Premise 1: Newspapers generally report on studies with more dramatic-sounding findings.

Premise 2: Newspapers report more often on smaller observational studies than larger studies. (The reliability of either type is irrelevant).

Conclusion: Smaller studies must be more likely to have dramatic-sounding results than larger studies (since newspapers report on smaller studies more often).

The way I understood it, the flaw here is basically the "fails to consider an alternative." D presents the alternative: newspapers report more often on smaller observational studies because there are more of them (not because they are more likely to have dramatic-sounding results).

"Much as" means "much like." Read it as "the best jazz singers use their voices much like horn players use their instruments" instead of "[...] use their voices as much as horn players use their instruments." The Billie Holiday statement is then evidence for that claim. :)