For a two-sample t-test with unequal variances (Welch's t-test), the degrees of freedom $\nu$ is approximated by the Welch-Satterthwaite equation. If $s_1^2=16, s_2^2=36, n_1=10, n_2=10$, what happens to the degrees of freedom relative to a standard pooled t-test?