Academic journal article North American Journal of Psychology

The Contribution of Falsework Fading to the Stimulus Generalization of a Skill

Academic journal article North American Journal of Psychology

The Contribution of Falsework Fading to the Stimulus Generalization of a Skill

Article excerpt

In 1992, Schmidt and Bjork reviewed the literature on the training, retention and generalization of new skills. They suggested that the very attributes of practice that result in rapid acquisition are also those least likely to promote either long term retention or effective transfer to novel stimuli. One example of these attributes is how the training trials are scheduled. Consistent practice using repetitions of the same stimulus ("blocked" training) results in speedy acquisition, but lower levels of retention and generalization. Training using a set of related stimuli ("variable" training), while often producing slower acquisition, increases scores on retention and generalization tests. Schmidt and Bjork (1992) suggested that the evaluation of the overall success of training should be based on more than the speed at which the new skill was acquired. They believe that long term retention and transfer should also be considered as training criteria. Over the subsequent 16 years, 449 articles were published (56% of all articles on this topic since 1915), examining the benefits of variable training and testing theories attempting to explain its effectiveness.

One of the most common of these is Schmidt's (1975) schema theory. It states that practice using a single stimulus results in the encoding and storage of a specific stimulus-response unit, which makes retrieval of the response alone difficult when a novel stimulus is presented. Practice based on an array of different stimuli encourages the learner to abstract and store a general rule (schema) that describes the relationship between stimuli and behavior. This schema is created based on many types of information stored in memory during the learning process (some only temporarily), including environmental characteristics and the specific set of physical responses made. Once the schema is learned, it can be retrieved in response to any appropriate stimulus, including those not part of the initial training.

Other theories of procedural learning also propose that there are differences between the acquisition of stimulus specific knowledge, where one-to-one pairings produce automatic responses to stimuli (Logan, 1988; Schneider & Shiffrin, 1977) and strategic level knowledge, which (with practice) can be expressed independently from particular stimulus features (Anderson, 1982; Pellegrino, Doane, Fischer & Alderton, 1991).

Other types of generalization are also issues in training research. Studies on context generalization have indicated that an immediate shift to a new environment after training generally produces a decrease in the learned behavior. However, decades of animal studies have shown that context generalization is more likely over extended retention intervals (e.g., Perkins & Weyant, 1958). For example, Biedenkapp and Rudy (2007) found that footshock-induced fear training in rats produced context-dependent behavior after a single day, but that the contextual dependency was no longer evident after one week. This pattern (see Riccio, Ackil, & Burch-Vernon, 1992, for a review) has been traditionally explained as the result of context forgetting--the initial context that guided the acquisition of the new behavior is forgotten faster than the new behavior itself. This forgetting allows the behavior to be performed in contexts that were initially disruptive.

Bertsch and Lamb (2008) found evidence (among human participants) that instead of context forgetting per se, it was actually the relationship between a new behavior and the original contextual cues that appeared to weaken over time. They gave subjects practice reading reversed text under one set of context conditions (font size) and then tested them under another. Changing contextual features immediately after training slowed reading speed. After one week however, even though all participants were able to report that the font size had changed, their mean reading speed was no longer affected by that change. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.