Miscellanea 
306 
The next step was to find the correlation of rank in the total examination with marks in 
arithmetic. There resulted 
p„,= --8805* 
whence we deduce r"'M='901. 
All four correlations ?Vi >''bdi '>'"bd f-id r"',„i differ by less than -01, or less than half the probable 
error of r(„j. The differences therefore have no significance, or the values are interchangeable. 
Equally clearly r"^^ and r'",,^ are as stated above substantially equal which involves the equality 
of pi,c and pad to the same degree as stated above. 
Lastly I correlated total marks with rank in total examination and found 
p„6= --9739. 
Multiplying this by 1-023,278 we have >•;,{,= -997, 
which is as near the desired result of perfect correlation as we can expect to get on the assumption 
of a normal distribution of the variate. 
As far therefore as this isolated example is concerned the method appears to be satisfactory. 
The results reached in both our illustrations suggest that the correlation of rank and variate 
for the same character may be of considerable value for control purposes in dealing with exami- 
nation results. 
IV. A Study of Women Delinquents in New York State by Mabel R, 
Fernald, Mary H. S. Hayes and Almena Dawley with a Statistical 
Chapter by Beardsley Ruml and a preface by Katherine Bement 
Davis. Published by The Century Co., New York City, 1920. 
By ETHEL M. ELDERTON. 
At the present time there seems to be an increasing demand for statistics in connection with 
social problems and a desire to collect data sometimes without a very clear idea as to how the 
material when collected can be used but with the pious hope that it will be viseful in some way. 
The interesting memoir before us if it had done nothing else would be extremely valuable as 
showing what is required before any social problem can be adequately studied from the statistical 
aspect, how investigators are hampered on all sides by a lack of comparative data and how after 
years of work and a most careful compilation of facts the main problem may l»e still unsolved. 
The authors of this book on women delinquents in New York State fully realize the necessity of 
such a study and the difficulties in the way. They start by pointing out that though there is 
abundant literature on the subject of the criminal population much of it is useless becatise it 
rests "at its best, upon the most casual and superficial observations, and at its worst upon what 
the writer thinks he would find on observation." The writers point out that when the object of 
imprisonment was simply the two-fold one of punishment for the crime and protection of the 
state from further crimes an exact study of the causes of criminal acts was of less imi^ortance 
than it is to-day when the object of the imprisonment of criminals is of a more humanitarian 
nature, namely not only to protect the community but to attempt to readjust the criminal to 
society. As far as we know this is the earliest extensive study of women delinquents using 
modern statistical methods. The first eflTorts of the investigators were directed towards obtaining 
information as to the distinguishing characteristics of women convicted in New York State, of 
their mental capacity and of the main facts of their personal and environmental histories. 
Physical and medical facts had to be disregarded almost entirely since adequate medical data for 
most of the groups of women studied could not be obtained. The first comparison which the 
authors wished to make was between the women delinquents and women iu general : Are the 
delinquent women a special selection out of the female population or are they a random sample 
of the total differentiated only by their criminal career ? Do they differ significantly in their 
mentality, in their environment, in their education ? Here the difficulty is, that as in so much 
literature dealing with medical and social questions only the special group is studied and not the 
general population, and though the United States census gives some information it is too general 
to be of much use, and the authors have to fall back on data collected for special groups of 
women other than criminals who may not be representative of women as a whole. A second 
comparison which might be made is with men criminals, but the study of men is almost entirely 
confined to men convicted of relatively serious offences and does not deal with those committed 
to workhouses, county penitentiaries and reformatories. Dr Goring's study of English criminals 
would cover practically none of this group ; the criminals he considered were felons while the 
* See footnote on p. 302 as to change of sign between p;,^ or and r"^^ or r'"(,j. 
