One uses a sheet with the stimulus words printed on it and with space opposite each stimulus word for the reaction.2 In a room reasonably free from distracting influences the subject is seated at a distance from the experimenter so as to be unable to see either the printed stimulus words or the reactions as they are recorded. He is instructed to respond to each stimulus word with the first word that comes to his mind other than the stimulus word itself or a mere different grammatical form of it, to respond with one word only and not with a compound word or a sentence or phrase. A few stimulus words not on the list may be given for preliminary practice, the reactions not being recorded; and when it appears that the subject understands the instructions the test may be begun. Should the subject in the course of the test give an unacceptable reaction, it is not put down, but the pertinent instruction is repeated, the test continued, and at the end all the stimulus words thus improperly reacted to and therefore remaining without a recorded reaction are given over again.

1 G. H. Kent and A. J. Rosanoff. A Study of Association in Insanity. Amer. Journ. of Insanity, July and Oct., 1910. - R. S. Woodworth and F. L. Wells. Association Tests. Psychol. Monogr., No. 57, 1911. - F. L. Wells. The Question of Association Types. Psychol. Review, July, 1912. - F. C. Eastman and A. J. Rosanoff. Association in Feeble-Minded and Delinquent Children. Amer. Journ. of Ins., July, 1912. - Isabel R. Rosanoff and A. J. Rosanoff. A Study of Association in Children. Psychol. Review, Jan., 1913. - E. K. Strong, Jr. A Comparison between Experimental Data and Clinical Results in Manic-Depressive Insanity. Amer. Journ. of Psychol., Jan., 1913. - Margaret Otis. A Study of Association in Defectives. Journ. of Educ. Psychol., May, 1915. - Ida Mitchell, Isabel R. Rosanoff, and A. J. Rosanoff. A Study of Association in Negro Children. Psychol. Review, Sept., 1919.

2 Printed forms for this test may be purchased in packages of 25 from The Morningside Press, 417 West 118th Street, New York.

In cases in which it is desired to use the association test for the purpose of detecting pathogenic subconscious ideas or complexes that may be suspected to exist, the examiner's familiarity with the case will suggest to him special stimulus words adapted to the particular case; these stimulus words may be given together with those regularly employed, being introduced, say, after every fifth or every tenth one. In such cases it is also advisable to record in each instance the reaction time in fifths of a second, taken by means of a stop watch; subconscious ideas or complexes are said to be indicated either by abnormal types of reaction or by instances of reaction time much above the average for the individual.

Classification Of Reactions

This test has been applied to one thousand normal subjects, and all reactions thus obtained arranged in frequency tables for all the stimulus words. These frequency tables are reprinted below.1

In the examination of a test record obtained by this method the first step is to compare it with the frequency tables and thereby distinguish the common reactions, which are to be found in the tables and which are for the most part normal, from individual reactions, which are not to be found in the tables and which include the great majority of those that are of pathological significance.

For the sake of accuracy, any reaction word which is not found in the table in its identical form, but which is a grammatical variant of a word found there, is classed as doubtful.

From amongst both common and individual reactions a fairly definite group can be separated out, the non-specific reactions. In this group are placed words which are so widely applicable as to serve as more or less appropriate reactions to almost any of the stimulus words. In the standardized procedure any of the following words, occurring as a response to any stimulus word, is classed as non-specific: article, articles, bad, beautiful, beauty, fine, good, goodness, great, happiness, happy, large, man, necessary, necessity, nice, object (noun), people, person, pleasant, pleasantness, pleasing, pleasure, pretty, small, thinking, thought, thoughts, unnecessary, unpleasant, use, used, useful, usefulness, useless, useless-ness, uses, using, woman, work.

1 Similar tables have been compiled for children: H. Woodrow and F. Lowell. Children's Association Frequency Tables. Psychological Monographs, No. 97, Princeton, N. J., 1916.

Inasmuch as the frequency tables do not exhaust all normal possibilities of word reaction, a certain number of reactions which are essentially normal are to be found among individual reactions. In order to separate these from the pathological reactions an appendix to the frequency tables has been compiled, consisting mainly of specific definitions of groups of words which, occurring as individual reactions, are to be counted as normal. (See p. 603.)

Derivatives Of Stimulus Words

Under this heading is classed any reaction which is a grammatical variant or derivative of the stimulus word: eating - eatables, short - shortness, sweet - sweetened.

Sound Reactions

In the standardized procedure a reaction is placed under this heading when 50 per cent of the sounds of the shorter word of the pair are identical with sounds of the longer word and are ranged in the same order.

Among sound reactions are occasionally found neologisms; for these a separate heading is provided.

Word Complements

Here is included any reaction which, added to the stimulus word, forms a word, a proper name, or a compound word.

Particles Of Speech

Under this heading are included articles, numerals, pronouns, auxiliary verbs, adverbs of time, place and degree, conjunctions, prepositions, and interjections.

The phenomenon of perseveration occurs in cases in which there is abnormal lack of mobility of attention. The names of the different types of reactions attributable to perseveration are given below in the classification table and are sufficiently descriptive; we shall refer here only to those which require further definition.