Integrity | 104.9 | 130.1 | 117.5 (Little Cooperativeness | 125.0 | 113.5 | 119.3 Agreement) Cheerfulness | 130.0 | 112.0 | 121.0 Kindliness | 120.0 | 125.7 | 122.9 ------------------------------------------------------------ It is to be noted that certain traits show small divergence in both cases. Thus intellect, quickness, originality and efficiency have low measures of variability, both for the sorority members and for the men of science. The average percentages of these four traits are, in the order named, 93.6, 89.0, 86.2, and 83.7. These, it is to be observed, are the traits which are likely to yield objective products. The more personal, social and moral traits, however, such as cooperativeness, unselfishness, kindliness, cheerfulness, and integrity, show large divergence of the individual judgment with both groups. The average measures of variability for these traits, in the order named, are 119.3, 110.5, 122.9, 121.0, and 117.5. There is another group of traits which, while showing only about average variability with one group, show close agreement in the other: such as will, judgment, perseverance, leadership and breadth. These, it is clear, are more nearly like the objective than they are like the personal traits. Then there are several traits which, while showing only average variability with one group, show large divergences in the other, such as courage and independence. These would seem to be more nearly like the more strictly personal traits. Norsworthy points out that the traits about which inquiries are commonly made in recommendation blanks sent out by teachers' agencies, employment bureaus, and employers, tend to be those on which, according to her results, individual opinion is least reliable. Traits such as originality, judgment, clearness and quickness, on which judgments are most unanimous and consistent, are usually omitted from these blanks. This indicates the desirability of a more careful examination into the general validity of this type of judgment. Here, then, as in all the other topics that we have had occasion to discuss, we find that our present knowledge is far from adequate to meet the demands of practical life. Available results are tentative only, but they are so suggestive as to afford a series of interesting problems for further investigation. The validity of judgments of associates varies with the judge, with the trait in question, and with the person who is being estimated. But it does not vary at random; it varies in what seem to be fairly definite, common, and determinable ways. That we do not know more about the precise nature of these variations means merely that few persons have taken the trouble to inquire into the matter.