It's kind of like being off with earnings estimates, except a whole lot worse. Investors expect that earnings estimates are just that -- estimates. They're not expecting top-line data from a phase 2 trial to be revised. The data is the data.
Except when there's an "error in the application of the agreed statistical analysis plan," as Targacept explained yesterday. Seems someone hit the wrong button on the computer program that calculates the statistics, or something like that.
TC-6987 was given along with an inhaled corticosteroid such as GlaxoSmithKline's
On the surface, hitting only one of the endpoints isn't a major issue. It was only a phase 2 trial and the drug just barely missed on the pre-dose measurement. A larger trial with the same increase in FEV1 would surely be statistically significant.
But investors should be a little worried about a company that can't get its data straight before presenting them to investors. Sure, mistakes happen, and I wouldn't exactly say this has reached the level of distrust as at Sequenom
The worry here is that this could become a systemic problem that might result in a sloppy regulatory application. An "error" at that point will result in a larger stock price drop than the 12% fall Targacept experienced yesterday.
Investing against the grain can be risky, but also very profitable. Find out the sector smart investors are in right now. Grab a copy of the Fool's free report: "The Stocks Only the Smartest Investors Are Buying."