A Further Report on the Midas Touch

A Further Report on the Midas Touch ERNEST TAVES, L. A. DALE, and GARDNER MURPHY In the April, 1943, issue of this JourNAL two of the present writers reported experimental ESP results which were characterized by various decline effects (1). These effects we attributed to a “Midas touch in reverse” on the part of the subject, the experimenter, or both. It was shown that when subjects called a series of decks of ESP cards, scores were significantly above chance for the first three decks, descending to the chance level thereafter. This effect we attributed to the Midas touch in respect of the subject: as the subject becomes less naive for a given experiment, he loses his ability to do above-chance scoring. We also presented evidence in support of the view that the Midas touch may work in respect of the experimenter; the present paper, however, does not deal with this aspect of the problem. We want to stress the fact here that the decline phenomena mentioned above were not a priort expected in the data in which they were found. On the contrary, the experiments reported on in the first paper had been designed to test an entirely different hypothesis. But the Midas touch effect became so obvious that it could not escape detection. On the other hand, decline effects were expected to appear in the data about to be presented, and the entire experiment was specifically planned to facilitate their emergence. The question is, Does the Midas touch hypothe- sis possess predictive value? Our results lead us to believe that it does. This new experiment will now be described. Experiment BA - Subjects: It was necessary, of course, to use naive sub- jects in this experiment, since it is supposed that the Midas touch makes itself apparent in the first few runs made by a given subject. Accordingly GM* asked Dr. B. F. Riess * The entire experiment was carried out under GM's direction, and was made possible by moneys available to him from the Hodgson Fund of Harvard University.

112] Journal of the American Societ y for Psychical Research of Hunter College if he could make an appeal for volun- teers in the latter’s psychology classes. Permission was kindly granted and early in March, 1943, GM addressed the classes and asked for volunteers for a “psychological experiment.” The parapsychological nature of the experi- ment was not mentioned since it was desired that the sub- jects be as naive as possible. A total of 45 volunteers was accepted and appointments were scheduled. All of the subjects, therefore, were girls of college age. None of them had previously taken part in formal ESP experimentation. Methodology and Procedure: Three methods of calling were used in the experiment. Two of these are the AE and AJ methods described in another paper appearing in this Journat (2: pp. 17ff. and also 1: pp. 71-72); the third method is similar to method AE, except that the subject does not receive knowledge of results after each individual guess. The letters “wkr” are used to signify “without knowledge of results,” so that the new method is called AEWwkr. Each subject was used for three sittings, each of which lasted from fifty minutes to an hour and a quarter, depend- ing on the individual subject’s preferred speed in calling, At each sitting the subject was required to call nine decks of standard ESP cards: three decks by one method, three decks by another method, and the final three by the third method, Each subject always did a group of three runs without the experimenter’s entering the room, At the con- clusion of a group of three runs, and before going on to the next method, the experimenter discussed the scores, and other matters of interest, with the subject. The order of methods used was rotated throughout the three sittings; that is, 15 subjects in their first sittings used the methods in this order: AE, AJ, AEwkr; 15 subjects followed the order AJ, AEwkr, AE: finally, the last 15 subjects called in the order AEwkr, AJ, AE. For each subject, in her second and third sitting, these orders were rotated so that each girl began one sitting with each one of the three methods. We planned this rotation of tasks so that a direct comparison of the methods themselves could be idas 1 {113 A Further Report on the Midas Touch : oe ere used, moreover, In order to gi made. Three methods w rity to appear. A subject i rtur et = cas ne Ki ane still be considered naive for who has called, BP > rding to the Midas touch pons oe SS cate frst run of all Seale ed ee ead ede higiest average. We also exe Shem ao nes Sms of each method would be eee Sag pes gr pa run of these would be higher eer : wey a d. It would follow from this, therefore, ie secon add better at her first sitting than a_i : “es ne Mea it was for this reason that we ha sc gt : oe die experimental sessions. These sessions Saeat : ximately weekly intervals. ee TD ces xperimenter throughout the se ton oa “ sitions of experimentation (cites paki a ed scinake material, Be actly the same el mechanical recording, etc. ) i= - ey 1 ( SCALE OF REWARDS conspicuously posted on < and in most cases LD cal t began her first sitting. e office of the A.S.P.R., This scale of rewards was subjects’ panel of the a= puns attention to it before the sub) tbs d at th - tt he subject appeare ree runs, the : he the seenpledicn of = group alee eek A and ¢ . rith her on ; , Neri conversed Ww! : j mal atmos Sorettort was made to create a friendly, infort real effo aoe

phere; cigarettes a 4 nd candy w was sl , qe ” y ere offered Ate Ab ae . for “high” scores. It is iaiposetide pbs able” attics; ee extent this attempt to create: Favors gene ioe Was successful. Most of the em pressed iperns pips enjoyed the experiment and cl titans Sennen ~ participate without remuneratio se asks: abene oc $s. Nothing was said, except in a ve r in for the mos . ance expectation,” and the subj ig ce Rest big part naive on this point iit ies ts: In the entire AS cards were cz re experiment 1215 : vidual nik As by 45 subjects—a total of sare ed of SOL tts A plus deviation of 51, giving a mean sean sionihea vas obtained. This over-all mean score ~ cance. result is without he avera € (that is, une genre of the first run of the first sitti 5.42. Since th © subjects were at their most nai itting value is not wc. were, Of course, only 45 Sarat one wee line ae ‘ In itself statistically significant eae = Tn Tal = bypothesis stated a priori nt, but it is in first coe bos Presented the deviations obtained { lee d second three runs called or the See 1: pp. 72 ff.), , by each method Table I Deviatio ations of first three : runs compared wi F vith deviations s of second three runs (by Expt. BA. (by each method ) called by naive subjects 135 135 135 135 135 135 It will be seen f the ay. 3 rom the above tabl : (de ——— score of the first run for sath oe he predicted, (+6 on = +40) than the average 5 panied ia ) and third (—6) runs score for the second A Further Report on the Midas Touch {115 None of the three methods used yielded a significant mean score; the obtained means for these three methods (N for each method is 405) are shown below: 4.975 5.044 The deviations and mean scores obtained on the first, second, and third sittings are as follows: First Sitting: sss 5.09 (+40) Second Sitting ..csscsessesenenssrrensens 5.005 (+ 2) 5.022 (+ 9%) Third Sitting ....sscsssesererersseereeeeees In Table II we present the complete Midas touch mate- rial to date, Included in this table are the data shown in Table I above, the original Midas touch material, and all the data collected under the original conditions (that is, exclusive of Experiment BA) subsequent to the publica- tion of the first Midas paper. Table II Deviations of first three runs compared with deviations of second three runs called by naive subjects, all data (August, 1941, to April, 1943). Thus, only eight times in ten thousand similar experi- ments (each experiment extending over a period of twenty months) would such a result be expected on a chance basis; this result is, therefore, highly significant by standard statistical criteria. Certain other analyses were made on the data. The following may be mentioned : (a) Salience (in respect of hits on the first card of each deck) ; preponderance of hits

ad a in the first half of each deck ; three, four, fi iree, » live, etc., successive hits): pica sania : ive hits); and abno be scat (preponderance of extreme scores) as ati sis quare. Null results were obtai ‘i aves »btained from these An isol : iki ue aE ae scoring must, however, be ,» Miss i : Foca . L., calling her firs some argh deck of her first sitting), obtain a nei wlll hate is cards of the deck being co HY 3 sed. 2 unbroken s t 5 hi ; ee sequence of 15 hits ae - ng iS - ao Abie be expected to occur by cel aan! wt vere runs. It will be recalled that all renee anged in accordance with random Discussion of results: It is clear thz og three possesses significant adv rr ie plac Pi east under our conditions of the mean score obtained b is without omg: antage over any other— xperimentation. Further ’ statistical <3 jects under all ws statis sienific: : all conditio tistical significance. It will be noted however! that the scoring level on Sitting ¢ ; s One is highe gee = ake as eae, ae Rei hie ean rst sitti : Sa — were ae Maer _ We believe that the most j ti i gr Rate mportant contributi hed Gar in the fact that a definite type of tale soe ae fneenuee mae a tendency to aa aiveté, with a s 1 ee a anapeat scoring. In other words, we Price tis eee we have put, the salt on the tail of that ‘ nec e Ih se experiment in ESP. 4 ilies bay — an mor obtained by Miss L. L Ngan Gs oe ds against correctly calling 15 cards aad aa _ - omical), it may be of interest further phonies = ‘A apr attending this run, as well as the antec nie rk o the subject. LD was experimente Gat fee i en but on the occasion of Miss L. T 8 ce tn cilia ae that ET was present in the roo eco operating the machine and GM i room at the A.S.P.R. office. When the first five average score ve chance (5.42). as hypothesis. grouping of hits (runs of A Further Report on the Midas Touch {117 or six guesses registered as correct, LD asked ET closely to observe the rest of the run, This he did. At the termina- tion of the run under discussion GM’s advice was sought on how to proceed. He suggested that there should be no change from the usual procedure as described above; that, in other words, Miss L. L. should continue her task of calling the remaining two AEwkr decks without interrup- tion from us. After these decks had been called (yielding scores of 4and 6) LD entered the room and, without telling the subject of her unusual score, engaged her in conversa- tion. Miss L. L.’s introspections were as follows: In the “17” run she had started out by looking around the room and certain objects there, which she associated with the five symbols, had successively caught her eye. For instance, when the coils of the radiator seemed to “attract” her she chose the wave; dry cell batteries on a near-by shelf indicated the circle; the corner of the table indicated the star; the square was symbolized by a chair-bottom, and “something she saw in the design of the rug” meant the cross. (Actually there is nothing in the design of the rug, which is of an Oriental type, that even remotely resembles the form of a cross.) Then, as the run progressed, this procedure began to seem too stereotyped, too “automatic,” and another approach to the task took its place. She felt that she should call, say, a circle, and then she looked around the room until she found an object the outline of which would “justify” her impression. She further stated that after a while (whether towards the end of the “17” run or at the beginning of the next run could not be exactly ascer- tained) she “got bored with looking around the room and just tried to imagine what symbol Mrs. Dale was looking at, but thought it silly.” She was not told at the end of this sitting, as were all the other subjects, what her exact scores were, nor was she paid the reward for her score of 17 until after the last sitting. Her total deviation for Sitting One was +14; Sitting Two, +8; Sitting Three, —4,. When Experiment BA was completed, Miss L. L. was kind enough to come on two occasions to take part in further experi- mentation. The results obtained were of a chance nature.

118] Journal of the American Society for Psychical Resear Summary: An experiment was designed to facilitate appearance of decline phenomena (Midas touch effects 45 subjects called, during three sittings each, a total ) 1215 decks of ESP cards, using three different methods o calling. Suggestive, but not statistically significant, declin effects were obtained, these effects being similar to th ¢ previously noted. When the results of the present experi ment are pooled with the results published in the April issue of this JouRNAL, the P-value, based upon the positive deviation of the first three decks called by naive subjects, is .0008. The Midas trend in the present data is less pro- nounced than in the previously reported material, but the total result remains statistically significant. REFERENCES 1. Taves, Ernest, and Dale, L. A., The Midas Touch in Psychical Research, — Journal A.S.P.R., Vol. XXXVII, April, 1943, pp. 57-83. 2. Murphy, Gardner, and Taves, Ernest, Current Plans for Investigations in ig Research, Journal A.S.P.R., Vol. XXXVI, January, 1942, BP