I've been running an experiment on WordRoom for iOS since the launch. At first we charged for the app and the usage from iOS numbers rose to the top compared to the other platforms. A month later we desided to release WordRoom for free. The result of this move, after an initial peak, was that usage dropped way below the other platforms.
I was sure this was just a coincident and left it free for the entire summer. Usage numbers dropped to below 5% iOS.
With nothing to lose I switched to a paid app once again and the players who buy the app return again.
My thesis from the beginning of the project seems valid in a non-statistically secured way. Players who pay for an app, even if it's just 99 cents, play that game a lot more since you don't want to uninstall a paid app.