To begin with the fake answer sheets were constructed, these consisted of 9 fake answers (refer to appendix 1 for how fake answers were generated) with a name and sex beside each one, a tenth space was provided for the participant’s estimate. The fake answers in the high condition ranged from 623-663, in the low condition the fake answers ranged from 409-449. Once the answer sheets had been constructed each participant was approached with the condition they were going to be placed in already in mind and asked “would you like to take part in a competition?” if the participant replied “yes” the Participant was given an answer sheet.
They were then informed “you have been given an answer sheet on which you must provide an estimate for the number of beans you believe to be in this jar, the information already on the sheet is of estimates of other participants”. They were then shown the jar of lentil. After handing back the answer sheet the participants were debriefed as to the true nature of the ‘competition’ and asked not to discuss the experiment with any fellow colleagues for the rest of the afternoon.
Controls and ethics Counterbalancing control used in that experiment was conducted on one afternoon, the aim of this was to prevent tutors discussing the experiment with other tutors who potentially could have been participants therefore ruling out any order effects. The fake answers in each of the conditions remained the same for each participant, this was to keep things standardised and so ruling out the possibility that differences in estimates were due to differences in fake answers. This was coupled with standard instructions.
Teachers who were held to have knowledge of psychology were excluded from the study; this was to control the potential of the Hawthorne effect. Clearly where deception is used there will always be a lack in informed consent but this is justifiable using a cost-benefit analysis. It was decided the deception was of a very minor nature and was not enough to amount to cause emotional disturbance in anyone.
Also the estimate that was to be obtained from the participant was considered trivial information, and when debriefed some participants who were deemed as conforming claimed they were giving a genuine independent estimate suggesting they didn’t feel deceived. On the other hand the benefits that were to be gained from conducting the study were seen to be significant compared to the deception. The research was providing information on conformity ‘today’ which could be compared to conformity ‘yesterday’ to give an idea of how conformity has changed over time.
Results The raw data (answers) had to be prepared so that it could be analysed to find a result using the Mann-Whitney-U-Test, this was appropriate as an independent design was used. In order to use the test I ranked all of the data, from the lowest to the highest, rank 1 being the lowest. In some cases the participants estimated the same answer in which case the rank given to that score became the mean of the ranks involved. A one tailed test was used to establish significance because the experimental hypothesis was directional. With 16 participants in each of the conditions the tabled value for significance at 0.05 was at 60. These results reject the experimental hypothesis and retain the null hypothesis (refer to appendix 2 for how this conclusion was reached).
Discussion The results have been shown to be insignificant on the Mann-Whitney U test; this means in general participants did not conform. This result has not confirmed my Experimental Hypothesis which states participants, would show evidence of conformity. In the introduction it was related that conformity had decreased since the original studies, nonetheless results were always shown to be significant even in follow up studies that were carried out two decades later e.g.
Larsen (1974). However, in the Perrin and Spencer (1980) studies they found only one conforming response in 396 trials, as well as the historical difference another factor which was said to influence the results was the sample, which consisted of students doing advanced studies, and thus were able to resist conformity pressures. When they conducted the same study using youths on probation they found similar results to that of Asch. Could this criticism also be applied within this context, and if so the education of the teachers would have been an operative cause to the low levels of conformity which provides support for the basis of the experimental hypothesis but would mean the sample used in this study was unrepresentative of society in general.
It was also cited in the introduction that persons from individualistic cultures are less likely to conform and so it is the case that the lack of conformity may have been due to this rather than education or maybe even both. The fact that what affected conformity was undistinguishable threatens the reliability of the investigation. If the study were to be replicated this problem could be overcome by selecting another set of participants who have no degrees and contrasting their results with those who do.
Since the task was ambiguous the type of social influence that was being investigated was Informational, the fact that the presence of other people was created through fake answers creates implications for the investigation and may threaten the validity of the results since the participants may have reacted differently with the presence of other participants physically. Instead the investigation should be considered as a variation of traditional conformity studies which gives an insight to a factor that affects conformity, i.e. the absence of participants physically.
Another factor that possibly threatened the internal validity of this study was the inability to control confounding variables, which is often common of field experiments; this is a problem because the dependant variable may have been affected by these factors thereby reducing the validity of the results, for example participants may have took the task more seriously if they weren’t preoccupied with their work. Whether participants did not conform is also questionable; the test for significance takes into account every estimate thus estimates that were significantly lower or higher than the others swayed the test for significance rendering it insignificant. It is also true that the number of participants that were selected was very limited and they all had the same occupational background this reduces the external validity of the investigation.
On the basis of the above, alterations that would be appropriate for a follow up study could include: 1. Use of a larger and more representative sample for example instead of just teachers the sample would include people from a range of occupational backgrounds, ages and ethnicity. This would provide more of an insight on the factors that affect conformity i.e. culture, education, age etc.
2. Creation of an unambiguous situation. This would allow contrasting of both types of social influence which would provide a broader understanding of levels of conformity and allow us to distinguish which of the two is a stronger influence in our willingness to conform. 3. Involve real life participants that are there to represent social influence this is because in this investigation the presence of people was made through fake answers which inevitably produces a different reaction in some participants than if they were to encounter real persons.