What guides a researcher when deciding between using either linear, logistic, or Cox proportional hazards regression as an analysis tool

The outcome variable type (continuous, binary, or time -to- event)

Explanation:

The outcome variable type is the type of variables involved in a research. The researcher would put the type of variable involved in his research into consideration in deciding what regression model to apply in his research. For example, if the type of variable in his research are continuous variables(continuous variables are variables that may be any value within a range and may be infinite), he would use linear regression

i'm not sure i quite understand, could you provide more information?