Re: Differences between 'trend line' and 'correlation'?
- From: jlclyde <jlclyde@xxxxxxxxxxxxxx>
- Date: Fri, 4 Jan 2008 10:17:59 -0800 (PST)
On Jan 4, 10:31 am, Donal P <Don...@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
I'm using Excel 2003 under XP Professional to examine the relationship, if
any, between two sets of data - let's call them 'audit score' and 'profit'..
I have 74 data points for each set. When I graph the data points in an XY
scatterplot and then add a trendline to the data series, I get a slope for
the line of .033. However, when I use the correlation function in data
analysis, I get a .0185.
I thought these two numbers should be the same. Although the difference is
not very large, the difference gets noticeable when I look at certain subsets
of the data; on one subset, the numbers come out identical to the fourth
decimal place.
Any thoughts?
If you are trying to find the correlation between two things, you are
going the right way. Do the XY scatter plot and add the trend line
and the R squared number. This is the amount that Y is directly
affected by X. In Six Sigma, you say that there is strong correlation
when it gets to around .75. Anything lower then that, you obviously
will ahve noise or other varaibles also working to raise or lower the
Y.
Jay
.
- Prev by Date: Re: EXCEL read a text string as scientific notation
- Next by Date: Re: Chart Wizard
- Previous by thread: VBA Consolidate Data
- Next by thread: Re: Differences between 'trend line' and 'correlation'?
- Index(es):
Relevant Pages
|
Loading
