Answer: D) The linear model shows a strong fit to the data
The actual strength of the relationship is unknown unless we have the actual values of each data point (so we can compute the correlation coefficient r), but the residuals are randomly scattered about both above and below the horizontal axis. This means we have a fairly good linear fit. If all of the points were above the line, or all below the line, or all residuals fit a certain pattern (eg: parabola), then it wouldn't be a good linear fit.