Table 2.
Definition | Symbol | Estimate ψ̂ = D̂/C | |
---|---|---|---|
Covariance | D | D̂= |π11π22 − π12π21| | |
Association | ρ | D̂/Q(1 − R) | |
Correlation | r |
D̂/ |
|
Regression | b | D̂/R(1 − R) | |
Frequency difference | f | D̂/Q(1 − Q) | |
Delta | δ | D̂/Q(1 − R − Q + RQ + D) | |
Yule | y | D̂/[2Q(1 − Q)R(1 − R) + D(1 − 2Q)(1 − 2R) + 2D2] |
D̂ is an estimate with expected value D and information KD.
KD = n/[Q(1 − Q)R(1 − R) + D(1 − 2Q)(1 − 2R) − D2] (ref. 34, equation 3.8).