Linear Regression Project Part II Statistical Analysis 1. Run r...

Linear Regression Project
Part II Statistical Analysis

1. Run relevant descriptive statistics and graphs using Minitab. (no more than 30% of the paper)
• Create a histogram for each of the quantitative variables.
• Create a boxplot for each of the quantitative variables.
• Calculate the descriptive statistics to study the center (Mean and Median), dispersion (Standard Deviation and Range), distribution (Skew and Kurtosis), and position (re: outliers - if any and by what criteria) for each of the quantitative variables. Describe what these values tell you about each of the variables.
• Decide not only whether the data are symmetric or skewed, but whether the data are sufficiently symmetric to make the assumption of normality. Give your reasoning as well as stating your decision about the shape of the data.

2. Create and analyze a scatter plot.
• Use Minitab to create a scatter plot with the regression line.
• Use the scatter plot to decide if there is a significant linear relationship between the variables.
• Categorize the correlation as strong/weak/insignificant.
• If a strong or weak correlation exists categorize it as positive or negative.
• Analyze points in the scatter plot that appear not to follow the trend of the regression line. Try to discern why these points do not follow the trend and whether the points should be kept in the analysis.

3. Analyze numeric measures of correlation
• Find the correlation coefficient, r and explain its relevance.
• Explain the relevance and practical meaning of R-squared.
• Examine whether evidence of a relationship between two variables exists using the p-value.
• Explain the relevance of the slope of the regression line.

4. Predictions
• Find the linear regression equation.
• Elaborate on the reliability of predictions. Should the equation be used for predictive purposes?
• If appropriate, select a few values of the predictor variable and make the prediction using the regression equation. If not, explain what approach you would take to make predictions.