lumped element simulation in microwave office

i've simulated several lumped elements circuits in microwave office (e.g. branchline coupler at 1.4GHz) and the simulation result seem suit my application.
but when i implement practically using lumped element capacitor and inductor, the measurement results seems do not match the simulation result.
anyone know why?

Number of things:

a) You may be using ideal lumped elements whereas real parts have Q and series/parallel resonances.

b) Pad capacitances and part/board interactions are not taken into account.

c) Your measuremnt setup may not be calibrated or de-embedded properly.

These are good models:

http://www.modelithics.com/

i'm using capacitor and inductor model which is accessible online using awr microwave office 2003 and the vna is well calibrated.