Measurements of twins / by Edward L. Thorndike.

  • Thorndike, Edward L. (Edward Lee), 1874-1949.
Date:
[1905]
Catalogue details

Licence: In copyright

Credit: Measurements of twins / by Edward L. Thorndike. Source: Wellcome Collection.

    42/80 (page 30)
    Table 7 Data for and results of correction for attenuation 'rg = the raw coefficient from the best obtainable deviation measures. “ trial 1 in one twin (a) with trial 1 in the other twin (b). u 2 11 “ “ 11 *< u 2 “ ** <* n t< 11 <( j “ ,l n n <« »i 2 “ “ <i n «* “ “ 2 “ “ “ “ “ “ i “ •< a if a “ “ 1 with trial 2 in the same individuals. rc4_ = the corrected r, using the formula rc = [(r^ + r2 + r3 + r4)/4]/r«. »*•*-=“ “ “ “ “ “ rc=[(r,+r,)l2]lrs. re4v == “ “ “ “ “ rc=[1.414(ry)-(r1+r2+r3 + r4)/4]/.414. rC2v=“ “ “ “ “ “ re= [1.414(r^)—(r2 + ra)2]/.414. = the probably true r, from the data at hand. ra rx r s *4 r» ^C2 — ^c*v rc2V JR 1. A test .634 .597 .616 .643 .626 .875 .71 .69 .67 .70 .69 2. Word test .569 .595 .428 .655 .78 .71 .71 3. Misspelled .754 80? 4. Addition .677 .638 .651 .693 .595 .860 .74 .75 .77 .75 .75 5. Multiplication .700 .596 .709 .600 .557 .772 .79 .85 .91 .81 .84 6. Opposites .783 .674 .793 .438 .682 .649 .90 .90? Comb, of 1, 2 and 3 .697 Comb, of 4, 5 and 6 .815 The corrected coefficients for the 9-11 and 12-14 year groups of Table 2 were obtained by the same method. I choose the ‘ empirical ’ or method of correction for the word and op- posites tests because we can not be sure that the two trials test the same thing. I choose the correction from rx and r2 rather than from r1} r2, r3 and r4 for this same reason. § 12. The Sources of Possible Error NUMBER OF CASES The unreliability of the measures of resemblance due to the num- ber of pairs is not large enough to make any of the general deduc- tions of §§ 3-7 insecure. If we use the formula P. E. (of the divergence of the true from the obtained r) = .6745(1 — r^/Vwd + r2), the P. E. for a raw coefficient of .60 from 50 pairs would be approxi- mately .045. If we use the formula, regarded by some as truer, P. E. (true-obtained r) = .6745(1 — r2)/\fn, the P. E. is approxi- mately .06. Since no claim is made that the general tendency of twins to resemblance in all mental traits is measured, it is needless to ask how unreliable the general tendency to resemblance in these six traits will be as a measure of their tendency to resemblance in the entire group of mental traits as a whole. It may be interesting to. note, however, that if these six were a random sampling from mental traits in general and if the corrected coefficients had each a P. E. (true r — obtained r) of .10 or less (as they almost surely do), we Data. < = r3 — >4 = r» — U (t u l i