What's in a meme? Internal letter transposition, word length and eye-voice span

What's in a meme? Internal letter transposition, word length and eye-voice span

To investigate if internal letter transposition (ILT)
decreased reading performance and memory output in a rehearsal-free memory task using the eye-voice span (EVS) as the measure. The major source was an internet meme (a rumour, urban legend or concept propagated via email)
suggesting that ILT did not affect reading performance, provided the initial and final letters in words remained unchanged. An Interrupted Reading Task (IRT) tested the
concept of the original meme. Reading speed, number of words output and output duration were measured by analysing sound files recorded from the IRT. The IRT presented text passages on a computer screen to participants, who immediately began reading them aloud,
until a predetermined point at which the experimenter caused the text to disappear from view, replacing it with a random pattern mask, while participants continued reading aloud for as long as they could. Text passages were designed so that the words following the interruption
were either mostly long (3 or more syllables) or short (one syllable). Additionally, passages were either presented as unmodified text or as ILT stimuli, where the internal letters were randomised whilst leaving the initial and final letters intact. Word length was manipulated within subjects, and ILT condition between
subjects in a mixed ANOVA. Recorded output represents the contents of the EVS, which relates to short term memory capacity. Significant main effects obtained for both the ILT manipulation (ILT worse) and word length (long
words worse) on the measures of number of words output and output duration, as well as a significant interaction showing the ILT further reduced output measures when passages contained longer words. A significant difference
in reading speed was evident, with ILT passages read much slower than unmodified text. Data and analyses support both the DRC model of reading (Coltheart et al., 2001) and a global account of word recognition (Frankish & Barnes,
2008) as well as reiterating the robust and wellknown
Word Length Effect (WLE; as noted in Baddeley et al., 1975) – however the IRT is a rehearsal-free task and as such the results do not support rehearsal accounts of the WLE. The content of the original meme is, unsurprisingly,
refuted by these data.

Only abstracts published in conference proceedings, as supplied here.