Social attention with real versus reel stimuli: toward an empirical approach to concerns about ecological validity