30 



MASS. EXPERIMENT STATION BULLETIN 310 



Accuracy of Medical Examhiations 



The accuracy of data in which the personal equation enters so largely as in 

 medical examinations is always open to question. An attempt was made, 

 therefore, to obtain some idea of the dependability of the records by observing 

 the accuracy with which the physician would duplicate a number of examina- 

 tions. Whether fortunately or unfortunately, a week or ten days was allowed 

 to elapse between examinations. As a result the influence which colds and acute 

 infections may have on the records is show n more clearly than the relative accuracy 

 of the physician. 



The conditions as given in the twenty-six duplicate examinations were corre- 

 lated with those found in the original ones. With so few cases the coefficients 

 obtained are significant only in indicating that in no instance did a high degree 

 of agreement exist (Table 8). Further study of the two sets of records showed 

 that practically every child had a cold at the time one examination was given 

 or between the two or shortly before the first. Less correlation was indicated 

 in conditions more apt to be influenced by acute illnesses as lungs and conjunc- 

 tivae, or in those involving a change in color, which suggested that the lack of 



VP 



p- 



Orlglnal Essmlnatlon 

 P+ F- P P+ (}- 



0+ 



P+ 



P 



g P- 



<=> P+ 



Chart 5. Correlation of tlic Condition of tlie Tonsils Found in 27 Duplicate Examinations 

 with that in the Original. 

 F In most instances lack of agreement could be explained by colds, indicating that the physician's 

 work was guile accurate. Crosses (-f ) indicate that the difference may be explained by a cold. 



