RANK and CORRELATION

Rebecca · 11-25-2005, 06:35 AM

Greetings. I realize this is a simple statistical question, but I don't have
a statistics package (or any knowledge of statistics), but I am using EXCEL
2003. I would deeply appreciate your help because I can't figure out on my
own how to do this in EXCEL.

I have a column of numbers in random order (5, 33, 4, 76, 32, 46, etc.).
The lower numbers are "better" than the higher numbers (they are scores on a
test; higher numbers mean more mistakes). I have another column of numbers
also in random order (not ranked). The lower numbers are "better" than the
higher numbers (they are scores on another test; the higher the number, the
more the mistakes). I want to see if there is a correlation between the two
columns of test scores: that is, I want to find out if those who did well on
the first test (those with lower numbers) also did well on the second test
(those with lower numbers). Could you please tell me in simple English how
to do this?

Gary''s Student · 11-25-2005, 08:10 AM

Hi Rebecca:

Excel has a built-in function to help you. See:

http://office.microsoft.com/en-us/as...090231033.aspx

This article also outlines the mathe behind the function.
--
Gary's Student

"Rebecca" wrote:

> Greetings. I realize this is a simple statistical question, but I don't have
> a statistics package (or any knowledge of statistics), but I am using EXCEL
> 2003. I would deeply appreciate your help because I can't figure out on my
> own how to do this in EXCEL.
>
> I have a column of numbers in random order (5, 33, 4, 76, 32, 46, etc.).
> The lower numbers are "better" than the higher numbers (they are scores on a
> test; higher numbers mean more mistakes). I have another column of numbers
> also in random order (not ranked). The lower numbers are "better" than the
> higher numbers (they are scores on another test; the higher the number, the
> more the mistakes). I want to see if there is a correlation between the two
> columns of test scores: that is, I want to find out if those who did well on
> the first test (those with lower numbers) also did well on the second test
> (those with lower numbers). Could you please tell me in simple English how
> to do this?

Rebecca · 11-25-2005, 08:25 AM

Thanks, but do I have to rank the students first. Also, the link you gave me
was not very helpful. Remember, I'm a new user. Therefore, could you
explain how this should and could be done? Thanks.

"Gary''s Student" wrote:

> Hi Rebecca:
>
> Excel has a built-in function to help you. See:
>
> http://office.microsoft.com/en-us/as...090231033.aspx
>
> This article also outlines the mathe behind the function.
> --
> Gary's Student
>
>
> "Rebecca" wrote:
>
> > Greetings. I realize this is a simple statistical question, but I don't have
> > a statistics package (or any knowledge of statistics), but I am using EXCEL
> > 2003. I would deeply appreciate your help because I can't figure out on my
> > own how to do this in EXCEL.
> >
> > I have a column of numbers in random order (5, 33, 4, 76, 32, 46, etc.).
> > The lower numbers are "better" than the higher numbers (they are scores on a
> > test; higher numbers mean more mistakes). I have another column of numbers
> > also in random order (not ranked). The lower numbers are "better" than the
> > higher numbers (they are scores on another test; the higher the number, the
> > more the mistakes). I want to see if there is a correlation between the two
> > columns of test scores: that is, I want to find out if those who did well on
> > the first test (those with lower numbers) also did well on the second test
> > (those with lower numbers). Could you please tell me in simple English how
> > to do this?

Gary''s Student · 11-25-2005, 11:50 AM

Post a small sample (say ten pairs) of data and I'll give your some formulae.
--
Gary''s Student

"Rebecca" wrote:

> Thanks, but do I have to rank the students first. Also, the link you gave me
> was not very helpful. Remember, I'm a new user. Therefore, could you
> explain how this should and could be done? Thanks.
>
> "Gary''s Student" wrote:
>
> > Hi Rebecca:
> >
> > Excel has a built-in function to help you. See:
> >
> > http://office.microsoft.com/en-us/as...090231033.aspx
> >
> > This article also outlines the mathe behind the function.
> > --
> > Gary's Student
> >
> >
> > "Rebecca" wrote:
> >
> > > Greetings. I realize this is a simple statistical question, but I don't have
> > > a statistics package (or any knowledge of statistics), but I am using EXCEL
> > > 2003. I would deeply appreciate your help because I can't figure out on my
> > > own how to do this in EXCEL.
> > >
> > > I have a column of numbers in random order (5, 33, 4, 76, 32, 46, etc.).
> > > The lower numbers are "better" than the higher numbers (they are scores on a
> > > test; higher numbers mean more mistakes). I have another column of numbers
> > > also in random order (not ranked). The lower numbers are "better" than the
> > > higher numbers (they are scores on another test; the higher the number, the
> > > more the mistakes). I want to see if there is a correlation between the two
> > > columns of test scores: that is, I want to find out if those who did well on
> > > the first test (those with lower numbers) also did well on the second test
> > > (those with lower numbers). Could you please tell me in simple English how
> > > to do this?

Rebecca · 11-25-2005, 12:00 PM

24 50
32 25
33 13
50 39
109 35
51 44
28 31
53 56
38 26

In the first column, low scores are "better," and that's the case for the
other column, too. Do I have to RANK the columns first before I can
establish a correlation? I really appreciate your help.
"Gary''s Student" wrote:

> Post a small sample (say ten pairs) of data and I'll give your some formulae.
> --
> Gary''s Student
>
>
> "Rebecca" wrote:
>
> > Thanks, but do I have to rank the students first. Also, the link you gave me
> > was not very helpful. Remember, I'm a new user. Therefore, could you
> > explain how this should and could be done? Thanks.
> >
> > "Gary''s Student" wrote:
> >
> > > Hi Rebecca:
> > >
> > > Excel has a built-in function to help you. See:
> > >
> > > http://office.microsoft.com/en-us/as...090231033.aspx
> > >
> > > This article also outlines the mathe behind the function.
> > > --
> > > Gary's Student
> > >
> > >
> > > "Rebecca" wrote:
> > >
> > > > Greetings. I realize this is a simple statistical question, but I don't have
> > > > a statistics package (or any knowledge of statistics), but I am using EXCEL
> > > > 2003. I would deeply appreciate your help because I can't figure out on my
> > > > own how to do this in EXCEL.
> > > >
> > > > I have a column of numbers in random order (5, 33, 4, 76, 32, 46, etc.).
> > > > The lower numbers are "better" than the higher numbers (they are scores on a
> > > > test; higher numbers mean more mistakes). I have another column of numbers
> > > > also in random order (not ranked). The lower numbers are "better" than the
> > > > higher numbers (they are scores on another test; the higher the number, the
> > > > more the mistakes). I want to see if there is a correlation between the two
> > > > columns of test scores: that is, I want to find out if those who did well on
> > > > the first test (those with lower numbers) also did well on the second test
> > > > (those with lower numbers). Could you please tell me in simple English how
> > > > to do this?

Gary''s Student · 11-25-2005, 12:40 PM

Rank just tells you where a given piece of data is relative to the rest. So
in the first column 24 ranks first, 28 ranks second, 32 ranks third, etc.

You can arrive at the rank just by sorting. You don't have to rank before
correlating.

If your data is in A1 thru B9, then:
=CORREL(A1:A9,B1:B9) calculates the coefficient of correlation. For you
sample this is : 0.161485681

--
Gary's Student

"Rebecca" wrote:

> 24 50
> 32 25
> 33 13
> 50 39
> 109 35
> 51 44
> 28 31
> 53 56
> 38 26
>
> In the first column, low scores are "better," and that's the case for the
> other column, too. Do I have to RANK the columns first before I can
> establish a correlation? I really appreciate your help.
> "Gary''s Student" wrote:
>
> > Post a small sample (say ten pairs) of data and I'll give your some formulae.
> > --
> > Gary''s Student
> >
> >
> > "Rebecca" wrote:
> >
> > > Thanks, but do I have to rank the students first. Also, the link you gave me
> > > was not very helpful. Remember, I'm a new user. Therefore, could you
> > > explain how this should and could be done? Thanks.
> > >
> > > "Gary''s Student" wrote:
> > >
> > > > Hi Rebecca:
> > > >
> > > > Excel has a built-in function to help you. See:
> > > >
> > > > http://office.microsoft.com/en-us/as...090231033.aspx
> > > >
> > > > This article also outlines the mathe behind the function.
> > > > --
> > > > Gary's Student
> > > >
> > > >
> > > > "Rebecca" wrote:
> > > >
> > > > > Greetings. I realize this is a simple statistical question, but I don't have
> > > > > a statistics package (or any knowledge of statistics), but I am using EXCEL
> > > > > 2003. I would deeply appreciate your help because I can't figure out on my
> > > > > own how to do this in EXCEL.
> > > > >
> > > > > I have a column of numbers in random order (5, 33, 4, 76, 32, 46, etc.).
> > > > > The lower numbers are "better" than the higher numbers (they are scores on a
> > > > > test; higher numbers mean more mistakes). I have another column of numbers
> > > > > also in random order (not ranked). The lower numbers are "better" than the
> > > > > higher numbers (they are scores on another test; the higher the number, the
> > > > > more the mistakes). I want to see if there is a correlation between the two
> > > > > columns of test scores: that is, I want to find out if those who did well on
> > > > > the first test (those with lower numbers) also did well on the second test
> > > > > (those with lower numbers). Could you please tell me in simple English how
> > > > > to do this?