Kennaway paper; correlations

CHUCK TUCKER 970421
   I am behind on my reading but I did notice comments on this paper
   (does anyone mind posting the paper on CSGNET- I don't have all
   those fancy browers - still have gas powered computer!) but I
   wondered how what is said by Kennaway that is not implied by
   the stuff below that I have posting on all the lists that talks
   about statistics (especially sociology lists). BTW, if you want
   some examples of the misuse of statistics just examine ANY sociology
   sociology journal. Regards, Chuck

                 A TABLE SHOWING THE RELATIONSHIP BETWEEN

                     SEVERAL DESCRIPTIVE STATISTICS*

···

____________________________________________

                   r r2 k2 k E
               ____________________________________________

                 1.00 1.00 .00 .00 100 %
                  .9995 .999 .001 .032 97 %
                  .9987 .997 .003 .054 95 %
                  .995 .99 .01 .099 90 %
                  .954 .91 .09 .299 70 %
                  .90 .81 .19 .435 56 %
                  .87 .756 .244 .493 51 %
                  .865 .748 .252 .50 50 %
                  .80 .64 .36 .60 40 %
                  .75 .56 .44 .66 34 %
                  .71 .50 .50 .70 30 %
                  .65 .42 .58 .76 24 %
                  .60 .36 .64 .80 20 %
                  .55 .30 .70 .83 17 %
                  .50 .25 .75 .87 13 %
                  .45 .20 .80 .89 11 %
                  .40 .16 .84 .92 8 %
                  .35 .12 .88 .94 6 %
                  .30 .09 .91 .95 5 %
                  .25 .06 .94 .97 3 %
                  .20 .04 .96 .98 2 %
                  .15 .02 .98 .99 1 %
                  .10 .01 .99 .995 0 %
                  .00 .00 1.00 1.00 0 %

DEFINITION AND INTERPRETATION OF THESE STATISTICS**

All of these measures describe two variables (X, Y) within a particular
sample:

     r is a correlation (or coefficient of correlation) which
     describes the linear association of one variable with another.
     It can also be characterized as "... a relative measure of the
     degree of association between two series " of values for two
     variables. It varies between 1 (perfect positive correlation) to
     -1 (perfect negative correlation). The closer this measure is to
     a perfect correlation the more confidence one has in "predicting"
     the values of one variable from another variable.

     r2 is a measure of "explained" variance (or coefficient of
     determination) which describes "shared" variation or the amount
     of variance that one variable is "explained" by the other
     variable or the proportion of the sum of y2 that is dependent on
     the regression of Y on X. The larger the numerical value of this
     measure the more confidence one has in "predicting" the values of
     one variable from another.

     k2 is a measure of "unexplained" variance (or coefficient of
     nondetermination) which describes "unshared" variation or the
     amount of variance that one variable is NOT "explained" by the
     other variable or the proportion of the sum of y2 that is
     independent of the regression of Y on X. The smaller the
     numerical value of this measure the more confidence that one has
     in "predicting" the values of one variable from another.

     k is a measure (called coefficient of alienation) which
     describes the lack of linear association of one variable with
     another or the ratio of the standard error of estimate to the
     standard deviation of the variable. The smaller the numerical
     value of this measure the more confidence one has in "predicting"
     the values of one variable from another.

     E this measure is computed by (1-k)100 and is called an "index
     of forecasting efficiency" (Downie and Heath, 1965: 226) and
     indicates the "improvement" for a prediction by knowing the
     coefficient of correlation (r) for two variables as contrasted
     with knowing nothing about the linear association of the two
     variables. For example, with a coefficient of correlation of .71
     one can "predict" the values of one variable from another 30%
     better (on the average) than one could "predict" those values
     WITHOUT any knowledge of the relationship between the two
     variables OR one has decreased the size of the "error of
     prediction" by 30% (on the average) by knowing that the
     correlation of the two variables is .71.

                                  REFERENCES

          Arkin, Herbert and Raymond R. Colton. 1956. Statistical
     Methods. College Outline series, Fourth Edition, Revised.

          Downie, N. M. and R. W. Heath. 1965. Basic Statistical
     Methods. Second Edition. New York: Harper and Row.
     ______________________________
     *compiled by Charles W. Tucker with the encouragment and
     assistance of the Control Systems Group (especially Gary Cziko)
     and the comments of Jimy Sanders. Other comments appreciated -
     N050024 AT UNIVSCVM.BITNET

     **It should be noted that these descriptions and interpretations,
     especially those involving "predictions" are limited to a
     particular sample; if another sample is not a random sample from
     the same population then predictions about the other variable
     ("Y") will be unpredictably worse than the original sample.

[From Bruce Gregory (970421.1120 EST)]

                 A TABLE SHOWING THE RELATIONSHIP BETWEEN

                     SEVERAL DESCRIPTIVE STATISTICS*

Thanks! This is really helpful for those of us who are
"statistically challenged"!

Bruce

[Hans Blom, 970424]

(CHUCK TUCKER 970421)

Chuck, in your TABLE SHOWING THE RELATIONSHIP BETWEEN SEVERAL
DESCRIPTIVE STATISTICS, you give a simple "translation" between r,
r2, k2, k and E which essentially shows that these measures are all
equivalent. Is that right? If so, why are they all in use and not
just one of them?

Greetings,

Hans