Linear mixed models have become a popular framework for analyzing data that have complicated design and structured covariance. It is known that exact test is not available for fixed effects except for balanced data and special covariance structures. Test statistics based on asymptotic distribution of parameter estimates are useful for large samples. However, these tests can be unreliable in applications with small sample sizes. Kenward and Roger (KR) proposed a scaled Wald-Type statistic and a method to approximate its small sample distribution. This presentation discusses the motivation and approach of the Kenward-Roger method. This presentation also describes caveats of the Kenward-Roger method as it applies to a nonlinear covariance structure and how these caveats motivated the latest improvement for the Kenward-Roger approach (Kenward and Roger 2009).