Tremaine and Weinberg ( TW ) proposed a conceptually simple procedure relying on long-slit spectroscopy to measure the pattern speeds of bars ( \Omega _ { p } ) in disk galaxies .
Using a simulated galaxy , we investigate the potential biases and uncertainties of TW measurements using increasingly popular integral-field spectrographs ( IFSs ) , for which multiple pseudo-slits ( and thus independent measurements ) can be constructed with a single observation .
Most importantly , to establish the spatial coverage required and ensure the validity of the measurements , the inferred \Omega _ { p } must asymptotically converge as the ( half- ) length of each pseudo-slit used is increased .
The requirement for our simulation is to reach \approx 1.3 times the half-light radius , but this may vary from galaxy to galaxy .
Only those slits located within the bar region yield accurate measurements .
We confirm that the position angle of the disk is the dominant source of systematic error in TW \Omega _ { p } measurements , leading to under/overestimates of tens of percent for inaccuracies of even a few degrees .
Recasting the data so that the data grid aligns with the disk major axis leads to slightly reduced uncertainties .
Accurate measurements are obtained only for well-defined ranges of the bar angle ( relative to the galaxy major axis ) \phi _ { bar } and the inclination angle i , here 10 \degr \la \phi _ { bar } \la 75 \degr and 105 \degr \la \phi _ { bar } \la 170 \degr and 15 \degr \la i \la 70 \degr .
The adopted ( pseudo- ) slit widths , spatial resolution , and ( unless extremely aggressive ) spatial binning of IFS data have no significant impact on the measurements .
Our results thus provide useful guidelines for reliable and accurate direct \Omega _ { p } measurements with IFS observations .