Words are made of letters, and yet sometimes it is easier to identify a word than a single letter. This word superiority effect (WSE) is observed when singly presented written stimuli are presented very briefly or degraded by visual noise. It is unclear at which level in visual perception this effect arises, and how general it is. This study aimed to investigate the perceptual basis for the WSE: Is it due to a lower threshold for perception of words, or a higher speed of processing for words than letters? Furthermore, we wanted to investigate if this word advantage is also evident when multiple stimuli are presented simultaneously: Are words treated as units or wholes in visual short term memory? Using methods based on a Theory of Visual Attention (TVA), we measured perceptual threshold, visual processing speed and visual short term memory capacity for words and letters, in two simple psychophysical experiments. Using briefly presented single stimuli (words and letters), we show that the classical WSE is specifically reflected in perceptual processing speed: words are simply processed faster than single letters. Intriguingly, when multiple stimuli are presented simultaneously we find a different pattern: Letters are perceived more easily than words, and this is reflected both in perceptual processing speed and short term memory capacity. So even if single words do come easy, they seem to enjoy no advantage in visual short term memory.