INSTRUCTION IN LEARNING A TEMPORAL PATTERN ON AN ANTICIPATION-COINCIDENCE TASK

Share Embed


Descripción

The effect of instruction on learning a temporal pattern in an anticipation-coincidence task

Authors: Albinet, C. & Fezzani, K. Correspondence about this paper should be sent to Khaled FEZZANI., UFR STAPS Université Paul Sabatier, 118 Route de Narbonne 31062 Toulouse Cedex 4, France. Email : [email protected]

Abstract: Using a computer-simulated anticipation-coincidence task, the main aim of the study was to examine the effect of the type of instruction on learning a temporal pattern. In this task, participants must learn to anticipate the appropriate time for launching a projectile to hit a moving target. The experiment involved three instructional conditions. In the first condition, the Explicit-Rule discover Instruction condition, participants were informed that target speed could change from trial to trial and that change is controlled by a regular pattern. Their task was then to search, to identify and to use such pattern in order to enhance their anticipation. In a second condition, the Explicit-Informative Instruction condition, participants were, however, allowed to examine attentively, before practice, the regular pattern. Participants were also explicitly recommended to use the pattern that they observed to ensure a better interception of the target. Finally, in a third condition, the Implicit Instruction condition, participants are only informed that their task is to hit, or at least, to place the projectile as near as possible to the target. No additional information was provided about the target behaviour. The obtained results indicated that learning of the temporal pattern was more important in Implicit than in Explicit-rule discover Instruction condition. However, the study revealed that the Explicit-Informative Instruction condition produced unambiguously the highest amount of learning. Overall, the study highlights the role of information over guidance in the understanding of the effect of the instructions on learning. Finally, we discussed the implications of these results on the comprehension of the variability of the impact of the instruction on learning.

Key-words: Instruction, learning, temporal pattern Running head: Learning and Instruction

The effect of instruction on learning a temporal pattern in anticipation-coincidence task

The main objective of the present study was to evaluate how the type of instruction, provided before practice, can affect learning of a temporal pattern in the context of an anticipationcoincidence task. Anticipating when rapid and discrete events may occur represents an important component of adapted behaviour. Under some conditions, participants must learn to match their actions, or the effects of their actions, with the occurrence of some specific events. Learning under such conditions reflects the detection and the use of the temporal pattern characterizing the occurrence of successive events. Following the work of Pew (1974) and Nissen & Bullemer (1987), we can globally identify at least three different learning conditions corresponding to the type of instruction provided before practice of a particular task. In the first condition, participants are explicitly recommended to search, to detect and to actively use the critical regular features of the task in order to enhance their performance. Naturally, participants in this condition are not informed about which kind of features they may search and use. In a second condition, however, participants receive full information about the critical regular features of the task. They are also recommended to use such information when they are faced with the task. Finally, in a third condition, participants are not informed about the critical features of the task and are even not recommended to search them. Usually, instruction only indicates how to produce specific responses to specific stimuli. To simplify, we will call the first, the second and the third condition the Explicit-rule discover, the Explicit-informative, and the implicit instruction condition, respectively. As expected, numerous studies (Cohen, Ivry & Keele, 1990; Frensch, Buchner, & Lin, 1994; Willingham, 1999) have revealed that learning is highly sensitive to the type of instruction. The amount of learning varies depending on the type of instruction provided before practice. Less intuitively, some studies revealed that learning could be more pronounced under the Implicit Instruction condition than under Explicit-rule discover or even under the Explicit-Informative

The text body 2/13

Instruction Condition (Bright & Freedman, 1998; Hardy, Mullen, Jones, 1996; Magill, 1997; Magill & Clark, 1997; Masters, 1992; Maxwell, Masters, & Eves, 2000; Maxwell, Masters, Kerr, Weedon, 2001). Reber, 1989). This implies that the active intervention of the instructor, represented by the use of Explicit-Informative and Explicit-Rule discover instructions, produces a paradoxical negative effect on participants’ learning. The general explanation given to such effect suggests that, depending on the type of instruction, a participant may engage in a different learning process. Under Explicit-instruction conditions (rule discover or informative), participant may engage in a learning strategy similar to that used in problem-solving tasks (Newell & Simon, 1972) such as hypothesis testing and verification. Presumably, under such conditions and in each trial, the participant would verify if the observed event is predictable or not by the past events and, should this be the case, what kind of event such event would predict in return, and so on. Such strategy may interfere, and presumably inhibit, the efficient processes of incidental (i.e., non-intentional) extraction and use of the regularities which, however, are probably used under Implicit Instruction conditions (Hayes & Broadbent, 1988; Green & Flowers, 1991; Magill, 1997; Masters, 1992; Shea, Wulf, Whitacre, & Park, 2001). We may note, however, that other studies did not confirm the fact that the use of Explicit-rule discover or Explicit-Informative instructions produces a negative effect on learning (Green and Shanks, 1993; Shanks, & St-John, 1994; Singer, Lidor, & Cauraugh, 1994). For example, Curran & Keele (1993) reported that learning and performance were more important in the condition in which participants received full information about the pattern used in a serial reaction time task. In the same vein, using a golf-putting task, Bright & Freedman (1998) replicated Masters' (1992) basic findings, but they argued that the results are better explained in terms of differences between the learning and testing conditions. Consequently, they stated that their study gives no support for the claim that Explicit-Rule discover and Implicit instructions produce qualitatively and quantitatively different learning processes.

The text body 3/13

Taken as a whole, the general observation is that the literature concerned with this issue seems to be far from univocal. Hence, the main objective of the present study was to provide additional evidence that may clarify the relation between the type of instruction and learning. We note also that the preceding studies used tasks in which critical stimuli can be colours (Willingham, Nissen, Bullemer, 1989) shapes (Mayr, 1996), letters (Hartman, Knopman, & Nissen, 1989), digits (Hoffmann & Koch, 1997), or positions (Fezzani et al., 2000; Nissen & Bullemer, 1987). However, as suggested by Seger (1994), little is known about how participants would detect and use temporal patterns and how instruction could affect learning such task feature. Accordingly, to address this issue, the present study uses an anticipation-coincidence task in which temporal events are the critical regular features guiding responses. In this task, participants must learn, with practice and as a function of the type of instruction, which is the appropriate time for launching a projectile in order to hit a moving target.

Method Participants. Forty-two students (27 females and 15 males) from the University Paul Sabatier participated in the experience and were randomly assigned to experimental conditions. Their mean age was m=23 (SD= 3.6) and all had normal or corrected vision. Task and apparatus. A computer-simulated anticipation-coincidence task (see Figure 1) in which participants observed a target moving horizontally a cross a video screen from the right to the left. Five fundamental manipulations characterized the task. First, during the target trajectory, the target was occluded. Participants only see the target in the starting and stopping positions. The appearance of the target in the starting and stopping position was accompanied by a sound. Second, the duration between the appearance of the target in the first and second position was 2000 ms, 2500 ms or 3000 ms. Third, participants responded by pressing the space bar to launch a projectile (a black square) which moved upward with a constant speed from the bottom to the top of screen. 950 ms were necessary for the projectile to intersect the target. Fourth, participants had to anticipate the exact

The text body 4/13

moment when the target will appear next time to launch the projectile at the right moment. This implies that, to successfully hit the target, participants must wait 1050 ms, 1550 and 2050 ms before launching the projectile corresponding respectively to the durations 2000 ms, 2500 ms and 3000 ms. Longer or shorter latencies than the recommended duration produced errors. Finally, a program controlled the experiment and collected the Temporal Errors that correspond to the difference between the appropriate time (1050, 1550 and 2050 ms) and the time actually needed by participants to launch the projectile. Temporal errors can be negative or positive depending on whether the participants responded earlier or later than the appropriate time, respectively. For example, if the target in the particular trial took 2000 ms to move from the starting to stopping position and if the participant responded 1500 ms after seeing the target in the start position. The response will produce a temporal error of + 450 ms corresponding to the difference between the appropriate time for launching the projectile (1050 ms) and subject response time (1500 ms). Temporal errors were collected with an accuracy of 1 ms. . ______________________________________________ Figure 1 may be inserted here ______________________________________________

Procedure. Participants were tested individually and were assigned randomly to one of the three experimental conditions. In the Explicit-Informative and Explicit-rule discover instruction conditions, participants were informed that there were different durations and that the order of these durations was governed by a regular pattern which was predictive of the suitable time to launch the projectile. They were also encouraged to detect such pattern and to use it order to enhance their hitting performance. We may note, however, that only participants in the Explicit-Informative Instruction Condition were allowed, before carrying out the task, to observe and examine attentively

The text body 5/13

the regular pattern. Finally, and in comparison to the two preceding conditions, only the participants in the Implicit Instruction Condition group were informed that their task was to hit the target, or at least to place the projectile as near as possible to the target. No other information was provided about the target behaviour. In the three conditions, participants carried out 10 blocks of 45 trials (see Figure 2). Target behaviour in blocks 1 to 9 was controlled by a 9-element sequence. To illustrate; if we assign digits ascending from 1 to 3 to the three possible durations, the sequence may be presented as follows: 3/3/1/2/2/3/1/1/2. Three different versions of the sequence were used. Block 10 was a test block. In this block, participants were faced with a new sequence of nine elements, which may be presented as follows: 3/3/2/1/1/3/2/2/1. Overall, the experiment lasted about 45 min. All participants immediately received a post-experimental interview about their impressions and knowledge of the regular features of the task. Results In the following analyses, the absolute value of Temporal Error will be used. Preliminary analyses did not indicate any significant trend to produce negative or positive Temporal errors as a function of the manipulated factors. The absolute value of temporal errors were then averaged among durations and calculated for each participant and for each block of 45 trials. For simplicity, we will use the term Temporal Errors instead of the Absolute value of Temporal Error in the following analyses. The means of Temporal errors obtained in each condition are presented in Figure 2. Learning was assessed by examining (1) the effect of sequence practice (Blocks 1-9) and (2) the effect of introducing a new sequence in Test block (block 10). The temporal errors obtained in the practice blocks were entered into a two-way analysis of variance (ANOVA) with Instruction type (Informative, rule discover, and implicit) as a between participants factor, and Training Blocks (blocks 1-9) as a within participant factor. The analysis yielded that temporal errors varied significantly as a function of Instruction type F(2, 39)= 11.51, p
Lihat lebih banyak...

Comentarios

Copyright © 2017 DATOSPDF Inc.