More accurate than what? If you're implying it would be more accurate than profiling Muslims, I think you are incorrect. Muslims as a group make up 1% of the US, but a generous, biased study [0] found that they are responsible for 48% of deaths from terrorism since 9/11. From this it follows that the average Muslim is responsible for approx 48 times more deaths than the average American of any religion.
And this is without considering whether the TSA considers categories that have more predictive power than simply being Muslim, e.g. following a particular kind of Islam.
Nitpick, Deaths due to TERRORIST attacks. Not arbitrary deaths....
"From this it follows that the average Muslim is responsible for approx 48 times more deaths than the average American of any religion."
^ This statement is false as it assumes all deaths are from terrorist attacks, based on your previous statement of 48 times more likely to commit TERRORIST attacks, of which has a silly low death rate compared to ANYTHING else in the US.
And this is without considering whether the TSA considers categories that have more predictive power than simply being Muslim, e.g. following a particular kind of Islam.
[0] http://securitydata.newamerica.net/extremists/deadly-attacks...