outline
    • What is Rank Correlation.
    • How it is calculated.
    • What are its merits and demerits.

    BBA 104 2
    BBA 104 3
    Rank correlation
    • Since Karl Pearson’s coefficient of correlation is
    based on assumption that population is normally
    distributed there is need for a measure of
    correlation that involves no assumption about
    the parameters of the population.
    • It is possible to avoid making any assumption
    about the population by ranking the observations
    according to the size and basing the calculations
    on the ranks rather than upon original
    observations.

    BBA 104 4
    Spearman’s Correlation
    Coefficient

    where ,
    D = difference of ranks between paired itema in two
    series.
    N = number of paired observations
    BBA 104 5
    When ranks are given

    Rank
    Correlation

    When ranks are not When equal ranks are


    given given

    BBA 104 6
    When Ranks are given
    • Steps when ranks are already given :
    1. Take the difference of the two ranks i.e. (R₁-
    R₂/ R₂-R₁)
    2. Denote above difference as D
    3. Square these differences.
    4. Add up the differences to obtain ∑D2
    5. Apply the formula

    BBA 104 7
    Method of teaching A B C D E F G
    Rank by College A 2 1 5 3 4 7 6

    Rank by college B 1 3 2 4 7 5 6

    Method of Rank of A Rank of B D D2


    teaching
    A 2 1 1 1

    B 1 3 -2 4

    C 5 2 +3 9

    D 3 4 -1 1

    E 4 7 -3 9

    F 7 5 +2 4

    G 6 6 0 0

    Total 28
    BBA 104 8
    R= 1 - 6(28)
    73 -7
    = 1 – 168
    343-7
    = 1-168/336
    = 1-.5
    = + 0.5

    BBA 104 9
    When ranks are not given
    • Steps:
    1. Assign the ranks.
    2. Ranks can be assigned by taking either the highest
    value as 1 or the lowest value as 1. in either approach
    , one has to follow the same method in both the
    series.
    3. Take the difference of the two ranks i.e. (R₁-R₂/ R₂-R₁)
    4. Denote above difference as D
    5. Square these differences.
    6. Add up the differences to obtain ∑D2
    7. Apply the formula

    BBA 104 10
    X 53 98 95 81 75 61 59 55

    Y 47 25 32 37 30 40 39 45

    X R1 Y R2 D D2

    53 8 47 1 7 49

    98 1 25 8 -7 49

    95 2 32 6 -4 16

    81 3 37 5 -2 4

    75 4 30 7 -3 9

    61 5 40 3 +2 4

    59 6 39 4 +2 4

    55 7 45 2 +5 25

    Total BBA 104 160 11


    R= 1- 6*160
    83- 8
    = 1 - 960
    512-8
    = 1 - 960
    504
    = 1-1.905
    = - 0.905

    BBA 104 12
    When Ranks are Equal
    • Sometimes , it is essential to rank two or more individuals
    or entries as equal. In this situation it is mandatory to give
    each individual an average rank. For example: if two
    individuals are ranked equal at third place then they will get
    (3+4) = 3.5.
    2
    i.e. when two or more items are to be ranked equal, the rank
    assigned for purposes of calculating correlation coefficient is
    the average of the ranks which these individuals would have
    got had they differed slightly from each other.
    Adjustment is also done in the formula for the same rank.

    BBA 104 13
    BBA 104 14
    Item A b C D E F G H
    Rank 5 3 1 6 2 3 8 7
    1
    Rank 3 4 1 8 2 5 5 5
    Item 2 1
    Rank Adj R1 Rank 2 Adj R2 D D2

    A 5 5 3 3 +2 4

    B 3 (3+4)/2= 4 4 -.5 0.25


    3.5
    C 1 1 1 1 0 0

    D 6 6 8 8 -2 4

    E 2 2 2 2 0 0

    F 3 (3+54)/2= 5 (5+6+7)/3= -2.5 6.25


    3.5 6
    G 8 8 5 (5+6+7)/2= +2 4
    6
    H 7 7 5 (5+6+7)/2= +1 1
    BBA 104 15
    6
    R= 1- 6{∑D2 + 1/12(22- 2)+1/12(33 -3)}
    83- 8
    = 1- 6{19.5+1/12(8-2)+1/12(27-3)}
    504
    = 1- 6{19.5+.5+2}
    504
    =1-132/504
    =1-.262
    =+0.738
    BBA 104 16
    Question
    Marks in Accountancy Marks in statistics

    15 40
    20 30
    28 50
    12 30
    40 20
    60 10
    20 30
    80 60

    BBA 104 17
    Merits /Demerits
    • Easy to understand easy to apply
    • When data is of qualitative nature, it is the
    best method to apply.
    • Only method when ranks are given & not the
    actual data.
    • It cannot be used for finding out correlation in
    terms of grouped data.
    • Not suitable to use when n exceeds 30 as
    calculations become difficult.
    BBA 104 18
    Summary
    • After studying student is able to answer ;
    What is Rank Correlation.
    How it is calculated.
    What are its merits and demerits.

    BBA 104 19

    You might also like