Coincident ozone measurements by the Solar Backscattered Ultra Violet (SBUV) and Stratospheric Aerosol and Gas Experiment (SAGE) II instruments and by ozonesondes and the Umkehr method are used to estimate stratospheric ozone variability and standard uncertainties of these different measurements. Below 20 km over northern midlatitudes, estimated measurement uncertainties for SBUV(/2), sondes, and Umkehr are similar (~8% for 0 to 20 km integrated ozone), although only sondes have a high vertical resolution there. From 20 to 28 km., the estimated uncertainties (4--6%) for all four measurement sources are substantially smaller than the ozone variability in winter (10--15%), but they are comparable in summer (~5%). Above 28 km, sonde uncertainties are larger than or comparable to the ozone variability, and much larger than uncertainties of SBUV(/2), Umkehr, or SAGE II data. Umkehr measurement uncertainties at 24--32 km are about 5% and are lower than sonde uncertainties (7--13%) at these levels. SBUV(/2) data are used to evaluate differences between different types of ozonesondes, and to show that correction by total ozone measurements noticeably reduces ozonesonde uncertainties. The latitudinal dependence of ozone variability and instrument uncertainties is studied using pairs of collocated SBUV(/2) and SAGE II measurements. There is good correlation between these measurements over middle and high latitudes. Over the tropical region, the correlation coefficients are modest (about 0.5) but significant in all layers except at 28--31 km. It is noted that where ozone variability is comparable to instrument uncertainties, comparison with a climatology based on a large number of observations may provide better insight into instrument performance than intercomparison of a small number of quasi-coincident measurements.