Summary and Critique Effects of retention intervals on receiver operating characteristics in artificial grammar learning

Documentation:

Tunney, R. J., & Bezzina, G. (2007). Effects of retention intervals on receiver operating characteristics in artificial grammar learning. Acta Psychologica, 125, 37-50.

 

summary

 

        In the past three decades, research showed that recollection and familiarity played an important role in recognition.  Indeed, they made progresses on the discovery, which is from observing memorization in AGL to understanding dual process in AGL.  However, the insight of implication in the relationship between recollection and familiarity toward AGL was not brought out.  Thus, the aim of this study, beside reviewing the past thirty years literatures, is to provide further evidence for episodic memory in AGL by examining the effects of retention intervals on recollection and familiarity.  Also, the dissociation between recollection and familiarity that involves in AGL is also examined in this study.  A experiment was administered for the purpose of this study.  In the experiment, twenty-nine people of the University of Nottingham community participated in this study.  Twenty-four were female, and five were male.  Sequences generated by the AGL were presented to participants for ten times in separated blocks.  Later on, participants would take the test by pressing either "yes, I have seen the sequence before" or "no, I have seen this before" on the computer.  And, tests were administered in three different day; that is, the immediate day the receive the knowledge, seven days after, and two weeks after.  Also, participants were asked to answer how confident they were in accuracy of their decision on a ten point scale after each tests.  After analyzing the collected data by applying ROC curves, the results showed that the findings of effects by using recollection and familiarity were the same.  Furthermore, according to the slope of zROC curves, the recognition decisions for the first tests were based on recollection more than familiarity; whereas, the last tests recognition decisions were based on familiarity more than recollection.  The limitations were discovered in the discussion part as well.

 

Critique

 

        After reading this journal article, there are still pros and cons that I am going to illustrate in the below.

 

        Basically, the value of this study can not be ignored.  Reviewing the past thirty years data is quite clear that readers would have a complete idea about the motivation.  And, there is no doubt that experiments were conducted precisely since the arrangement of the three-times tests items indeed matched the purposes of this research.  Besides, by using zROC curves, results of collected and analyzed data is presented and explained thoroughly as well as the discussion part.

 

        Still, there are some flaws needed to be improved.  First of all, the literature review part seems to be separated into two parts, which one of them was put together with the introduction part, and the other one was put into the discussion part.  This way, I find it hard to scan for key points since the article is not well-organized.  Besides, the group of participants were really too small, and the proficiency of those participants was not included in the article even if it may cause little influence to the result or the whole study.   More importantly, whether the result, done by artificial grammar task, could be put into the environment of natural language learning or not still requires further evidence because natural language is considered to be more complicated.