Final answer:
The incorrect statement refers to a misinterpretation that a significant positive coefficient in logistic regression causes a direct increase in the probability of success by the coefficient amount with a one unit increase in the variable.
Step-by-step explanation:
The statement that is NOT correct about interpreting the coefficients of logistic regression is: "If a variable has a significant positive coefficient, we can say that the probability of success increases by the amount of the coefficient as the variable increases by 1 unit." This statement is incorrect because a significant positive coefficient in logistic regression indicates that the log odds ratio increases, but this does not translate to a direct increase in probability by the amount of the coefficient. The change in probability is not linear and depends on the logistic function.
In logistic regression, if a variable has a significant positive coefficient, it is indeed correct that the log odds ratio increases by the coefficient amount when the variable increases by 1 unit. Similarly, an increasing log odds ratio means the odds ratio also increases, which in turn suggests an increased probability of the event occurring. Conversely, a decreasing log odds ratio indicates a decreasing odds ratio and a lowered probability of the event.