Question

From a scatter plot of data r² = 0.0039 and no correlation is shown between exam marks and hours spent studying. The line of best fit of y_p= -0.197x + 81 shows a negative slope, showing that the more you study, the lower the exam mark is. What changes to the data set could be made in general to account for this error.

Answer 1

This could be due to presence of outliers in the data. Correlation coefficients are very sensitive to the presence of outliers in the data and this could be the reason we have got vague results. You could do the following changes in the dataset to account for the errors:

1. Check if there is any incorrectly measured or incorrectly entered data point in the set. This could happen and you would have to remove those points.

2. Remove the outliers and then check the results because presence of an outlier could change the results significantly. Just do remember to always state in a footnote that you have removed an outlier.

3. Another option is to try a different model, like an exponential model or a polynomial model with a higher degree.