Formulae and Tables: ∑π₯ π₯Μ = π π ∑ π₯ 2 −(∑ π₯) , π =√ π‘= π1 +π2 −2 Μ Μ Μ Μ Μ Μ Μ Μ π1 −π 2 −(π1 −π2 ) 2 2 π1 π2 Paired t test: t = πΜ −(π1 −π2 ) , π π ⁄ π √ Grand mean π₯Μ = , π‘= 1 1 + π1 π2 2 π 2 π 2 ( 1+ 2) π1 π2 π 2 π 2 ( 1⁄π )2 ( 2⁄π )2 1 + 2 π1 −1 π2−1 π π √ , π·πΉ ≅ π π √ 1+ 2 π₯Μ −π0 , π‘ π ⁄ π √ Μ Μ Μ Μ Μ Μ Μ Μ π1 −π 2 −(π1 −π2 ) π§= π(π−1) (π1 −1)π 12 +(π2 −1)π 22 π π = √ 2 = π₯Μ −π0 π ⁄ π √ , DF : n-1 , DF: π1 + π2 − 2 , ∑π πΜ = π π 2 π ∑ ππ2 − (∑ ππ ) π π = √ and π1 π₯Μ 1 + π2 π₯Μ 2 +β―+ππ π₯Μ π π(π−1) , π = π1 + π2 + β― + π1 + π2 +β―+ππ ππππ = π1 (π₯Μ 1 − π₯Μ )2 + π2 (π₯Μ 2 − π₯Μ )2 + β― + ππ (π₯Μ π − π₯Μ )2 πππΈ = (π1 − 1)π 12 + (π2 − 1)π 22 + β― + (ππ − 1)π π2 ππππ πππΈ ππππ ππππ = π‘= π−1 πππΈ = πΉ= π−π πππΈ , DF = n-1 ππ ~πΉ(π − 1, π − π) π½Μ − π½ ~ π − (π + 1) ππΈ(π½Μ ) [π½Μ : Estimated regression coefficient π½: Population coefficient, ππΈ(π½Μ ): Standard error of estimated coefficient] Logistic Regression: The marginal effect of quantitative π₯π variable: ππ ππ₯π = π½π π (1 − π) Alternatively, ME of π₯π at value x is: ME = P(Y = 1 | Xk= x+1) − P(Y = 1 | Xk = x) Marginal effect of a binary dummy variable D: ME = P(Y = 1 | D=1, X) −P(Y = 1 | D=0, X) Chi Square Tests: π 2 = ∑ (ππ −πΈπ )2 πΈπ [O: observed freq, E: expected freq] [DF: (r-1)(c-1) for test of independence and k-1 for goodness of fit test] π ππ€ πππ‘ππ × πΆπππ’ππ πππ‘ππ πΈπ₯ππππ‘ππ πππππ’ππππ¦ = , Expected frequency = np πΊππππ πππ‘ππ