For several years the authorities of the Naval Academy have been conducting experiments to determine some means of measuring accurately what the psychologists would call native intelligence. They have made these tests not as an academic study of psychology but wholly as a practical effort to determine:
(a) Whether native intelligence can be measured accurately by means of a short, simple test.
(b) Whether such measurements can properly be used as a bar to entrance, or as a contributory cause for dismissal after admission.
(c) Whether the entrance requirements are adequate.
(d) Whether the scholastic standards of the course are sufficient.
(e) Whether there is a relationship between native intelligence and misconduct.
(f) Whether there is a relationship between native intelligence and relative standing for the four years of the course.
The results of the tests already completed indicate that intelligence is measurable. They require the serious consideration, eventually, of refusing admission to candidates of lower intelligence; show that the psychometric mark is a valuable assistance in diagnosing a case of scholastic failure; prove that the standards of the Academy are in keeping with what the country has a right to expect from students in its pay; indicate that the entrance requirements as recently revised are adequate; show that there is little relationship between intelligence and misconduct, probably because the entrance requirements are a bar to students of really low mentality; and show a marked relationship between high native intelligence and an ability to graduate near the top of a class.
Only in a college where detailed records are kept is it possible to compare native intelligence with factors other than a mere ability to graduate.
The daily marking of midshipmen necessitates minute record keeping in order that, at the end of the four-years' course, the authorities may say that Midshipman A is more intelligent than Midshipman B and deserves to be promoted ahead of him. These accurate records provide a fertile field for the study of psychometric relationships.
Early in his plebe year each midshipman is required to undergo a simple test which consists of but five blocks of questions and takes less than half an hour to complete. No particular education is necessary in order to pass the test. No study or general knowledge is required; in fact, none is desired. For the measurement of natural intelligence presupposes a mind untutored in the particular exercise.
Among the tests used are the "dissected sentence," the "missing word," the "cancellation," the numeral test for memory, and the "code" test. All of these were furnished originally by Dr. A. W. Stearns, a psychologist who had a wide experience in this sort of work in the Navy during the World War. They have been altered from time to time but not in their major features. In the dissected sentence test words must be transposed into a sensible sentence. For instance:
Seldom — forever — good — lasts — luck which is, of course:
Good luck seldom lasts forever.
In what may be called the "missing word" test, certain words are omitted and must be supplied. A very simple example is:
The boy fell out of the — and — his hand.
Visualizing a code and then writing a short sentence in it is one of the hardest of the tests. The partial code on the following page is not the one used at the Naval Academy but is typical and will serve by way of example.
It can never be said with certainty that a midshipman of superior intelligence will succeed or that one of less natural capacity will not. There are external factors common to a percentage of each psychometric group such as dislike for naval restrictions, home sickness, love, and physical disability, pulling or pushing the midshipmen and upsetting any absolute measuring of one individual's chance of graduating. The records show that these factors exert a reasonably equal pressure on all groups and therefore they were disregarded in the analysis of groups. Midshipmen who left the Academy by voluntary resignation or for physical disability were removed from the compilation of psychometric marks and were not charged as either successes or failures.
The remainder of any one class were divided into six groups (A to F), determined by the marks made on the psychometric tests. Classes which were examined later were divided into eleven groups (A to K) to get a finer differentiation between groups. "A" group contained those midshipmen who made psychometric marks between 95% and I00%, and the lower groups—B, C, D, etc.—were scaled down in increments of 5%, except that in the six-group division all those midshipmen who made below 75% were placed in group "F," while in the eleven-group division the increments of 5% were continued down to group "K," which included all those midshipmen making marks below 50%.
In order to have the data in an easily readable form, graphs have been constructed for the respective classes, showing the relationships between native intelligence and the various factors studied. Three typical graphs are reproduced herewith, which show at a glance that native intelligence can be and is being measured. They also indicate the reasons for claiming the results of a previous paragraph. They show that the men who make the higher marks on the psychometric tests have, as a group, a markedly better chance of completing the course than the men whose psychometric marks place them in the lower groups.
STUDY OF A CLASS AFTER ONE YEAR AT THE NAVAL ACADEMY
See graph-A
Curve 1 shows the original class distributed according to the marks made on the psychometric test.
Curve 2 shows the distribution of those remaining in the class after the completion of one year (their fourth-class year).
Curve 3 shows the distribution of those who failed during the year, arranged according to their psychometric marks.
Ordinates for curves 1 and 2 are percentages of class. Ordinates for curve 3 are percentages of respective groups. The letters "A," "B," "C," etc., represent psychometric groups. The midshipmen who made a mark of 95% or better on the psychometric examination compose "A" group. The other groups represent 5% each on the psychometric examination on a descending scale except "K" group which contains all those midshipmen who made examination marks below 50%. Eight hundred seventy-two of this class were examined shortly after admission. After one year 683 remained.
EXAMPLE:
(a) To determine first point on curve i (original distribution), 872 midshipmen were examined of whom 50 made psychometric marks between 95% and i00%, which placed them in group "A." Thus 50/872 of the class, or 6%, were in group "A."
(b) To determine first point on curve 2 (distribution after one year), 683 of the original 872 remained in the class after one year; 49 of these 683 were "A" group men. Thus 49/683, or 7%, of the class were "A" group men.
(c) To determine first point on curve 3 (curve of failures), group "A" originally contained so midshipmen. After one year 49 of them remained and the one who left did not fail. None failed. Thus the percentage of failures for "A" group was 0/49 — 0%.
The following table gives the data for curve 3 and is inserted for the benefit of those who dislike reading curves.
Group | Percentage of group failed |
A | 0 |
B | 7 |
C | 10 |
D | 13 |
E | 20 |
F | 16 |
G | 19 |
H | 25 |
I | 44 |
J | 56 |
K | 53 |
NOTE: That there is a nearly regular increase in failures varying as the psychometric mark.
From an analysis of these curves the worth of the psychometric test may be proved. They show that:
1. Few men fail who are able to make good marks on the intelligence test.
In the class of 1927 none failed from group "A" and only 7% from group "B."
2. Many fail from those who make poor marks on the psychometric test.
In one year at the Naval Academy there were eliminated by scholastic failure 44% from group "I," 56% from group "J" and 53% from group "K."
3. The entrance requirements have eliminated most of the slower, inaccurate thinkers of low native intelligence.
Only 17 out of 872, or 2%, received grades of less than 5o% on the intelligence tests.
4. Any group of men contains a few of superior intelligence, a large number of average intelligence, and a few of inferior intelligence.
Whereas there were 6% in group "A," there were 16%, 17% and 15% of the class in groups "C," "D," and "E," respectively. Group "J" contained but 2.5% of the class. Yet each of these groups covered the same variation (5%) in examination mark. Only 19% of the class made grades between 69 and 50, while 60% of the class made from 89 to 70.
5. The psychometric test does measure intelligence insofar as ability to secure passing marks goes.
Relatively few failed from groups "A"-"D" and more from groups "E"-"K," which was expected. The tests do measure intelligence and the midshipmen of the lower groups are a considerably poorer risk than those of the higher groups.
6. The groupings may be used to indicate the probability of a student graduating.
If the curves be plotted showing the percentage of the class remaining at any time, it is seen that the percentage increases in the higher groups and decreases in the lower groups. The curves are almost parallel and equidistant "D" to "H," indicating that in these groups there is an average chance to graduate. Those in groups "A" to "D" have a splendid chance, and those in "I"-"K" a very poor chance. So far as the Naval Academy is concerned, these "I," "J," "K" men are the round pegs in square holes. Few of them can be shaped to fit and the process is uneconomical, and unsatisfactory to both the government and the man. There is much to recommend denying them the chance to try, and inviting their attention to lines of endeavor more in keeping with their intellects.
STUDY OF A CLASS AFTER TWO YEARS AT THE NAVAL ACADEMY
See graph B
The curves are constructed similarly to those in graph A. Curve 1 shows the distribution of the 699 members of the class who took the psychometric test shortly after admission in 1922. Group "A" contained 128 midshipmen. 128/699 = 18% = first point on this curve. (The test was easier than the one given the Class of 1927 though similar.)
Curve 2 shows the distribution of the 467 of the original 699 who remained after two years. After two years group "A" contained 107 midshipmen. 107/467 = 23%,—that is, 23% of the class were "A" men. This gives the first point on curve 2.
Curve 3 shows the distribution of those who failed during the first two years. Group "A," originally 128, has lost ten midshipmen by physical rejection or other reasons not scholastic failures, which reduced the group to 118. Eleven have failed scholastically. Thus 11/118, or 9%, have failed. (First point on curve 3.)
To show that the above conclusions are not the result of a particular test, the case of the Class of 1926, after two years in the Naval Academy, as in graph B, may be compared with the Class of 1927 after one year, as in graph A. The test given 1926 was less difficult than that given 1927, which shows in the greater percentage of 1926 men in the higher groups, but the curves show the same tendency of few failures in the "A" and "B" groups, and many in "E" and "F."
Group | % of class in a group | % of failures in groups | ||
1927 | 1926 | 1927 after 1 year | 1926 after 2 years | |
A | 5.7 | 18.3 | 0 | 9.3 |
B | 13.1 | 21.9 | 6.7 | 19.6 |
C | 16.2 | 21.8 | 9.8 | 22.5 |
D | 17.7 | 16.6 | 12.9 | 39 |
E | 14.9 | 11 | 19.7 | 41.2 |
Mark below 75 on test | 32.5 | 10.4 | 24.4 | 44.3 |
ANALYSIS:
1. In the two upper groups approximately one fifth of the Class of 1927 (18.8% of class) there were 153 midshipmen of whom 7, or 4.6% failed. In the four lower groups (marks below 65% on test) there were 105 men of whom 42, or 40%, failed. Thus, a man whose psychometric mark is below 65 has 8.7 times the chance of failing that a man has whose mark was go or more. Think that over.
2. With the limitations of the entrance examinations it is seen that few candidates of low native intelligence can qualify for admission and the majority of failures after admission come from those who do.
3. The percentage of the class in "A" group has increased some 5% since admission, while the per cent of the class in "F" group has dropped about 3%. The curve is swinging up in groups "A," "B," "C" and down in groups "D," "E," "F"—about a pivot between "C" and "D." The next curve will show how this continues during the last two years of the course. The curve of the Class of 1926, after two years, is on its way to superimpose itself on that of the Class of 1923 at graduation.
STUDY OF THE PSYCHOLOGICAL DISTRIBUTION OF A GRADUATING CLASS AFTER FOUR YEARS AT THE NAVAL ACADEMY
See graph C
This curve represents the distribution of those members of the Class of 1923 who were able to complete the course satisfactorily and graduate. Note how the curve has pivoted-8o% of the class are in groups "A," "B" and "C," and only 20% in the three lower psychometric groups—and, there is a larger percentage of "A" men than "B," and more "B" than "C," etc. The slope of "F," perhaps, is due to that group not having an equal range of psychometric marks with the other groups, and because the percentage is based on so few men.
Experience shows that a study made at any given point in the course gives a certain answer regardless of which class is being studied. With this in mind note the slope of the 1927 admission curve, the pivoting upward for group "A" in the 1926 curve after two years, and the marked upward swing of group "A" at the end of four years as shown on the 1923 curve. This is evidence that the course, in general, eliminates those of lower intelligence and permits the graduation of those of higher intelligence, which indicates that the standard for graduation is satisfactory.
Graduates of the Naval Academy are listed according to their final multiple for the four years. The curve constructed for analyzing the relationship between intelligence and relative standing at graduation is somewhat complex. Suffice it to say that 53.5% of the first sixth of the class in relative standing came from "A" group, and that in this first sixth there were no midshipmen from groups "D," "E" or "F," again proving the point.
These tests do not prove that the "A" men will become the best officers; only time can settle that, but they do show that the "A" man gets the most out of his four years' novitiate. Having the most, and other things being equal, he ought later to give the most.
Whether the man of superior natural intellect can stand the strain and responsibilities of the service as well as the average man is something also for the future to disclose. If he can—and knowledge is supposedly power—he has an initial advantage over those of lesser endowments, which can be overcome only by the hardest kind of work. Character, judgment, leadership, health, assiduity all are external factors of importance which vary with individuals. But of two men equally well endowed with them, the one of "A" group intellect has a tremendous advantage over him of "K" group whether his career be on the sea or on the land.
Life runs on averages, not individuals. Knowing that only one or two of the midshipmen in the lower groups can expect to graduate, why should any in that classification be admitted? Suppose that fifteen or twenty of them in a class of some eight hundred do succeed in cramming enough to pass the entrance examinations, perhaps on a second or third attempt. There is an economic loss to the government in admitting, paying, subsisting, and paying officers to teach them when not more than a very few can be graduated into the service. Their more capable classmates are held back by them. Many man-hours are wasted in extra instruction and re-examinations for men who cannot absorb a two five's worth. Then, too, there must be considered the injustice to the men themselves in taking them into a career in which they cannot succeed, only, eventually, to brand them failures. Nor can it be ignored that they are blocking the appointments of better men from their districts.
True, the one man from this group who may graduate might be a Nelson or a Farragut, but that's a far cry. It is impossible to Put a quart of water into a pint measure.
In one of the recently graduated classes only one man in the lowest psychometric group succeeded in winning a diploma. He failed to be commissioned because of defective vision. Perhaps he ruined his eyes in a fine struggle with a course too difficult for his mental equipment. He is out of the service now, and the time and money spent on him might have been spent more wisely on a potential ensign—which he was not, as the psychometric test demonstrated. The quart of water went into the pint measure but only after the pressure had started a seam.
The Naval Academy regulations now require a continued study of the psychometric relationships and comparisons. In a few years' time there should be a wealth of data on hand. It is not at all an illogical prophecy that some sort of a psychometric test will then be added to the requirements for admission.
The psychological measure of an individual cannot be taken as a final index to his future, but the measure of a group can be taken accurately and the success or failure of the average of the group predicted with certainty. When the time does come for barring candidates of insufficient natural capacity it will be wise to set the minimum mark allowable no higher than is necessary to block those who will be unable to complete the course, for strain oft-times finds the shoulders of the slower thinker a sturdier resting place than those of his more intellectual brother—and service and strain are seldom separated for long.