The reason Obama doesn't argue that raising taxes is potentially good for our current economic situation is that winning this election is given primacy over winning that economic argument. And he's smart enough to conclude that talk of raising taxes, even if it'll lead to desired economic results for our country, won't appeal to voters because of years of “taxes are bad” type conditioning/memes.

In Gladwell's case, though he fashions his writing as intellectual fare, it's more aptly characterized as entertainment writing for intellectuals (or wannabes as the case may be). His entire enterprise, whether he realizes it or not, is about appealing to the intellectually curious, but only if it's first an entertaining read. It's difficult to constantly come up with theses that fulfill that criterion. This inevitably leads to data selectivity and/or biases. I don't think he's being intentionally dishonest, it's just that his desire to entertain the reader can (and does) obfuscate his desire to be academic or scientific.