In wind tunnel testing you need to match several similarity parameters; the most well known are the Reynolds number and the Mach number.
Since the Reynolds number depends on flow speed, viscosity and physical size while the Mach number also depends on flow speed, it looks like both can only be matched when the physical size of the model remains unchangedphysical size of the model remains unchanged. But when fluid pressure and temperature are also changed, it is indeed possible to get both close to the real value. This requires an actively cooled, pressurized wind tunnel.
Such tunnels are expensive and have a long waiting list, so other tunnels exist which match both parameters less well. In that case, the researcher needs to pick which is more important and select one of the many facilities in existence.
For preliminary development, it is in most cases sufficient to test at a much lower Reynolds number and then apply correction factors, just as it has been done in the early days of aviation. If, for example, you need to find the lateral stability of a design or test its spin behavior, a smaller Reynolds number is not critical. In order to trip the boundary layer where it would have its natural transition between laminar and turbulent flow, the flight surfaces of such models have artificial roughness applied at the right chord location, so the flow is made as similar as practical.