Simple Linear Regression (SLR) Model and Assumptions SLR is a statistical method used to model the relationship between two numerical variables, x an...
SLR is a statistical method used to model the relationship between two numerical variables, x and y. It assumes a linear relationship between these two variables, meaning that the y variable changes proportionally with the x variable.
Assumptions of Linear Regression:
Linearity: There is a linear relationship between x and y. This means that the points on the scatter plot of these two variables will form a straight line.
Homoscedasticity: The variance of the residuals (the vertical distance between points on the scatter plot) is constant across the entire range of values for x. This means that the points are scattered around the line in a consistent pattern.
Normality: The residuals are normally distributed, meaning that they are randomly scattered around the line of best fit. This assumption is important for ensuring that the least-squares regression line is unbiased.
Independence: The observations are independent of each other. This means that the value of y for one observation does not affect the value of y for any other observation.
Implications of Non-Compliance:
If any of these assumptions are not met, the results of the linear regression model may be misleading. For example, if the residuals are not normally distributed, the confidence intervals may not be accurate, and the coefficient estimates may be biased.
Examples:
Predicting the sales of a product based on its price.
Analyzing the relationship between the number of patients and the number of treatments they receive.
Evaluating the effectiveness of a new drug by comparing its effectiveness to a placebo.
Additional Notes:
The ordinary least squares (OLS) method is used to find the line of best fit that minimizes the sum of the squared errors between the observed data points and the predicted values from the linear regression model.
The R-squared value is a measure of how well the linear regression model fits the data. A high R-squared value indicates that the model fits the data well, while a low R-squared value indicates that the model does not fit the data well