The startle effect is real, I’ve witnessed a few instances of it. There are obviously different degrees of “startle” . Fortunately, in the last 7 years, my airline has been designing training scenarios to train for startle effect. For example, for years our V-1 cut was between V-1 and V-2, on the runway, now we get the engine failure airborne at different altitudes and configuration levels. One of the best scenarios is applicable to this event. It’s a go-around at altitudes from above missed approach altitude to 1000 ft AGL. It forces you to think about where you are in the approach regime and adjust automation as necessary.