Eq. (12.2.4) is based on the assumption that Z has approximately a normal distribution. Occasionally, the normal approximation is not good enough. In such cases, one can let
v = σ2/∊2(1− γ). (12.2.6)
To be precise, let Z be the average of v independent random variables with meanμand variance σ2. Prove that if v is at least as large as the number in Eq. (12.2.6), then Pr(|Z − μ| ≤ c) ≥ γ .