Title-The effects of disfluency on memory AbstractTo test the the disfluency effect, in accordance with the Diemand-Yauman et al (2010) study, we tested the effects of disfluent versus fluent font in memory, and therefore learning. This could therefore have applications in a learning environment. The independent variable in this case was the type of font, with the dependant being the amount of correct answers in recall. The experiment was conducted on six Bath Spa students, free from learning disabilities and sound or corrected vision. We found that the results were statistically significant, with t= 4.9497, p<0.
0078, two-tailed, finding that disfluency had a positive effect on recall. This could potentially impact further research in learning and cognitive capabilities. Introduction Diemand-Yauman, Oppenheimer and Vaughan’s (2010) study, researches the effects of disfluency on learning, therefore with practical applications to the education and retention of memory to participants. For instance, Diemand-Yauman et al,.(2010) set out to measure the effects of disfluent language, consequently impacting the educational results.They set out by devising two tasks, aimed at testing whether increasing the difficulty of the interpretation of the texts for participants, through a disfluent condition (using the Comic Sans font) and a fluent condition (using the Arial font). The expectation was that by changing the font for the participants to read pieces of text off, pps.
Would therefore focus on the information more thoroughly, as it is harder to read, they would henceforth retain the information better, and experience a more long-term retention (Bjork, 1994). This theory can however be contradicted, with research into education and retention, many believe that an easier approach of encoding information is the preferred way of teaching, as students have a reduced amount of extraneous variables such as, difficulty understanding the font and/or concept, therefore increasing the success of the session, hence benefitting the learner. (Sweller & Chandler, 1994). It may be the case that this is a more common method of teaching and learning, however the research by Oppenheimer et al.
(2010) may contradict this.Craik & Tulving (1975) found adding cues of cognitive loads enhanced cognitive engagement on a particular subject, hence benefitting learning. They were called, “desirable difficulties”. These “desirable difficulties” engaged areas in the brain responsible for processing, allowing the individual to hence retain the information better, and for longer. Researchers Diemand-Yauman, Oppenheimer& Vaughan (2010), researched this theory by gathering 28 pps, (aged 18-40) and asking them to learn facts about aliens in 90 seconds (1 minute and 30 seconds).
This presented a cognitive stimuli for the pps. However the independent variable was the kind of font, (either disfluent or fluent). After 15 minutes of a cognitive ‘breather’/break, the pps.
Were then asked questions about the information they had learnt previously. Oppenheimer et al (2010) found that those in the disfluent condition recalled 14% more information than those in the fluent,(72.8% compared to 86.5%). This therefore shows the results of the disfluency affect.
Oppenheimer et al., (2010)study further supports the affects of the disfluency affect with a second, related study. In this case, the disfluency affect was tested in a field experiment, rather than a lab. This adds a different dimension to the affects of disfluency on learning, as it can be seen as being more applicable to the real-world. The experiment consisted of 222 high school student pps.
(aged 15-18) being separated into 2 categories, a disfluent and control group (again using clear and disfluent fonts). However, slightly differing from the previous study, after the students looked at the disfluent or control worksheet/powerpoint (the stimulus) they were then asked if the disfluency affected their motivation, analysed by using a survey on a scale of 1-5 for the questions. The results showed that students in the disfluent condition scored higher in assessments than those in control condition. This study also showed that the disfluency affect can be seen across numerous areas of study, for example, the hard sciences and the humanities. In accordance with Oppenheimer et al (2011) study, and the “desirable difficulty” affect as described by (Craig& Tulving, 1975), I aim to test the affect of the disfluency affect on 6 pps. By using ‘Ebrima’ font for the fluent condition, compared to ‘algerian’ for the disfluent condition. I predict to find a difference between the disfluent condition and the fluent condition, as research evidence in this area of study has given key insight into the theory.
Method Design:The research employed an independent measures, experimental design, with the participants. Allocated to both the disfluent (‘algerian’) and fluent condition (‘Ebrima’).The participants were randomly allocated however there were some ground rules set out in order to choose the participants. The independent variable in this experiment, was the type of font, operationalised through the disfluent or the fluent text in the statements the participants had to remember.
The dependant variable was the number of correctly recalled facts about mermaids the participants remembered. In terms of making sure the experiment followed the Ethical guidelines necessary, through the BPS Code of Ethics. All participants were asked to give their permission to be involved in the study, meaning they were knowingly taking part of their own accord. As well as this, participants were given the right to drop out of the study throughout any stage, and participants were given a debrief sheet after the experiment was undertaken, to avoid deceit issues in the experiment.Participants:Participants in this study were selected from certain characteristics to make the study as representative as possible, in order to lower the effects of extraneous variables. Six participants in total were chosen for the study. To participate, as far as inclusion material, the participants must be Bath Spa University Students, with normal or corrected to normal vision. It was decided however that participants with learning difficulties or reading impairments such as dyslexia were excluded.
Within the six participants, three were allocated to either the fluent or disfluent groups, an individual groups, experimental design. It was decided as well, that participants be students, as they may have similar cognitive abilities due to age, and being present in education at the time, possibly affecting the results. Participants when chosen, were then randomly allocated to either the fluent condition or the disfluent, then the results were able to be collected.Materials/Apparatus In order to conduct this experiment, after successfully choosing and randomly allocating your participants, you will need a few materials. Firstly, you will need to acquire some fictional facts on mermaids. A good basis is the following website (seen in appendix). You will then need a stopwatch to time the participants.
This will provide a distraction period between the tasks. The stopwatch used was the following, ‘Digital Sport Stopwatch Timer Chronograph Athletic Watch with Clock Alarm, Calendar and LCD Display, Blue’, by ‘Leap’, sold by Willbond Tech, found at (seen in appendix ), however any stopwatch can be used. After the time of ten minutes is finished, you will need a Christmas word search. To acquire this, go to (https://www.pinterest.co.
uk/pin/509751251548049928/, by https://printables4kids.com/getting-ready-for-christmas-word-search/christmas-wordsearch/, 2006).ProcedureParticipants were collected, three of which allocated to the fluent group, the other three then allocated to the disfluent group. Participants were asked individually to complete the task.Participants were asked to memorise a list of fictitious facts about mermaids, from the following lists of questions. They were given 90 seconds (or one minute and thirty seconds) to learn as many facts as they could in the given time scale, using the stopwatch. The facts used were as follows.
-They can only breathe underwater for some time then must come up for air-Their tails are no longer than 7 feet-Mermaids can sing songs luring sailors towards them-All mermaid scales are fluorescent- Mermaids have gills -Mermaids have pet catfishParticipants were then timed, using the stopwatch for ten minutes, to provide a distraction. Participants then undertook a Christmas word search. After this, they were then asked to recall as many facts as they could remember from the previous list of facts.The facts were presented in the fluent condition though the use of the font ‘Ebrima’, and the facts in the disfluent condition were presented in ‘algerian’.
Results Table 1- Mean/median/mode for conditions 1 and 2 mean median modeCondition 1-fluent 3.33 4 4Conditon 2- disfluent 5.67 6 6 In table 1, we can see that the mean, varies between condition one and condition two, and as predicted, the disfluent condition has a higher mean. Looking at figure 1 we can see a visual representation of the data, showing the expected . It is assumed the data is normally distributed, and the standard deviation found was 1.211060142.
In order to analyse the results, a t test was used. This form of data analysis was used, as the differences between condition one (fluent) and condition two (disfluent), as the design is unrelated and in individual groups design. The t test was used on the data of conditions one and two, this was also used as there was two sets of data.The results were t= 4.9497, p<0.0078, two-tailed. The N value being 3. By looking at the critical values for t, we can see that the critical value in this case was, 2.
78, (at the 0.05 confidence level), meaning the results were statistically significant at the 95% level. This means the difference between fluent and disfluent conditions is therefore significant and is statistically supported.
Discussion In this study, we can see that the original assumptions of the study were in fact satisfied by the results. We have evidence that the disfluency effect does alter recall skills, though cognitive activities. From the results, we can be 95% sure that the data is statistically reliable, meaning this provides support for the theory.In comparison to Diemand-Yauman et al (2010) study, we can see that there are similarities to this study, ‘the effects of disfluency on recall’ and the Diemand-Yauman et al (2010) study (‘effects of disfluency on memory).
The results found in Diemand-Yauman et al. (2010) study, show that they were also statistically significant differences. (t(220) =3.38, p< .001).
This provides evidentiary support for the disfluency argument, as students performed higher in the disfluent condition. The study demonstrated the effects of disfluency on memory, through recall. It is suggested by Alter et al., (2007) that this is due to an increase in cognitive processing in the brain. However, there are other possibilities that can take shape in affecting memory.
For example, Begg et al,. (1986), suggested the importance of ease of processing, especially with clear compared to blurred words. This contradicts the previous assumption, as it suggests people will remember clearly marked stimuli easier than hard to read stimuli. Yue et al,. (2012) found that ‘judgements of learning’ (JOLs) (Yue et al.
, 2010) were actually higher for fluent and clearer words. This experiment conducted shows that disfluency does in facts have an effect on memory and hence recall of information. However, in evaluating the experiment, it is hard to be sure the results are entirely due to the effects of a heightened cognitive processing. Rhodes and Castel (2008), found that recall was unaffected by word size. Additionally, there are many other studies, for example, (Glass, 2007), that discredit the theory of the disfluency effect, suggesting that as difficulty of perception increases, memory decreases. In this experiment, it is hard to know whether the results are due to the disfluency effect, they could be due to other reasons.
Even though the analysis of the data shows that it was statistically reliable, the results may have been skewed by an unknown confounding variable. For example, the one group of participants may have had a higher IQ than the other, meaning the results may have changed, or the participants could have had individual differences such as energy for example on the day of testing. There is no way to be sure, but if repeating the experiment, a few changes could be made. For example, the participants could undertake an intelligence test before the experiment. This may change the results as those with a higher intelligence may perform better in the experiment than those with a lower intellectual capacity.In addition to this, it would be beneficial to the investigation ‘the effects of disfluency on memory’ if there were more participants to test.
There were six participants tested in this experiment, meaning the results may not be generalisable. Although the result was significant, it would be useful to repeat the experiment with separate participants, to highlight any differences.