[lemon_3nigma]

My question is: why is the H of an alpha carbon shifted more upfield when compared to a H of -CH2 beside the O on the 1NMR spectrum? This was an explanation to one of the NMR questions I was doing.  To better clarify, use the link for graphical representation.

http://www.steve.gb.com/images/molecules/esters/ethyl_propanoate.png

I always thought the compounded effect of the pi bond overlap of the C=O and the electronegative carbonyl oxygen deshielded more of the surrounding than just the O itself in the ester group....can someone help explain what I'm doing wrong?