What are the problems with predicting a dichotomous dependent variable using OLS?
The reasons why OLS cannot be used for a dichotomous dependent variable is because it may lead to predicted probabilities of the DV that lie outside the limited values of 0.0 and 1.0. Additionally, the relationship between the IV and DV may be nonlinear. Also, the error terms do not have a constant variance.
Is OLS estimation applicable for estimating the dichotomous variable as the dependent variable?
Introduction to both Logistic Regression and Ordinary Least Squares Regression (aka Linear Regression): It is similar to a linear regression model but is suited to models where the dependent variable is dichotomous. It’s coefficients can be used to estimate odd ratios for each of the independent variables in the model.
When to use OLS regression in a scale?
OLS regression is only appropriate when the assumptions of OLS regression are met (normally distributed errors, etc). As mentioned, if you have scales built from the sum of several 5-point questions, you might be able to get away with using OLS.
Do we need normal distribution of dependent variable when working with?
1) It is not the distribution of the variable that needs to be normal (or, better: Gaussian). If a distribution matters at all (e.g. in the Newman-Pearson framework of hyposesis testing) then it is the distribution of the residuals.
Can you use OLS to analyze 5-point Likert scale?
For 5-point scales, this is probably not true. (E.g. “agree” might be closer to “neutral” for most people than it is to “strongly agree”). In practice, if Likert scale data will pretty well meet OLS assumptions, people will often use it as the best option.
Can a probit model be used as an odds ratio?
Probit coefficients are essentially uninterpretable – given a probit model I would report average marginal effects for this very reason. Of course most people improperly interpret odds ratios as probabilities which is a big no-no.