Testing for and Correcting for Serial Correlation

Correcting for positive serial correlation will most commonly result in:

Incorrect. The change in _t_-statistics and p-values move in opposite directions.

Correct! The problem of serial correlation (either positive or negative) is found in overstated _t_-statistics and understated p-values. Correcting for serial correlation with the use of robust standard errors will reduce the _t_-statistics and increase the p-values to more reasonable levels.

Incorrect. The problem of serial correlation is overstated _t_-statistics to begin with; a solution would not include further increases in _t_-statistics.

lower _t_-statistics and p-values.

lower _t_-statistics and higher p-values.

higher _t_-statistics and lower p-values.