Applied Statistical Methods - Solution 9

Problem 1: Repeated Observations in Predictors

Use the dataset on blood presure and pulse frequency and fit linear regression models with response variables SYS, DIA and PUL on the mean of the manually measured pulse frequencies. Check the distribution of the residuals with a dot-plot of the resiudals of the regression model on the fitted values. For which of the responses (SYS, DIA and PUL), the highest absolute regression coefficient can be found? The dataset is available at

[1] "https://charlotte-ngs.github.io/asmasss2024/data/20240429_bp_data.csv"

Tasks

• Read the data

• Compute mean of manual pulse frequencies

• Fit a regression model of SYS on mean manual pulse frequency

• Plot residuals

• Fit a regression model of DIA on mean manual pulse frequency

• Plot residuals

• Fit a regression model of PUL on mean manual pulse frequency

• Plot residuals

• Ranking of the regression coefficients

Problem 2: Median Predictor Variables

Fit the same type of regression models, but use the median of the predictor variables instead of the mean.

Tasks

• Read the data

• Compute median of manual pulse frequencies

• Fit a regression model of SYS on median manual pulse frequency

• Plot residuals

• Fit a regression model of DIA on median manual pulse frequency

• Plot residuals

• Fit a regression model of PUL on median manual pulse frequency

• Plot residuals

• Ranking of the regression coefficients