A study examining the correlation between caffeine and consumption and the speed of running during an athletic event has yeilded a value of +.65 for pearson r. While examining a scatter plot of the data, the researcher finds two data points that are very different from the rest of the data. For both of these points high levels of caffenine occur with high running speeds. Would the magnitude of Pearson r most likely increase, decrease, or remain unchanged if the researcher re-calculated after excluding these extreme outliers? Why?