[Brendan Halpin]

You may be thinking in spreadsheet terms: once qx is defined it doesn't change when the previous cases are dropped -- unlike in a spreadsheet where deleting cells that a formula depends on will change the result. Play around with the code when you get access to Stata again and it should become clearer.

[maartenc]

Thanks to all of you for helping me! I checked it again and apparently I made my conclusions to early. The dropped observations are indeed still included in qx. I only looked to the first observations where it looked liked they changed and panicked to early. Brendan I am glad you understood my 'problem'.

Best, Maarten