Question

Consider the scatter diagram below:

200 150 > 100 1 50 0 0 20 40 09 80 100 X

This most likely illustrates the problem of:

homoskedasticity

serial correlation

multicollinearity

heteroskedasticity

Answer 1

Answer. (b) serial correlation

Explanation: Serial correlation means that the error terms are dependent on the previous error terms, and thus errors are not random. This means that on a scatter plot, we have runs of positive errors, and runs of negative errors. Whereas in a perfeclty random case, errors should not display any pattern. In this scatter plot, it can be seen that for a while there are only negative errors, then only positive errors, and then again only negative errors. So it is serial correlation problem. In a random case, errors should have been on both the sides of the line at all times.

Please upvote this answer.