Investigating this experiment would be interesting if the same trends were evident today using sixth form students, because sixth form students are educated people, but arguably most subject to pressure of not wanting to stand out amongst a group. This experiment will examine whether people conform to other peoples estimations by asking how many ink dots they think are on a piece of paper, with condition A being with the knowledge of other peoples false predictions which they will believe are true predictions, and condition B being with no knowledge of any predictions at all.
The aim of the experiment is to investigate the extent to which conformity occurs in the presence of previous estimations.
Alternative hypothesis: There will be a greater level of conformity in terms of participants estimates in the presence of stooges estimates.
Null hypothesis: There will be no significant difference between conditions A and B, any difference will be due to chance alone.
Method
Design
The design of the study was a laboratory experimental research method using a independent measures design, however the participants were matched as much as possible. Independent measures were used to avoid order effect and demand characteristics. The independent variable (IV) was the presence of stooges estimates in condition A and the dependent variable (DV) was the estimations of the participants.
Before conducting the study, a pilot study was conducted to eliminate any problems that may have occurred; fortunately there were none.
3417 dots were chosen because it is a high number and therefore provides an ambiguous situation because the correct answer is not clear and research shows that participants are more likely to conform in ambiguous situations.
Ethics
To include participants in the investigation, ethical considerations had to be made. Participants were briefed before the experiment, which told them what it entailed and what they were required to do. The briefing usually tells participants the aim of the experiment in full, however for the cause of the experiment, slight deception was needed, so the total aim could not be disclosed. This is because half of the participants had to be led to believe that the stooges estimates put before them were real. This emphasises how essential de-briefing is as an ethical consideration so participants can be told the absolute full nature of the study. Participants were also made aware that they did not have to take part in the experiment and that they could leave at any time.
Participants
The 24 participants of the experiments were independent measures. This was done to avoid order effect. The sample used was a stratified sampling method. Attempts were made to match participants as much as possible but not enough for them to be called matched pairs. All participants were 17 years old, attended the same 6th form, had similar academic abilities and in mathematics according to their AS level exam results they were of extremely similar abilities, and they were matched accordingly with the same sex. No participants studied psychology. This was a precaution took to avoid participants guessing the nature of the experiment.
Material
The materials used were 1 piece of paper with 3417 dots drawn on, another piece of paper with stooges estimates on and a table in which participants estimates could be recorded. (refer to appendix)
Procedure
First of all the procedure involved the experimenter, briefing both sets of participants. Both groups needed to be briefed separately (refer to appendix) as 1 group needed slightly different information as they were to be deceived slightly, as this was necessary to the investigation. The first group of participants were informed that they were to look at the piece of paper given to them for 15 seconds only and then they were to give their best estimates of the number of dots that they thought was on the page. Participants were also made aware that the numbers on the other piece of paper given to them was previous estimations of other students. Students were to look at the paper individually. This was ‘condition A, in the presence of false estimates’. False estimates were close to actual answer. In ‘condition B, no guidance from previous estimations’, participants were not given stooges estimations. When the experiment was finished all participants were debriefed together.
Briefing
This psychological experiment is part of my year 13 coursework.
The aim of this experiment is to measure social influence and I would like you to take part (as my participants).
When called I would like you to enter the room and observe the piece of paper given to you with dots on for 15 seconds only. The other piece of paper has previous students estimations on it. Then you are to write down your estimate of the number of dots you think are on the paper on a separate piece of paper provided. You are asked not to discuss the experiment with other participants, in particular your estimation. You do not have to take part if you do not wish to and you are allowed to leave the experiment at any time. Your information will not be used for any other purpose and all information is confidential.
Debriefing
I would like to take this opportunity to thank you all for taking part in this psychological study.
The absolute full nature of the experiment was to measure conformity levels of people in the presence of previous estimates. For the cause of the experiment you had to be deceived slightly. People who were in ‘condition A’ were given another piece of paper that had stooges estimations on. This was to see if you conformed to these estimates. The paper actually had 3417 dots on.
All your information given to us is totally confidential and will not be used for any other cause.
Appendix
Contingency table
Chi Squared Test
Calculation can be represented by the formulae
x² =∑ ( 0-E )²
E
Expected frequency (E) = RT x CT ÷ GT
Cell 1 = 12 x 8 ÷ 24 = 4
Cell 2 = 12 x 6 ÷ 24 = 3
Cell 3 = 12 x 10 ÷ 24 = 5
Cell 4 = 12 x 8 ÷ 24 = 4
Cell 5 = 12 x 6 ÷ 24 = 3
Cell 6 = 12 x 10 ÷ 24 = 5
Observed frequency (O) - Expected frequency (E) ² ÷ expected
Cell 1 = (6 – 4) ² ÷ 4 = 1
Cell 2 = (5 – 3) ² ÷ 3 = 1.33
Cell 3 = (1 - 5) ² ÷ 5 = 3.2
Cell 4 = (2 - 4) ² ÷ 4 = 1
Cell 5 = (1 – 3)² ÷ 3 = 1.33
Cell 6 = (9 – 5)² ÷ 5 = 3.2
11.06
Degrees of freedom (df) = (2 – 1) (3 – 1) = 2
In this case the minimum level of significance accepted was 0.05.
The observed value of 11.06 is greater than the critical value of significance of 2.71 at 2 degrees of freedom, therefore the null hypothesis can be rejected and the alternative accepted. Results are still significant at the 0.0005 level as the observed value is still greater than the critical value of 10.83.
Results
The Chi Squared test was chosen as the experiment involved independent samples and it is a test that clearly shows he differences in categorized data and shows the significance of the results. The results gained were very significant. Using the Chi square test has indicated that there was less than a 5% possibility chance that the results were due to chance alone.
5% is considered appropriate because it is not too harsh and it is not too lenient. It also allows for chance.
The observed value of 11.06 is greater than the critical value of significance of 2.71, therefore the null hypothesis can be rejected and the alternative accepted. Results are still significant at the 0.0005 level as the observed value is still greater than the critical value of 10.83.
The bar chart clearly shows that when estimations were present in condition A, there was a clear similarity of estimations which implies that participants conformed to stooges estimates. In condition A there was only 1 estimate more than 1000 away from the actual figure, compared with 9 estimations in condition B.
Discussion
The alternative hypothesis which states that ‘there will be a greater level of conformity in terms of participants estimates in the presence of stooges estimates’ is retained. Using the Chi square test has indicated that there was less than a 5% possibility chance that the results were due to chance alone. Results would actually still be significant at the 0.0005 level. There is still however a slight chance that the null hypothesis can be correct but it seems safe to conclude that there is a greater level of conformity in terms of participants estimates in the presence of stooges estimates. 50% of estimates in condition A seemed to conform to stooges estimates as they were all within a close range of 500. There was only 1 outstanding estimate in condition A in the ‘poor estimate’ category and the remaining 5 all gave ‘reasonable estimates’ which suggests they may have conformed to a certain extent. The ‘poor estimate’ could have occurred for numerous reasons; it could have been simply a poor estimate, person may have wanted to give a bizarre estimate, or participant may have ignored stooges estimates. The results in condition B were very dispersed as expected, with two participants making good estimates, 1 making a reasonable estimate and 9 making poor estimates. The range in condition B was 999600 which demonstrates how disperse the results were.
The results from this experiment support the studies of Asch (1956) and Crutchfield (1955). It would seem that the same trends are still present today more than 40 years after Aschs study.
In Aschs study the average subject conformed to the group response on 32% of the trials and 74% of the subjects conformed at least once. This situation not only emphasises how a lot of people do not want to stand out amongst a group, but also how people look to others for guidance. It is not clear to determine what participants actually conformed , or whether they really did estimate without conforming, however it would be fair to say in this experiment 91.6% of participants in condition A seemed to conform to stooges estimates to some extent at least as this amount were all within 1000 of the actual answer whereas in condition B only 25% of participants were within 1000 of the actual answer. Going on this these results are actually more obvious than those obtained by Asch.
A criticism of this study could be that there is a possibility it could have been influenced by experimenter expectancy. Participants may have been affected by this. If I was to do this again I would counteract this by using an unbiased experimenter to take the experiment. Therefore if participants did feel pressured in giving certain answers, now they would not. Another possible criticism could be that less dots should have been because the range of answers was so large, it was difficult to conclude precisely on the results. The problem could have been easier as in Asch’s study, as in this case of using 3417 dots it was difficult anyway. Maybe using 50 dots would have been more effective. As Chi squared was used, this is a test that proves to display results more clearly when the sample is larger.
It would be interesting to see if participants conformed to stooges answers that were not close to the correct answer. This could be considered as a follow up study. To follow these results up I could see if participants conformed to other peoples estimations if they knew these other people were of a high level of intelligence.
Conclusion
The main findings supported previous research and formed a basis for further research. Expected results were gained with no major limitations to the experiment. Participants were found to conform to stooges estimates as previous research suggested they would
Abstract
The main aim of this study was to follow up on previous research such as Asch’s(1956), and Crutchfield’s(1955) studies to see if the same trends of conformity would still be present today in 6th form students. The way this was done was by asking 24 6th form students of similar trends to estimate how many dots they thought was on a page of A4 paper. As there was 2 conditions (A and B) 12 of the students made their estimations in the presence of stooges estimations and 12 made estimations without this guidance. The aim was to see if the participants in the condition A conformed to the stooges estimates. It was found that in condition A 91% of participants seemed conformed to stooges estimations. The results were very significant.
