The laboratory confirmation of growth hormone (GH) deficiency (GHD) has been extensively studied. Multiple stimuli induce GH release, but insulin-induced hypoglycemia usually is considered the ‘gold standard’. Seventy-five to 90% of normal children have significant increments of hGH to any single test. Complete and partial syndromes of GHD have been defined, but some patients with a clinical appearance of GHD release hGH during provocative testing. Discordant results on varied tests may occur in the same child. Sequential and simultaneous tests have been attempted with diverse time patterns; testing sequence may significantly affect data interpretation. Persistent problems with GH provocative tests remain: normal data not strictly defined throughout childhood, multiple tests with discordant results, and substantial discrepancies of immunopotency estimates with different radioimmunoassays. Some children with ‘normal’ hGH increments during provocative tests, despite clinical GHD, may require short-term treatment with hGH to finally establish the diagnosis.