Citation indices potentially are very invalid because they are susceptible to manipulations and irrelevant weightings when utilized as indices of merit. The whole process of citing references is very idiosyncratic and a reflection of author biases because references can be cited on either side of any contention. On the basis of more than 30 years of editorial experience, a listing is made of ways in which eitation indices can be manipulated by factors irrelevant to merit. Genuine merit can be assessed only by long‐term validity and predictive usefulness after many replications by independent studies.

extremely high correlation of over .99 with the other two recall time groups. The t test reveals that a sample drawn from a population with zero correlation to one of the recall time groups would result in a correlation coefficient as high as those actually obtained less than one percent of the time. Therefore, a highly significant, positive correlation between the recall time groups is clearly shown. Using "expected symptoms'' versus "unexpected and equal symptoms'), there is a t least a .27 probability that the differences between comparable percentages in any two recall time groups are chance differences. Therefore, coinparable percentages in the recall time groups are not significantly different. CONCLUSIONSSince each recall time group is significantly correlated with the remaining ones and since comparable percentages are not significantly different, it appears that the length of time interval between the present and the recalled contrast period makes no difference in the reporting of past symptoms elicited from wives of alcoholics by the Index of Psycho-physiological Disturbance. The results suggest that similar techniques may be applicable for the analysis of past-present contrasts with other populations and other psychological tests.

No abstract

No abstract

No abstract