Firstly conceptualized by
Bjork (1995), the idea of desirable difficulties is one that claims performance
can be improved by making material harder to learn. The term disfluency describes
one such difficulty. It involves the manipulation of stimulus characteristics like
font style. Such manipulation has been stated as being a metacognitive cue – one
that alters an individual’s subjective perception of difficulty. Hence,
individuals are compelled to use more analytical and elaborative processing
techniques to appropriately tackle such problems (Alter et al. 2007; Yue et al.
2013). Alter et al. (2007) refer to theories regarding dual processing systems
(James, 1950; Tverski and Kahneman, 1974) to visualize this concept and assess
that the activation of either system depends on the perceived difficulty of a
task. System 1 has been theorised as working intuitively and without much
conscious effort whereas System 2 contrarily involves effortful and systematic engagement.
As disfluent material is subjectively interpreted as being more difficult, individuals
metacognitively initiate the use of System 2 (Seufert et al., 2017). This
considered, assuming that disfluent tasks result in better performance would be
logical. However, Kuhl and Eitel (2016) observe in their overview of literature
that while the disfluency effect has been demonstrated by some studies
(Diemand-Yauman et al. 2011; Eitel et al. 2014; Sungkhasettee et al. 2011),
others found it as having either no or even negative effects on performance
(Rhodes and Castel, 2008; Yue et al. 2013).
Moreover, there is
acknowledgement that disfluency is closely associated with cognitive load
theory (Lehmann et al. 2016; Seufert et al. 2017). The theory claims that
intrinsic (task-relevant) cognitive load (ICL) is mediated by task element
interactions (difficulty). Extraneous (task irrelevant) cognitive load (ECL) on
the other hand is affected by format of instruction, for example, the legibility
of text. The theory implies that decreases in word legibility increase
perceived ECL and thereby inhibit learning; this directly contradicts the
concept of desirable difficulties. Hence, perceived levels of ECL are assumed
as being positively correlated with disfluency levels (Eitel et al. 2014). Degrees
of ICL by contrast may be determined by the nature of a processing task –
whether it be of lower cognitive order i.e. recollection or comprehension, or
of higher cognitive order i.e. transfer (applying knowledge to a novel skill).
In relation to disfluency
being considered with respect to cognitive load, Lehmann et al. (2016) analyse
the importance of working memory capacity (WMC). They reported in their
findings that both considerable disfluency and higher order tasks result in a
cognitive overload and thereby a strain on cognitive resources. This hinders
any positive effect disfluency may have as System 2 activation is inhibited.
Therefore, the rationale of the present study was based on a question that
naturally ensued – could disfluency impact on performance persist if ICL and
ECL were minimised? As this idea was largely unexplored, and keeping in mind
the homogeneity of prior literature, it was hypothesized that disfluency would influence
visual recall task performance when ICL and ECL were relatively controlled for.