Why does it make little sense to look for heterogeneity of effect sizes in this example? Bauer and Döpfmer (1999) examined the efficacy of lithium as an augmentation to conventional antidepressants. They found nine placebo-controlled testing lithium for those cases that did not respond to other antidepressants. The data, as estimated from their paper, follow:

Cl -1.10 to 050 0.65 to 1.03 002 to 1.02 0.05 to 2.77 0.09 to 1.85 -0.34 to 1.36 -1.0I to 2.03 -0.02 to 2.68 -002 to 1.36 Stein 34 -0.30 0.19 0.50 1.36 0.97 0.5 051 082 0.86 0.51 1.39 0.99 087 1.55 1.38 0.70 Katona 6l 27 24 17 Baumann Kantor Heninger Joffe 15 0.76