(i)

A correlation coefficient close to +1 or -1 indicates a strong linear relationship between X and Y. The regression equation is therefore more reliable for forecasting.

(ii)

Working to a high number of decimal places gives spurious accuracy unless both the data itself is accurate to the same degree and the methods used lend themselves to such precision.

(iii)

Forecasting for values of X outside the range of the original data leads to unreliable estimates, because there is no evidence that the same regression relationships hold for such values.

(iv)

The regression equation is worthless unless a sufficiently large sample was used to calculate it. In practice, samples of about ten or more are acceptable.