A student runs a logistic regression in SPSS predicting whether someone buys a meal plan (1 = yes, 0 = no) from two predictors: Income and Housing (1 = on-campus, 0 = off-campus). The Omnibus Tests of Model Coefficients shows χ²(2) = 12.6, p = .002. The Hosmer–Lemeshow test shows χ²(8) = 4.1, p = .85. Which statement is most accurate?
The overall model is statistically significant and the Hosmer–Lemeshow result suggests the model fits the data reasonably well.
The model is not significant because Hosmer–Lemeshow is non-significant.
The model is significant, but the fit is poor because Hosmer–Lemeshow is non-significant.
Neither test is useful in logistic regression; you should only look at R².