This paper discusses over the air (OTA) testing for multiple input multiple output (MIMO) capable terminals with emphasis on wideband MIMO channel capacity analysis in a multi-probe anechoic chamber setup. In the literature, the spatial correlation simulation accuracy at the receiver (Rx) side has been used to determine the test area size for a limited number of probes. However, it is desirable that the test area size is defined in terms of data rate deviation of the simulated channel in the laboratory from that of the target channel model. This paper reports MIMO capacity analysis results for wideband spatio-temporal channel models, with emphasis on the impact of spatial correlation at the transmit (Tx) side, the channel model, and the spatial correlation at the Rx side on the capacity simulation accuracy. Simulation results show that the number of probes is irrelevant to capacity simulation accuracy when the spatial correlation at the Tx side is in the high region (e.g. $\rho>0.7$). Furthermore, when correlation at the Tx side is low, the spatial correlation accuracy is less critical with small correlation at the Rx side. The simulation results are further supported by measurements in a practical multi-probe anechoic chamber setup. The capacity simulation accuracy is shown to be a valid measure to determine the test area size.