Exit polls - don't trust 'em!

I am now putting on my statistician hat.

The networks make a big deal of 'exit polls' - querying voters as they emerge from polling place:
- "For whom did you vote?"
- "Do you describe yourself as Christian?"
- "What's your income bracket?"

... sadly - these barometers have very little reliability.
The folks conducting these 'polls' are more likely than not to approach only those with whom they feel comfortable. The 'sampling' provided is perhaps 'representative', but NOT random!

My bet would be that exit polls over-represent more-or-less affluent voters - folks who are well-dressed, well-groomed, sort of 'normal' looking when they emerge from polling place. This would go a long way towards explaining MSNBC's recent claim that wealthy voters were turning out in larger numbers than in previous elections.
[note: I've absolutely NO data to back up this suspicion... it's just a guess.]

The most famous 'poll' of this sort is reflected in pic below:

Yeah - they're a lot of fun to read, BUT - don't believe 'em!

... on the other hand: REPORTING on 'exit polls' just MAY influence future voters!

Assessing that elusive entity known as 'public opinion' is VERY difficult... and 'exit polls' are a particularly lousy way to go about it!
(... BUT: they are fun to read!... and they DO contribute to overall employment!!!)