Here’s an experiment idea:
It’s common in RPGs to have some sort of religion system, generally with a multitude of deities which are more-or-less at peace with one another, and whose followers typically do not try to murder each other at the first opportunity. Sometimes, you can offer donations to churches to get some bonuses, to karma or luck or whichever is applicable. Knowing the right time to pray is even a key gameplay element in Nethack.
This sort of behavior is, of course, also observed in the real world – people will often pray and make donations seeking to get something in exchange. The difference is that, in the game world, they actually get something out of it. But they don’t have to.
Consider an MMORPG with a religion system (whether it’s monotheistic or polytheistic is not important), and a series of temples spread around the world. Players can visit those temples, and consult a list of services that the temple can perform, along with their costs. Perhaps a player can get a 5% extra to-hit bonus for a donation of 100 gold. For 500 gold, he will get a +5% bonus chance to find rare items… Or so the temple claims. The player spends his hard-earned coin, and the game says something along the lines of “you feel lucky”. And nothing changes.
How many people would believe that? Would anyone conduct a systematic, scientific in-game research to evaluate how much of a difference said donations would provide? Even if some “skeptic” told other players that it’s a hoax, would they believe it? Perhaps they’ve donated once, and found a very rare item afterwards. Their minds would be making connections. What if this “truth” is spread in forums, FAQs and wikis… perhaps in the game manual itself? After months investing money in those things, wouldn’t the player feel even more compelled to believe that he wasn’t being cheated all along?
The idea could be developed further and let players take on the role of priests, although a mechanic would have to be designed to allow them to mess with the system without easily exposing its truth. For example, perhaps there’s a holy book defining how those bonuses work, in a cryptic (and possibly self-contradicting way) and leave it to the priests to interpret it and write the list of services. Temples that offered so much that it was visible that it didn’t work would lost trust, and temples that offered too little wouldn’t be able to compete. Some form of selection would eventually choose the best religion.
To my knowledge, no game has ever implemented such a system (if you know of one, please mention it on the comments). If it works as intended, analysis of player’s reactions to the system (and to the discovery that it was all their imaginations, if the developers ever decided to Word of God (pun unintended) it) could be very enlightening. Would that change how they perceive religion in the real world? Would such a study have any impact on understanding the psychology of belief? Perhaps not. But, if nothing else, it would be an interesting topic to bring up in a religion discussion.