Comparing Guessing Games with Homogeneous and Heterogeneous Players:
Experimental Results and a CH Explanation
Eugen Kovac, Andreas Ortmann, and Martin Vojtek
(March 2007)
Results of Mann-Whitney tests
	our HOM > HET		q=1/3: GKS > our		q=2/3: GKS > our		q=1/2: GKS <> our
Round	z-statistics	p-value	z-statistics	p-value	z-statistics	p-value	z-statistics	p-value
1	0.06	0.476	1.21	0.113	2.52	0.006	0.90	0.368
2	1.42	0.078	1.49	0.068	4.16	0.000	-0.33	0.741
3	2.29	0.011	3.07	0.001	3.92	0.000	-1.31	0.190
4	2.38	0.009	3.54	0.000	4.03	0.000	-0.58	0.565
5	2.74	0.003	3.61	0.000	4.25	0.000	-0.99	0.320
Test		Data			Alternative hypothesis
our HOM > HET		Our data HOM vs. pooled HET			Distribution in HET is closer to the equilibrium than distribution in HOM (1-sided).
q=1/3: GKS > our		Our q=1/3 vs. GKS q=1/3			For q=1/3: Our distribution is closer to the equilibrium than GKS's distribution (1-sided).
q=2/3: GKS > our		Our q=2/3 vs. GKS q=2/3			For q=2/3: Our distribution is closer to the equilibrium than GKS's distribution (1-sided).
q=1/2: GKS <> our		Our q=1/2 vs. GKS q=1/2			For q=1/2: Our distribution differs from GKS's distribution (2-sided).