Contents
Does OLS require normality?
OLS does not require that the error term follows a normal distribution to produce unbiased estimates with the minimum variance. However, satisfying this assumption allows you to perform statistical hypothesis testing and generate reliable confidence intervals and prediction intervals.
Does the dependent variable need to be normally distributed?
So is the normality assumption necessary to be held for independent and dependent variables? The answer is no! The variable that is supposed to be normally distributed is just the prediction error.
What is a continuous dependent variable?
If a variable can take on any value between its minimum value and its maximum value, it is called a continuous variable; otherwise, it is called a discrete variable. Some examples will clarify the difference between discrete and continuous variables.
When to use OLS with a binary variable?
If you’re trying to optimize squared error, then even with this binary variable: OLS is exactly what you need. If, on the other hand, you’re interested in minimizing log loss or hoping to get a calibrated probability – then you better use logistic regression. OLS with a threshold is a real classifier.
How are average answers used in OLS regression?
The author then takes the average of the answers to the three questions for each individual and then uses this individual average as dependent variable in an OLS regression with binary and continuous explanatory variables. This does not make sense to me from an interpretation point of view.
How is the OLS estimator used in OLS analysis?
Interpretive issues for the OLS estimator notwithstanding, the real issue here is in the treatment of an ordinal variable as if it were a variable on the ratio scale. By using standard linear regression analysis, the researchers are essentially treating the ordinal response as if it were a continuous quantity.
Can you use the OLS method with a threshold?
OLS with a threshold is a real classifier. It may not explicitly optimize a traditional classification metric, but you can definitely *evaluate* it with a traditional classification metric; and if you’re happy with the results, then … Yep. Most people will say no, but it’s not so. It all depends on what metric you’re interested in optimizing.