Figure 1

Mean differences across a pool of N groups causes spurious correlations. r _xy was obtained from combining N groups of simulated data; simulation parameters: μ_xy,1 = (α,0), μ_{x y,2} = (0,α), −10 ≤ α ≤ 10, and μ_xy,i = (2,2) for i ≥ 3; n _i  = 10, λ _i  = λ, 0.01 ≤ λ ≤ 0.1, ∀ i = 1,N, and 10 ≤ N ≤ 100.