In a regression with n = 50 observations and k = 3 predictors, the criterion for high leverage is: A)h_i ≥ .33 B)h_i ≥ .25 C)h_i ≥ .21 D)h_i ≥ .16

Question

Quizplus · Accepted Answer

In a regression with n observations and k predictors, the leverage value for an observation i can be calculated as hi = (S1U1B1i)^2/(SSE * (1 - Hii)). The criterion for high leverage is usually set at 2k/n, which in this case is 2 * 3/50 = 0.12. However, the actual cutoff value can vary slightly depending on the specific context and goals of the analysis. Option D corresponds to a limit that is slightly more stringent than the standard 0.12 cutoff, and may be appropriate in some situations where the potential impact of outliers or influential observations is particularly concerning.

In a Regression with N = 50 Observations and K