First, the graphs are out of phase at important points.  The big jump around Sep 2001 peaks well before the gas price graph hits its minimum (remember, higher the gas graph, lower the price).  This would imply that gas prices are following Bush's approval rating, whereas, the claim is that people look at gas prices, and the lower the price, the more they approve of the shrub.  But, inspection would state that this isn't entirely true.  

Second, there is at least one maximum in gas prices around Mar 2003 which is anti-correlated with a jump in Bush's approval.  

There are a few more things like this, but I picked out these two features to point out the obvious third factor.  9/11 happens to correlate much better with Bush's approval maximum, and Mar 2003 marks the beginning of the Iraq war, which itself caused an increase in gas prices for various reasons.  

I would say this correlation has more to do with the fact that Bush is mucking with political and economic factors which drive gas prices rather than Joe Sixpack deciding whether he likes Bush or not based on how much it costs to fill up his Durango.

This is your fire, so I don't think the Saudi's are going to be able to throw the election by dropping gas prices.  

Jeff

I understand about the gas price line being inverted.  I'm talking about the time order of things.  Yes, the two are correlated, but the line for Bush's approval changes before the line for gas price changes.  So the cause-effect relationship, if it exists, is the opposite from what the original post said.