A Study to Show the Effect of Conformity on Estimating the Number of Sweets in a Jar
A Study to Show the Effect of Conformity on Estimating
the Number of Sweets in a Jar
Abstract:
The research was designed to test the claim by Jenness that conformity levels increase in an ambiguous situation. The method was a Laboratory experiment and the design was an independent measures design and the participants were an opportunity sample of fourteen Sixth Formers (16-17 years old). Participants in the control condition had to individually guess the number of sweets in a jar. The average score was relayed to the group and they were then, as the experimental condition, asked to re-guess the number. The one tailed hypothesis was that estimates in the experimental condition would produce estimates closer to the group estimate.
The findings were that conformity exists in an ambiguous situation. The data was put to the paired t test because the data was interval and from repeated groups. The results were significant at the 0.05 level. One can conclude that ambiguity induces conformity and so the one tailed hypothesis was accepted.
Introduction:
Much work has been undertaken in the field of conformity. Psychologists such as Asch, Sheriff, Crutchfield and Jenness have all conducted valuable studies which can be used as a basis for conformity research.
Asch investigated if people yield to group pressure when the answer is obvious. He gave participants a simple perceptual task of matching one line with another. There was the sample line and three possible matches. It was very obvious which line matched the sample. He then set up the situation with confederates who on some occasions argued a blatantly wrong answer. 32% conformed with the confederates. Asch also found that the level of conformity was higher when the confederates weren't too rigid. He therefore declared that a more fluid influence is more likely to induce conformity. Ref1
Sheriff used a visual illusion, (the autokinetic effect) where a spot seen in an otherwise dark room appears to move when in fact, it is stationary. He then told subjects to state the direction and speed of the "moving light". Individual estimates varied greatly as it is individual perception. However, when the participant became a part of a group, a group norm developed. This is a far less strict test in comparison to Asch. It also was more efficient than Asch as only a few results were collected in each of Asch's experiments due to the replication of majority influence dictating the small ratio of participants: confederates.
Sheriff claimed that he had shown conformity. The individuals were experiencing informational social influence.
Ref2
Crutchfield's participants were asked a series of questions and shown an array of alternative answers as well as lights apparently indicating answers of the other subjects. As all participants were in individual 'booths', they could not discuss their answer. In fact, the answers were wrong on approximately half the trials. He found that invariably group answers were more extreme than individuals.
All of the above studies suggest that conformity exists due to informational social influence. This is when an individual looks to others to make decisions about how to behave. Someone may yield to group pressure because others are thought to possess more knowledge. Sometimes this conformity may be to experts, to the influence or image of an idol or in the cases of the psychological studies, to the way others or a group are behaving.
Jenness showed participants a jar filled with a large number of beans and asked the participants to estimate the number present in the jar when in fact it was very difficult. A group estimate was calculated from the individual answers and relayed to the group. This was then discussed. The participant's were then asked to re-view the jar and amend their answers if they believed it appropriate. Estimates were collected again and a second average was calculated. Jenness found that the experimental estimates converged closer to the group estimate showing that conformity exists in ambiguous situations. Ref3
Aims and Hypotheses
The aim of this experiment was to replicate Jenness and so see if conformity still occurs to that level today. The alternative hypothesis therefore formed was that estimates in the experimental condition would produce estimates closer to the group estimate due to ambiguity. Furthermore, it was expected that a higher level of conformity would occur in an ambiguous situation because of informational social influence. This is a one-tailed hypothesis.
A directional hypothesis was selected because based on background knowledge of Jenness -as a result of conformity- the estimates in the experimental condition will ...
This is a preview of the whole essay
Aims and Hypotheses
The aim of this experiment was to replicate Jenness and so see if conformity still occurs to that level today. The alternative hypothesis therefore formed was that estimates in the experimental condition would produce estimates closer to the group estimate due to ambiguity. Furthermore, it was expected that a higher level of conformity would occur in an ambiguous situation because of informational social influence. This is a one-tailed hypothesis.
A directional hypothesis was selected because based on background knowledge of Jenness -as a result of conformity- the estimates in the experimental condition will converge towards the group estimate.
The null hypothesis was that any relationship between an ambiguous situation and participant's estimates converging towards the group estimate is solely due to chance.
The research method selected was a laboratory experiment with a repeated measures design. A lab experiment is where an independent variable is manipulated while others controlled, to see effects on a dependent variable.
The advantages of a lab experiment are as follows. It produces results with high experimental validity- therefore, it is easy to gain the same results if replicated. This is to do with the fact that all variables can be strictly controlled and therefore it is easier to comment on cause and effect. It enables the use of complex equipment and is often cheaper and less time-consuming than other methods. However, Lab experiments lack ecological validity-this means that it is not true to life as it excludes variables. They are prone to demand characteristics - participants aware of experiment, may change behaviour and due to it being an artificial environment there is low realism.
A repeated measures design is where the same group of people are tested in different conditions; the same people are used repeatedly. Advantages include that they avoid the problem of participant variables and fewer people are needed in the experiment. However, order effects are more likely to occur.
The study also used a single-blind design: Participants did not know which condition (experimental or control) they were in. There is some deception in many psychological studies to stop participants guessing the aims and changing their behaviour.
Method:
The chosen research method for this research was a laboratory experiment. It was conducted in a school classroom where many confounding variables could also be controlled. It used a repeated measures design using an opportunity sample of fourteen 16-17 year olds from school, participating in both the control and experimental conditions, so that each participant had to do exactly the same task.
A laboratory experiment meant that even the room had to be considered. To eliminate confounding variables, a room with adequate lighting and moderate temperature was to be selected. The room was free from loud noise, which could stop concentration.
All participants were firstly shown the jar filled with roughly 1000 sweets. They were briefly shown as a group and then were only allowed to hold the jar for 3 seconds before submitting their individual estimate on a piece of paper. No discussion or talking was allowed.
After collecting the participants' estimates, all figures were written in full view of the entire group and an average calculated which was relayed to the group as a whole.
To test the experimental hypothesis, the group were then allowed to hold the jar again for a further 3 seconds and asked again to write down their estimate.
The results were collected and another average score was calculated and relayed to the group.
The first set of results were the control set and the second set were the experimental set. This was because the first set shows individuality and the second set shows conformity if the independent T-test is positive.
The IV was ambiguity and the DV was the participant's estimates
The group were addressed with a set of standardised instructions (see appendix 1):
To make sure that this was a true experiment, all of the variables must be controlled to gain accurate, unbiased results.
Some examples of variables faced, the problems linked and some solutions are as follows.
The longer the participant looks at the jar, the more accurate their estimates could be. To combat this, each participant was allowed to view the jar for the same set amount of time (3 seconds).
As some people can easily guess numbers, the solution was made to use participants of matched ability, for example those in the same class.
Triviality could cause some participants not to take give true estimates. To combat this, participants were told that the study was for coursework and should be taken seriously.
Due to Demand Characteristics, participants may have tried to help the experiment by acting in the way that they believe the experimenter wants them to. The experimenter could therefore use deception. In this case, participants were told that it was a study on "Spacial Awareness". However, this is an ethical issue which must also be dealt with.
Time of day is also an important factor to consider, as at the end of the day people can get more easily distracted than in the mornings due to fatigue. To reduce the problems linked with this, the experiment would be done at the same time of day if it were to be repeated.
If the participant has background knowledge of the subject, for example studied Jenness, they might know the study so could cause biased results. Therefore, to combat this, Psychology students were not used in this study.
Noise can distract the participants and so cause guesses to be inaccurate, to combat this, all outside noises were kept to a minimum. If the study were to be repeated, the same room would be used and therefore same noise level.
Finally, it must be clear that all participants understand the procedure, as with a lack of understanding, the participant's estimate could be inaccurate. As a result, Standardised instructions were used in this study.
Ethical Issues
By debriefing, the issue of confidentiality, deception and informed consent were addressed and if the participant, decided that they were uncomfortable with the outcome of their taking part, their results were removed. In fact, no one took advantage of this offer. (See appendix 2)
The volunteers were told of their right to withdraw from the experiments at any time. This was to hopefully prevent any psychological harm linked to the experiment
The data then was tested using the paired t test and the significance level tested to the 0.05 level.
Results:
The data was collected by asking the participants to write their answers on a piece of paper. These were then collected and the average was relayed to the group. (See appendix 3)
It was Interval data because it was data with equal units of measurement throughout the scale but without a true zero.
The mean was chosen as the measure of central tendency. This is done by dividing the total of participants scores by the number of participants.
The advantage of using the mean is that it is a sensitive measure of central tendency, however, if there are one or more extreme scores, the mean can be changed dramatically and therefore be misleading in a study. To combat this, often the most extreme scores at each side are neglected and the mean is then calculated.
A repeated measures design was used so that the participants aren't exposed to practice effects, which could make the study biased. However, this can mean that the group are all of the same ability so the study does not have external validity.
The Independent t test was used because it analyses a group of results to determine if a significance difference exists between the means of two sets of data. If a difference is detected, it means that there is either positive or negative conformity. A score of 0 means that there is no conformity whatsoever. The results are as follows:
3 degrees of freedom had a 1.1771 level of significance and T had a value of 1.177.
The data was tested at the 0.05 level (see appendix 4)
This shows that conformity to the majority was not high, but still existed in the experiment this therefore means that the original hypothesis is maintained and the Null Hypothesis is rejected
Another point to mention is that even though the level of conformity to the norm was low, every participant changed his or her guess.
Discussion:
Explanation of Findings
The research was designed to test the claim by Jenness that conformity levels increase in an ambiguous situation. The one tailed hypothesis was that estimates in the experimental condition would produce estimates closer to the group estimate. The null hypothesis was that any relationship between an ambiguous situation and participant's estimates converging towards the group estimate is solely due to chance.
The findings were that conformity does indeed exist in an ambiguous situation and therefore the experimental hypothesis was accepted and the null hypothesis rejected.
The results gained in this experiment are not as conclusive as those found by Jenness, however, factors such as the era in which he performed the study could provide a partial explanation. During the 1930's, people were encouraged to conform, whereas nowadays, individualism is embraced. This could therefore easily affect the participant's behaviour and could not be controlled by the experimenter.
Although the t-test did not prove to be as definite as the findings of Jenness, it is interesting to note that all participants except for one changed their estimate. (See appendix 5)
Relating the study to Jenness
Although a level of conformity existed the study, it was nowhere near the level experienced by Jenness in 1932. He found there was a large shift from the student's original estimate, towards the group estimate on their second individual guess. This study was seemingly quite decisive, but because Jenness instructed the students to make a group estimate, obedience could have been a factor in the students 'new' estimates. The study can also be linked to Crutchfield in that the group estimates were different to individual estimates. This study is linked again to Jenness and Crutchfield in that people are conforming due to Informational Social Influence- this is because as a result of the ambiguity in the situation, participants may feel that the group are more knowledgeable in this area so conform as they themselves have no fixed idea.Ref4
Limitations and Modifications
It could be said that the experiment lacked external validity as it used a group of 16 year olds from privileged backgrounds. It therefore is not representative of all 16 year olds and certainly not of the whole population. As the task was fairly trivial, perhaps the participants did not pay full attention to it. Another problem faced was that it was difficult to control the length of time in which a participant held the jar.
By using the set of standardised instructions before conducting the study, issues such as right to withdraw and informed consent have been eliminated. However, by hiding the true reasoning behind the study, this is deception which could possibly lead to psychological harm, which should be avoided wherever possible. However, it hopefully eliminates demand characteristics. Background knowledge could have produced biased results, therefore Psychology Students were not used as participants to reduce this problem.
Implications and Suggestions for Further Research
If the study was to be repeated, a bigger group of participants could be used to give a more accurate average, and to improve further, different groups of participants could be selected. For example, testing different age groups, e.g, 16 year olds as in the study, middle-aged men and women, and possibly even the elderly. This could also be used to test whether conformity in an ambiguous situation was more common in men or women of each age group. This would create a more detailed picture of conformity.
To rectify the idea that the study did not relate to all of the population, it could be replicated in different situations using different age groups.
To make it less trivial, the jar could be filled with money and it could be turned into a game for example, whoever guesses the closest would win the jar. Finally, during the task, there could have been a fixed set of verbal prompts to prevent the participants from being distracted. This could also help with reiterating that fact that each participant had to hold the jar for 3 seconds.
Another idea for the study could be that it could be tested whether or not conformity to the minority existed in an ambiguous situation by trying to get the participants to guess far higher the second time with the use of verbal prompts from the investigator.
References:
. Asch, S.E. (1951) Effects of group pressure on the modification and distortion of judgements. In H. Guetzkow (Ed.) Groups leadership and men. Pittsburg, PA: Carnegie
Asch, S.E. (1955) Opinions and social pressure. Scientific American, 193, 31-35
Asch, S.E. (1956) Studies of independence and conformity: A minority of one against a unanimous majority. Psychological Monographs, 70 (Whole no. 416)
2. Sheriff, M (1966) Group conflict and co-operation: their social psychology. London: Routledge & Kegan Paul
Sherrif, M (1935). A study of some factors in perception. Archives of Psychology, 27, 187
3. Jenness, A (1932). The role of discussion in changing opinion regarding matter of fact. Journal of Abnormal and Social Psychology, 27, 279-296
Appendix:
Appendix 1
Standardised instructions
Appendix 2
Debriefing
Appendix 3
Raw Data
Participant no
First guess
Second guess
562
431
2
870
400
3
500
853
4
560
543
5
530
874
6
372
570
7
597
401
8
222
420
9
530
390
0
400
400
1
86
716
2
278
530
3
600
365
4
500
450
Average
479
595
Appendix 4
t-test (Paired samples)
t= 1.1771 with 13 degrees of freedom
Control Group Experimental Group
Mean 479.1 595.9
Variance 31613 103827
No of participants 14 14
Appendix 5
A study to show the effect of conformity on estimating the number of sweets in a jar
Page1
By Rebecca Ingram