regression - Excel Trend line (SLOPE() ) and CORREL() yields different coefficients -
i'm trying use excel coefficient 2 financial market spreads using 2 methods on data series sprd1 against data series sprd2:
1) used scatter plot , added trend line, showing r^2 (0.4052) , coefficient (0.614). trend line should using slope() coefficient...
2) used =correl(sprd1, sprd2), showing 0.637; =rsq(sprd1, sprd2), yielding 0.4052.
i understand r-sq values should pretty close. why coefficents differ? i'm trying difference in terms of excel's embedded methods or assumptions on trendline , correl.
thank much!
while both rsq , correl work same equation

the value returned rsq square of result.
i.e. rsq()=correl()^2
slope, on other hand, not use (y-mean(y))^2, nor take square root of denominator:

so give different results, depending on mean of y
Comments
Post a Comment