In econometric applications, there are often many reasons why the errors may not be homoskedastic and normally distributed, so that t-ratio tests are, at best, only approximately valid. * True

* False