Education Revision Behaviourism

Concepts to consider

• learning
• motivation
• disposition
• capability
• performance

Different types of learning

1. Motor -- Physical skills and co-ordination
2. Affective -- emotional learning
3. Cognitive -- information and ideas

Two major theories

Behaviourism
Cognitivism

Classical conditioning

UCS means unconditioned stimulus
UCR means unconditioned response
CS means conditioned stimulus
CR means conditioned response

Food (UCS) --------------------Salivation (UCR)

Food (UCS) --------------------Salivation (UCR)
Bell (CS)

Bell (CS)-------------------------Salivation(CR)
Behaviourism involves the consideration of Stimuli Responses
These components are used to form the rules of conditioning

Metacognition -- awareness of cognitive strategies
An example of a cognitive theorist is Piaget (remember Samuel and Bryant? --of course you do!)
The third approach! --Humanism -- human individuality and uniqueness
Check out table 4.1 in Lefrançois for differences between Behaviourism, Cognitivism and Humanism.

Classical conditioning

Watson liked Precision, Rigor and objectivity (don't we all?)
He rejects terms such as mind feeling and sensation
Classical conditioning can explain children's dislike of school.

Classrooms should be attractive, and learning should be fun. Unattractive, cold schools reinforce school phobia. Difficult subject material should be presented in small steps. Bad habits can be eradicated by the child pinging an elastic band against their wrist every time they perform a bad habit (or even think about it). Extra school work should not be given as a punishment.

Teachers who make their students laugh teaches three things

1. To laugh
2. the subject
3. to like all things connected with the learning situation

John B.Watson (1878 - 1958)

- conditioned 11 month old Albert to be frightened of furry objects and animals. A loud noise (UCS) was paired with a furry animal (CS). The noise will naturally frighten the child (UCR), but after a few pairings the furry animal on it's own would produce the conditioned response (fear).

Watson found that children's fears could be removed by classical conditioning. He used three techniques:

1. Emotional flooding
2. Counterconditioning
3. Systematic Desensitisation

Emotional flooding is when the child is exposed to the feared object in a safe environment.
Counterconditioning is when a desirable reward is given whilst the feared object is present.
Systematic desensitisation is when the child is at first presented with a mild version of the feared object (perhaps a photograph from a distance). Once the child is used to this, a stronger version of the feared object is presented. Eventually the child would be in contact with the actual feared object.

Watson believed that people were born with three basic emotions:

1. Fear
2. Rage
3. Love

Watson thought that learning was achieved by trial and error, with unsuccessful responses being lost, and successful responses becoming more frequent. Teachers should use problem solving in order to allow their pupils to experience frequent and recent success. The following scientific approach is recommended:

1. Define the problem (For example, design a bridge, using paper and glue, that can take a specified weight)
2. Hypothesise possible solutions to the problem (The child might try 'blind alleys', for example, gluing single sheets together)
3. Conduct controlled trials (try the sheets)
4. Collect and analyse the data (the paper breaks)
5. Form conclusions, and eventually, perhaps after a few attempts, solve the problem (paper needs support. Roll up paper to make tubes.  Go back to hypothesising new solutions.)

Can You think of lesson that could use this technique?  What are the advantages and disadvantages?

Edwin R. Guthrie

Guthrie believed in the contiguity explanation of learning.
This means that if two events appear close together, in time or space, they will become associated with each other.

Applying Guthrie's idea, a student who chews gum whilst studying will associate the flavour of the gum with the material. In an exam if the student chews the same flavoured gum then the associated facts should be more easily remembered.

Guthrie believed that the CS (gum) needs only to be paired with the CR (the to be remembered facts), the UCS is not needed (In fact, there is not an UCS that leads to an UCR of recalling facts). Only one trial was necessary to form an association.

Guthrie did not believe that reward or punishment was necessary to change behaviour. All that is required is success in one of the following:

1. problem-solving The feeling of elation after solving a problem
2. task-oriented (helpful) behaviours Doing something that would prove to be useful, such as studying for an exam.
3. defence-oriented behaviours Activities that stop you getting into trouble or physical danger.

Guthrie proposed three methods for breaking bad habits:

1. Threshold
2. Fatigue
3. Incompatible Stimuli

Threshold entails getting the subject used to part of the new habit (perhaps one page of a novel each night). After awhile more reading could be undertaken, until, eventually, the subject is reading several chapters a day.
Fatigue would involve getting the subject to repeat the bad habit over and over, until they get sick of it. For example, a student who spent time staring out of the window, could be forced to spend a prolonged period of staring out of the window.
Incompatible Stimuli involves pairing an undesired behaviour with a more pleasant behaviour. For example students who are frightened of tests could spend time undergoing relaxation exercises whilst contemplating a forthcoming exam. A student who has trouble reading a text book in noisy surroundings could try reading an interesting novel for a while. The pleasure of reading the novel would help to forge the link between reading and noise being in the background. Reading the text book in a noisy environment is easier after this, because reading in a noisy environment is associated with pleasure, whereas listening to the noise has not been associated.

Edward L. Thorndike (1874 - 1949)

--

Connectionism

Stamping in S-R bonds = learning
Law of effect - response that occur just before a satisfying state of affairs tend to be learnt (stamped in).
Responses that occur before an annoying event, tend to be stamped out.
Law of effect - repetition or practice will strengthen the Stimulus - Response bonds.   Lack of practice would result in the S-R bonds being weakened.
Law of readiness -- learner has to be ready for learning (e.g. maturation and previous learning)
Rewards are more effective than punishments.

Punishment does not work because:

1. It teaches the child that they should be cleverer in not getting caught in future
2. It creates anger
3. It destroys the relationship between teacher and pupil

Thorndike's five principles or subsidiary laws of learning

1. Principle of multiple responses
2. Principle of set
3. Principle of Selective Response
4. Principle of Response by Analogy
5. Associative shifting principle

Principle of multiple responses -- many different responses are tried until success - trial and error. In Piagetian terms, failure to achieve a result means that accommodation to the new learning situation is required. Success leads to the assimilation of whatever brought about the success into the schema. Trial and error --let students emit a variety of responses

Principle of Set -- Another way of stating the law of readiness. Teachers need to create conditions that will make their student's enthusiastic and motivated to learn. Teachers will also need to bear in mind how much their students already know about the subject to be taught. Relate information to child's cultural background and immediate environment
Cultural background and immediate environment determines what will be satisfying or annoying.

Principle of Selective response (Law of prepotency of elements)-- The student learns to pay attention to important (salient) features of the situation, in order to conduct a task (e.g. at roundabouts, only paying attention to traffic coming from the right [in Britain, left in most countries]). people respond to the most significant or the most striking aspects of a stimulus, not necessarily to the entire situation.
This means underlining important points, use of colour, use of voice and gestures, repetition.

Principle of response by analogy -- Generalisation -- teach one specific thing, but it should lead to the learnt response being given to a similar situation. Transferring knowledge from one situation to another. Piaget would say 'assimilation'.

Associative shifting principle -- Students learn that a particular lesson is enjoyable, because it has been paired with enjoyable experiences. History is fun because we watch videos, go to interesting places, have interesting objects to look at, and the teacher tells some entertaining stories. Eventually, just reading history from a text book is satisfying.

B F Skinner

Skinner box --rewards given for response

Primary reinforcer
Food Drink Sex

Generalised reinforcer
Prestige Money Success
Both Primary and generalised reinforcers can be
Positive or Negative
Positive reinforcement (Reward)
teachers smile, praise, high grade, etc.

Reinforcers may be extrinsic or intrinsic. Extrinsic is a reward given by another person, whereas intrinsic comes from within the person (e.g. satisfaction).

Reinforcers may be social or material. Social reinforcers would include praise, whereas material reinforcers are concrete items such as sweets.

Negative Reinforcement

Taking a "bad thing" away
e.g. Letting the students off their homework, because they have worked hard recently

Punishment

Inflicting a bad thing (punishment) because the pupil did the wrong thing
Or loss of something good -- detention

Punishment does not:

illustrate desirable behaviours causes undesirable emotional side effects only suppresses undesirable response

Schedule of reinforcement

Intermittent reinforcement can be ratio or interval
Ratio can be fixed or variable
Interval can be fixed or variable

superstitious schedule -
fixed interval schedule without the animal having to make a correct response

Conditioning measurements

Rate of learning
Response rate
extinction rate - time taken for behaviour to die away

Continuous reinforcement
important for the initial stages of learning
Initial learning in the classroom requires more reinforcement than later learning
Rapid learning but rapid extinction

Random ratio -- slowest rate of extinction --information remembered better

Marquis (1941) babies on fixed interval schedule feeding, showed an interest in activity just before the interval was up.

Bandura and Walters (1963) - children seeking attention from parents tend to be randomly reinforced.

Shaping -reinforcement given for behaviour remotely similar to desired behaviour.
Closer approximations are necessary for future reinforcement.
Eventually only desired behaviour reinforced.

generalisation - applying response to similar situations

discrimination - only responding to certain appropriate stimuli.

Behaviour modification - the application of behaviouristic principles in education and therapy

4 points by Thorndike Applying Behaviourism to learning in schools

• 1 Punishment is not very effective for eliminating undesirable behaviour
• 2 Interest in work and in improvement is conducive to learning
• 3 Significance of subject matter and the attitude of the learner are important variables in school.
• 4 Repetition without reinforcement does not enhance learning

Social learning

Learning through social interaction, socially acceptable behaviours.
SOCIALISATION - The fostering of appropriate behaviours. An important part of bringing up a child.

Observational learning - learning through imitation --Bandura 1969

Imitation is reinforced by social approval (If not from teachers or parents, from peers (as in the case of copying 'anti-social' behaviour, eg swearing)
This conforms to operant conditioning and is known as direct reinforcement.

Model
Person whose behaviour serves as a stimulus for an observer's response

Can you write down a list of people who might serve as models for children, and say what behaviour you think would be modelled?

Symbolic models

• oral
• written instructions,
• pictures,
• cartoon or film characters
• religious figures
• content and characters in books and films

Exemplary models
- well behaved people are held up to be this.
Types of reinforcement in imitation

• direct reinforcement - praise like "he's doing it just like daddy"
• Consequences of behaviour - learning to say "milk" to get a glass of milk
• Vicarious reinforcement - watching others getting rewarded leads to behaviour.

Effects of imitation

modelling effect
Acquiring new behaviour as a result of observing a model.

Inhibitory - Disinhibitory effect
Ceasing or starting deviant behaviour as a result of seeing a model punished or rewarded for similar behaviour.
Eliciting effect - Engaging in behaviour related to that of a model.Acting tough , but not in the same way as a famous actor
Televised violence, that is true to life, may increase violence, but this depends on child's personality (Collins 1983) A quiet passive child would be less affected by violence
Prosocial TV programs can have a positive effect on behaviour (Cooke et al, 1975) eg Blue Peter, raising money for charity
Disinhibitory effect
A behaviour normally suppressed is allowed to become manifest as a result of seeing an adult rewarded for similar (violent) behaviour. (Bandura and Walters 1963) If others are seen to get away with it,then one is more likely to break the law (eg by parking on double yellow lines).

Facilitation refers to the model being rewarded allowing the observer to gain confidence to perform an activity (e.g. model successfully jumps off a diving board).

The offer of rewards for aggressive behaviour cancel out the effects of inhibition. (Bandura 1962).

---

Walters, Llewellyn & Acker, 1962 and Walters & Llewellyn, 1963

Group 1
Watch scene from `rebel Without a Cause'
2 youths in a knife fight

Group 2
Film with adolescents engaged in art work.
Both groups then give shocks of varying intensities to confederate in `learning' experiment (like Milgram)

Result

Group 1 gives more shocks at higher intensities than group 2

---

Social learning saves learning by trial and error.
Would you learn to drive a car this way?   No thanks, modelling is safer!

Consider these factors

Many attentional processes are under the teacher's control.
e.g. distinctiveness of stimuli, arousal level of learner, etc.

Modelled events are imitated , depending upon the following three factors:

1. Attentional Processes
2. Retentional Processes
3. Motor Reproduction Processes

Note

1. Attentional processes depend upon paying attention to salient events, distinctiveness, complexity and prevalence. Also observer characteristics such as sensory capacities, Arousal level, perceptual set and past reinforcement.
2. Retentional Processes depend upon the processing of visual and verbal codes.
   Cognitive rehearsal and imagined actions are found to improve performance
3. Motor reproduction processes are limited by the physical capabilities of the subject, and their ability to monitor attempted reproductions and to use motor feedback to make corrections
   Motor reproduction processes also depend upon motivational processes.

Evaluation of operant conditioning

· Built upon laboratory experiments with animals

· Mechanistic view of humanity – robotic slaves to the consequences of their environment

· Tangible and observable rewards not always necessary for learning to take place

· Learning theory does not take into account hidden, unobservable cognitive and emotional factors

· But sometimes it is not possible to ask people directly and observing behaviour then becomes a more reliable measure of learnt behaviour

· Aversive methods (punishment) not always effective as it tells a child what he should not do not what he should do.

Applications of behaviourism

• Stickers, stars, merit points
• Detention, standing outside the head’s office, suspension and exclusion, reprimands, unpleasant activities, withdrawal of benefits, corporal punishment (now against the law in the UK).
• Analysing disruptive behaviour
• Antecedents
• Behaviour
• Consequences
• E.g. child is ignored, starts disruptive behaviour, and is rewarded by attention from class and teacher. Child needs to be rewarded for on-task behaviour by being given attention
• Shaping involves praising a child for displaying a component of the required behaviour (e.g. sitting at desk for a short while), then demanding behaviour that is closer to the desired outcome before giving praise (e.g. sitting longer at the desk, working for a short while, working longer)
• Whole class teaching does not suit slow or fast learners, just the average paced learners. Better to have teaching machines. Machines deal with the basics and allow the teacher to have the time to give meaningful exchanges. Student works at own pace.
• Teachers must be consistent. Moody teachers can cause students to be sullen and passive in lessons, frightened to contribute in case they get into trouble. Haney, Banks and Zimbardo (1973) report mock guards using arbitrary control on the mock prisoners (e.g. punishing a prisoner for smiling at a joke and later punishing a prisoner for not smiling at a joke).
• Observational learning requires attention to be paid to the model (teacher) and for the student to use mental rehearsal in order to remember the complex actions. This is followed by performance and practice.

Evaluation

• Analysis of a child’s behaviour takes time and involves careful observation and record keeping.
• Teacher cannot give constant reinforcement; there are too many other students.
• Initially teaching machines were crude and cumbersome, but now programmes can be run on personal computers.
• Teachers do not have control over the most powerful reinforcers that affect student behaviour.
  E.g. peer acceptance and praise, parental approval, etc.
• Teachers do not reward specific responses, or do not set about bringing such responses.
  Teachers do not wait for a response to reinforce, as in the Skinner box.
• Palardy 1991, behaviour modification techniques ignore the causes of behaviour.
• Once extrinsic rewards stop the behaviour is likely to stop.
• Social rewards are more effective than material rewards. Material rewards are effective because they are also social rewards (e.g. as in a token economy).
• Negative reinforcement requires there to be a disliked activity in place in order for the child to be excused from it. The disliked activity might have negative effects upon the child, just as in the case of punishment.
• As the teacher decides what is rewarded or punished, behaviourists believe that teacher always knows best. The teacher justifies his or her actions as being in the interests of the child: it is an unequal relationship between teacher and pupil; this contrasts with the Humanist approach.

Task

Using learning theory, design a computer programme that teaches multiplication to 7 year olds and involves the use of a game where two armies battle it out.

Key Studies

Cognitive study

In one study, Wood, Bruner and Ross (1976) attempted to teach 3 and 4-year-old children to assemble a complicated block pyramid. It was felt that it was not until the age of 7 that a child would be in a state of maturational readiness and would be able to do this task without assistance. The children were instructed by their mothers in how to do the task. Some children benefited from instruction while others did not. So can language instruction enhance or accelerate cognitive development? It would seem that the issue is not whether the use of language enhances cognitive skills, but what factors regarding the language of communication are responsible for cognitive development. Wood, Bruner and Ross (1976) found the following:

Techniques that did not work:

• Strategies that had tutors showing the child what to do first (i.e. ‘Now watch what I do. Now you try it’) did not work. The authors speculated that this approach overloaded the child’s powers of concentration.

• Strategies that relied on verbal instructions (i.e. ‘Put the big one there, and the small one there’) did not work. Again the authors speculated that children did not understand the commands without the commands being acted out.

Techniques that did work

• Contingent instruction — that is, specific instructions geared to the child’s perceived need — seemed to be the most effective. This involved two main rules: when struggling, offer more help; and when succeeding, withdraw help.

Evaluation

1. Control – Different mothers, different expertise in explaining, different relationships

2. Usefulness – Teachers can use the results to improve teaching and learning

3. Ecological Validity – Own mothers and engaged in a typical child-mother activity (Good compromise between complete ecological validity {e.g. doing whatever naturally happens}) and over-control (Pyramid building seems to be a typical activity for parents and children)

Behaviourist KEY STUDY

Influence of models' reinforcement contingencies on the acquisition of

imitative responses

Bandura (1965)

Aim: To show that children will imitate a model to a greater extent if the model is rewarded for his (or her) behaviour than if he is punished or if there are no consequences.

Sample: 33 boys and 33 girls from Stanford University Nursery School.

Method: The children were individually shown a television programme that depicted a model attacking a Bobo doll for about 5 minutes. (A

Bobo doll is an inflatable figure with a heavy base; when knocked over, it swings back into an upright position.) For one-third of the children, the programme ended with the model being rewarded by another adult who came into the room, gave the model sweets and soft drinks, and praised him for being a 'strong champion'. The second group of children saw an ending in which the second adult comes in and shakes his finger at the model, saying: 'Hey there, you big bully. You quit picking on that clown. I won't tolerate it.' He then spanks the model with a rolled-up magazine. The third group saw the model beating up the Bobo doll, but with no consequences at all; no other adult came into the room. Immediately after watching the film, the children were put into another room containing a Bobo doll and observed by two judges through a one-way mirror.

Results: At first, children in the model-punished condition showed significantly less aggressive imitation than children in the model rewarded and the no consequences groups. When the children were then offered a reward of fruit juice and colouring books for reproducing the acts of the model, they all showed high levels of imitation and there was no difference between the three conditions.

Conclusions: As the children in all three conditions could be induced to imitate the aggressive acts by being offered reinforcers, Bandura concluded that they had all learned the behaviour by observing the model. However, prior to being offered these inducements, the fact that the children in the model-punished condition showed less imitation supports the notion of vicarious reinforcement, that is, children are more likely to imitate behaviour if they see the model being rewarded and less likely if they see the model being punished.

This demonstrates that reinforcements do not have to be given directly to a P, watching somebody else being reinforced can be enough (vicarious reinforcement).

1. Observed through one-way mirror reduces demand characteristics.
2. Well controlled e.g. same film
3. Ethics – teaching aggression.
4. Lacks ecological validity may not generalise.
5. Sample – middle class kids so this makes the results stronger.

Behaviourist

Higgins, J. W., Williams, R. L., McLaughlin, T. F., 2001, 'The effects of a token economy employing instructional consequences for a third-grade student with learning disabilities: a data-based case study', Education & Treatment of Children, 24, 1, 99-105

To investigate the effectiveness of using a token economy utilizing academic rewards on the behaviour of a student with learning disabilities.

A case study involving an experiment with a repeated measures design.

A ten-year-old boy from a single-parent family, who had been diagnosed as having learning difficulties. His IQ score was within the normal range but he was below age level in reading and writing and exhibited high rates of multiple disruptive behaviours.

Three disruptive behaviours were targeted for intervention in this study inappropriate talking, poor seat posture and out-of-seat behaviour, as they had been selected by his teacher and classroom assistant as the most problematical. The study was conducted in the boy's usual classroom, in which there were a further 19 students.

Baseline measures of the three disruptive behaviours were taken at the outset via observations for a number of 20-minute periods over 15 days. The frequency of occurrence of each behaviour was recorded at the end of every minute of the 20minute period. To allow for ease of observation and to minimize disruption to the rest of the class, the boy was moved to a position on the fringe of the class. The observer was sat approximately 2 feet behind and to the right of the participant.

In the experimental condition, the participant was immediately given a tick in a box if he had not displayed any of the target disruptive behaviours in the previous minute. A grid, into which the ticks were placed, was taped to the participant's desk,

thus enabling him to gain instant feedback. At the end of each session the ticks were counted up and divided by two. The figure then arrived at equated to the number of minutes available for the use of secondary reinforcements such as maths worksheets, computer time, one-to-one reading instruction, leisure reading and playing work

related games. The participant was allowed to engage in these in the first ten minutes of the following day, which was usually taken up with registration and other administrative matters, to ensure that he did not miss out on any teaching. This period of the study lasted for 12 days.

T en to 12 days after the end of the use of the token economy, the boy was

observed again to see whether or not the intervention had any lasting effects.

To ensure that the observer was recording the correct behaviours, inter-rater

reliability testing was conducted, with the participant's teacher also recording the

disruptive behaviours twice each during the baseline assessment and four times each during the token economy phase. Reliability was established to be 100 per cent.

The results of the study are shown in the table below:

	Inappropriate talking	Out-of-seat behaviour Poor seat posture
Baseline	6.0	1.9	11.0
Token economy	0.8	0.2	5.0
Maintenance	0	0	2.5

The results show that the use of a token economy reduced the occurrence of disruptive behaviours and, moreover, that this reduction was maintained after the withdrawal of the token economy.

Behaviourist

Self-Brown, S. R., Mathews, S., 11,2003, 'Effects of classroom structure on student achievement goal orientation', The Journal of Educational Research, 97, 2, 106-13

To investigate the effects of different classroom assessment structures on student achievement goal orientation. It was hypothesized that the use of a token economy (giving tokens as rewards that can be cashed in for prizes) would lead to the setting of performance goals (high standard of work), the use of contingency contracting (get a reward for achieving a specified goal) would lead to the setting of learning goals (high standard achieved at the end of the module) and not using either of these strategies would not differentiate the type of goals being set.

An experiment.

71 students from three different classes in a single primary school. The classes were randomly allocated to the three conditions. A fifth-year class of 25 students was allocated to the token economy condition, a fourth-year class of 18 students was allocated to the contingency contract condition and a class of 28 fifth-year students formed the control condition.

Students in the token economy group were given a contract which explained how and at what rate tokens could be earned (e.g. four play dollars for each A or B in maths) and the secondary reinforcers (sweets, computer time, pens, key rings) that the tokens could be exchanged for. This contract was in a folder that was kept on the students' desks. Also in the folder was a goals chart. This chart contained a list of student behaviours that could earn tokens and the number of tokens each behaviour would be rewarded with, as well as a column in which the students could record their weekly and long-term goals for maths.

Those in the contingency contract condition received a contingency contract which described how they would meet weekly with the researcher to set and discuss goals for maths. They, too, were given a folder with the contract in it to keep on their desks. The folder also contained a goals chart on which they could write their weekly and long-term maths goals. A gold star was placed next to a goal when it had been achieved.

Students in the control condition received only a goals chart on which to record their weekly and long-term maths goals.

For all students the setting of goals was done in a weekly meeting with the researcher. For the token economy and contingency contracting condition this involved discussion about these goals. In addition, it was during this meeting that the students in the token economy condition could exchange their tokens for the secondary reinforcers. For students in the control condition, no discussion, feedback or any information about their goals was provided; they simply used the time with the researcher to write their goals on their own.

The experiment was conducted over a five-week period and was in place for all maths lessons (which were conducted as they had been prior to the start of the experiment) during that time.

The students' goals were classified as either performance or learning goals by the researcher on the basis of a goal typology developed by Dweck.

The results of the study are shown in the table below:

Goal type	Token economy	Contingency contract	Control
Learning goals	0.75 goals set	14.27	5.36
Performance goals	4.95	5.55	5.61

The results uphold the experimental hypotheses that a token economy would lead to the setting of more performance than learning goals, contingency contracting would lead to the setting of more learning than performance goals and providing neither of these would not differentiate between the types of goals set.

Lack of Control – Contingency contract sample is one year younger and they might be more likely to be motivated by long-term goals.

Experimenter bias – Discussion with experimenter meant they could be influenced. Classification not done by a ‘blind’ independent experimenter.

Validity of measure – should have used goals achieved or ability in maths rather than goals set.

Demand characteristics – control group – not suggested to them that they should set goals and other groups could be responding to what they think they are supposed to do.

Cognitive

Aleven, V., Koedinger, K. R., 2002, 'An effective metacognitive strategy (Metacognitive means understanding one’s own thought processes): learning by doing and explaining with a computer-based Cognitive Tutor', Cognitive Science, 26,2, 147-79

To investigate whether self-explanation can be scaffolded effectively in a classroom through the use of a Cognitive Tutor, a piece of intelligent instructional software that supports guided learning by doing and explanation. While this paper reports two experiments, only one will be summarized here.

Experiment with a matched pairs design.

41 students of 15-16 years, from a suburban high school near Pittsburgh, P A, USA, who were studying a geometry unit on angles via one hour of teaching per week, with a second hour spent solving problems using geometry Cognitive Tutor software on a computer. Data on only 24 participants are used for analysis as some failed to complete the course in the allotted time and the teacher forgot to administer a post-experimental test to measure outcomes for others.

Before the experiment commenced the students were matched on their prior achievement on this course and this was used to allocate them to conditions. Before the start of the angles unit they were given a pre-test to assess their knowledge and understanding of angles. This acted as the baseline measurement for the effects of the two conditions.

The students were then required to work through the same geometrical problems on a computer over the course of the unit, which ran for a semester.

The two conditions in this experiment were 'the explanation condition and the problem-solving condition. The students in the explanation condition used a version of the Cognitive Tutor that required them to input an explanation for each step in their problem-solving process. Feedback is then given on both their solutions to the problems and their explanations. The students in the problem-solving condition (effectively the control condition) used a version of the software that did not require this input of explanations. The unit was deemed to be finished when the students met the Cognitive Tutor's predetermined criterion for mastery.

At the end of the unit a post-test to measure their knowledge and understanding of the unit content was given to the students. The differences between the scores on the pre- and post-tests were used as a measure of how much learning had been acquired by the students in each condition.

First, it was found that the explanation condition spent more, but not significantly more, time working on the computer than the problem-solving condition, but this was to be expected as they had an additional task (inputting their explanations) to do.

In terms of the differences between their pre- and post-test scores, the students in the explanation condition exhibited a significantly greater improvement than those in the problem-solving condition. The former also needed fewer problems to reach the Cognitive Tutor's mastery criterion than did the latter (102 versus 136 respectively).

Control – should have had the control group typing in something else to control for time.

Sample – huge attrition of sample could lead to biased results. May not generalise to students who give up easily.

Usefulness – teachers can use computer programs that feed back on strategies used, but may only be useful for motivated students as many gave up in the study.

Generalisation – may not apply to other subject areas such as those that employ divergent thinking rather than convergent.

Acknowledgements

Sally Gadson, Philippe Harari, Karen Legge and Linda Sherry, (2005), Heinemann Psychology A2 for OCR. ISBN 0-435-80671-8

Fiona Lintern, Merv Stapleton & Lynne Williams (2004) Study Guide for OCR Psychology: A2 Level, Hodder & Stoughton. ISBN 0-340-81626-0.

Lecture notes

Return to home page