2011A&A...536A..44E

We test the robustness of published time delays for 11 lensed quasars by using two techniques to measure time shifts in their light curves. We chose to use two fundamentally different techniques to determine time delays in gravitationally lensed quasars: a method based on fitting a numerical model and another one derived from the minimum dispersion method introduced by Pelt and collaborators. To analyse our sample in a homogeneous way and avoid bias caused by the choice of the method used, we apply both methods to 11 different lensed systems for which delays have been published: JVAS B0218+357, SBS 0909+523, RX J0911+0551, FBQS J0951+2635, HE 1104-1805, PG 1115+080, JVAS B1422+231, SBS 1520+530, CLASS B1600+434, CLASS B1608+656, and HE 2149-2745 Time delays for three double lenses, JVAS B0218+357, HE 1104-1805, and CLASS B1600+434, as well as the quadruply lensed quasar CLASS B1608+656 are confirmed within the error bars. We correct the delay for SBS 1520+530. For PG 1115+080 and RX J0911+0551, the existence of a second solution on top of the published delay is revealed. The time delays in four systems, SBS 0909+523, FBQS J0951+2635, JVAS B1422+231, and HE 2149-2745 prove to be less reliable than previously claimed. If we wish to derive an estimate of H₀ based on time delays in gravitationally lensed quasars, we need to obtain more robust light curves for most of these systems in order to achieve a higher accuracy and robustness on the time delays.