2. The Methods of the Psychology of Testimony. By "Testimony" or "Report" (Aussage) we mean the verbal expressionof a recollection; and by "Recollection" (Erinnerung) a complex of memorial ideas which has reference to a definite objective constellation of facts (Talbestand) in the past. The chief purpose of the study of testimony is the determination of its accuracy, i. e., the degree of its agreement with the actual constellation of facts, and of the conditions upon which this accuracy depends. All experimental methods must therefore permit a comparison of the testimony with the facts to which it relates. Pictures to be carefully examined and then later described from memory furnish the most convenient material and have been most frequently used so far; but since in practical life one has very often to do with reports regarding events, event-experiments (Vorgangsexperimente) have also been arranged. Besides these, reports of narratives, of extents of time and space and many other matters have also been worked with.
The comparison of the facts and the report can be made only when the latter has been analyzed into its single statements (Einzelangaben) and the percentage of right, wrong and indeterminate statements has been calculated; and, because of the varied character of the statements, it is necessary to calculate separately the proportions for particular categories (e. g., with reference to matters of color or of space relations) as well as to make a general calculation.
Next the conditions affecting the accuracy of the reports must be varied experimentally. The most important difference is here between the "narrative" (Bericht) and the "interrogatory" (Verhör, testimony given in response to questions, Whipple's "deposition"). In the case of the "interrogatory" form there are also different degrees of suggestion to be considered. Then the interval between the original observation and the report is to be varied; and finally the dependence of the accuracy of the report on the education, age and sex of the person under investigation is to be determined.
3. Numerical Results. The first experiments were made with pictures on mature students, both ladies and gentlemen. The reports were in the "narrative" form, without interrogation. The errors in the report, when made immediately after the observation, amounted to 5%, some weeks later to 10%. To distinguish those portions of the report with reference to which the subjects were very certain—so certain that they would be willing to take oath upon them—such portions were underscored. These portions showed a lessened tendency to error, but were not free from it.
A further experiment, with the picture of the living room
