TABLE I.

Statistics for proportion of correct localizations in Fig. 3. N = number of presented sounds (1–8); combin(12,N) = combination of 12 sources taken N at a time (number of total possible source combinations); n = number of reported sounds (1–4) with n:Fig. 2 being n from Fig. 2; P_N(Hits):Fig. 3 = the proportion of hits from Fig. 3; n_Hits = number of correct localization responses measured out of the n responses, P_n(Hits) = proportion of Hits based on n reported responses; TOT_c = combination of N things taken n at a time (i.e., number of total situations in which all n responses out of N possibilities were hits); and P(Cor):Guess = guessing proportion, i.e., getting all n responses correct = TOT_c/N by chance. Measures of localization performance using the procedure of experiment I are only possible if N is considerably less than 12. Consequently, the current experiment uses 8 as the upper limit.

N	combin(12,N)	n:Fig. 2	PN(Hits):Fig. 3	nHits	Pn(Hits)	TOTc = combin(N,n)	P(Cor):Guess
1	12	1	0.93	0.93	0.93	1	0.0833
2	66	2	0.84	1.68	0.84	1	0.0152
3	220	3	0.70	2.10	0.70	1	0.0045
4	495	3	0.55	2.20	0.73	4	0.0081
5	792	4	0.47	2.35	0.59	5	0.0063
6	924	4	0.45	2.70	0.68	15	0.0162
7	792	4	0.42	2.94	0.74	35	0.0442
8	495	4	0.40	3.20	0.80	70	0.1414