A study by Blackmore and Troscianko (1985) found sheep were more biased to probability judgements. In this study participants were asked to estimate the probability of repetition in two events. One was throwing a dice ten times, and the other was throwing ten dice at once, and they were asked to estimate the amount of repetition that would occur. The results showed that believers (sheep) in ESP were found to underestimate chance more than non-believers (goats). However this experiment did not differ in avoidance for sheep and goats, all participants tended towards underestimation. This study has been criticised by critics of parapsychology who argue that there is a lack of valid evidence for the study because of the lack of duplication of study, poor methodology and fraud.
Brugger et al. (1990) completed a series of experiments on selective random generation and participant’s belief in ESP. They hypothesised that sheep would be more likely to avoid giving repetitions in an experiment task than goats. As a result they confirmed their hypothesis and demonstrated the sheep/goat effect. The sheep/goat effect says that believers in ESP (sheep) are more likely to score above the mean chance expectancy than non-believers (goats) (Brugger et al., 1990) and that believers (sheep) would score significantly lower on repetition than non-believers (goats) in a subjective random generation task.
French (1992) reviewed literature into the sheep/goat effect and concluded that although there are many uncertainties, the available evidence suggests that sheep (believers) may indeed be prone to bias.
This has lead to this experiment which is based on the work of Brugger et al. (1990), which aims to look at people’s belief in Extrasensory perception (ESP) and the effect this has on the probability of chance in equal chance tasks. From the studies in ESP, believers in paranormal are expected to make greater underestimations than non-believers. More generally it could hypothesize that sheep (believers in ESP) will record significantly fewer repetitions than goats (non believers in ESP) on a selective random generation task. Under null hypothesis there will be little or no differences between the number of repetitions sheep (believers in ESP) and goats (non believers in ESP) make on a selective random generation task.
Method
Design:
The experiment was an independent measures design, using 194 participants, who were selected using an opportunity sample. Independent variable (IV) is whether participants believe in ESP or not. Dependent variable (DV) is the number of repetitions on a subjective random generation task.
Procedure:
Participants one at a time were asked to imagine a dice being rolled. They were asked to say the digits one to six at random so as ‘to make the resulting sequence as indistinguishable as possible from that of an actually rolled dice’. Participants were asked to repeat this 66 times, and all 66 responses were recorded. Immediately after completing this task the participants were asked to rate their belief in ESP on a six-point scale, ranging from strong belief (=1) to strong disbelief (=6). Each participant was asked to complete the same task.
Scoring:
The score for each participant is calculated by examining the sequence of responses and counting the number of consecutive repetitions of the same digit.
For example in the following sequence
1 2 2 5 6 3 4 1 1 1 6 3 4 2 1 5 ……..
There are 3 repetitions.
Participants were then classified according to their ESP belief.
Scores 1 or 2 were classed as ‘sheep’
Scores 3 or 4 were classed as ‘indifferent’
Scores 5 or 6 were classed as ‘goats’
The results were then put into a table (see appendix)
Ethics
During this experiment ethical consideration had to be considered and followed accordingly.
All participants throughout the experiment had the right to withdraw at any time, this was explained to them at the beginning of the experiment so no pressure was placed on them to complete the task. Confidentiality was achieved by keeping all results anonymous, and asking participants to complete the task only in front of the experimenter, this ensured that the participant was not under any group pressure and did not conform to the group norm, as well as ensuring confidentiality. All participants needed to be naïve to the hypothesis, as a result all participants were fully debriefed after the event on the nature of the experiment and how there results were going to be used within the experiment. Any deception that may have took place during the experiment was overcome with debrief, as this allowed the participant to ask any questions about the experiment and fully understand what they have just taken part in.
Results
Of the 194 participants who took part in the experiment, 37 were rated sheep, 60 were rated goats and 97 were rated indifferent.
Table 1.1 Participant scores in a subjective random generation task
Graph 1.1 Graph to compare mean number of repetitions
The Mean Chance Expectancy (MCE) for all participants in a sequence of 66 dice rolls, is mathematically calculated at 10.8 for consecutive repetitions in a sequence of 66 dice rolls.
65 x 1/6 =10.8
The mean number of consecutive repetitions for both sheep (mean=3.41, SD=6.69) and goats (mean=4.78, SD=18.37) was significantly smaller than the mathematically expected score of 10.8 consecutive repetitions in a sequence of 66 rolls.
Table 1.2 Results needed for statistical calculations
Sheep (mean=3.41, SD=6.69) scored fewer repetitions than goats (mean=4.78, SD=18.37). This difference is significant in the direction predicted by the hypothesis so the null hypothesis is rejected (t(95)=1.74, p<0.05, one tailed).
Discussion
The results showed that non-believers (goats) have more repetitions than believers (sheep). From the results shown in table 1.1 the sheep scored the lowest mean (3.41) for consecutive repetitions, followed by indifferent (4.57), and then goats with the largest mean (4.78). This suggests there is a positive correlation between the replication scores and the level of belief in ESP. Both these findings are supported by previous work done by Brugger et al. (1990), who found similar findings in his work. The results also showed that non-believers in ESP have a bigger variance than believers. Both believers and non-believers have a much smaller mean then the mean chance expectancy. This raises the question how big does the difference between sheep and goats have to be in order for the IV (belief in ESP) to have caused the DV (repetition)? This may be considered in future studies in the relationship between belief in ESP and repetition in a subjective random generation task.
As the results have shown, there is a significant difference between sheep and goats. However this could be a result of a type 1 error, where there is a difference and the null hypothesis is rejected but the null hypothesis is actually true. This study does have the support of previous work, such as Brugger et al. (1990), which suggests this isn’t the result of a type 1 error.
There are also several other weaknesses in this study which should be noted for future work. The sex of the participants was not considered during this experiment. It may however be useful if the study was to be replicated to do so, as the gender of the participant may have an effect on their belief in ESP. Similarly we should also consider the class, social status, age and ethnicity of participants, as these may be confounding variables and consequently affect the results.
Another problem is not all participants used the six-point scale (for rating belief in ESP) in the same way. Some participants may believe stronger in some aspects of ESP but not as strong in others, consequently they may not know where to rate themselves on a six-point scale. Without standardising the scale and explaining each point on the scale fully people will rate themselves differently, which therefore makes it a less accurate measure of belief in ESP.
The study sufferers from a lack of ecological validity and mundane realism, this is due to study taking place in an artificial environment with a task which is not a task that takes place on an everyday occurrence. This could be overcome by conducting a field/observation experiment where participants are observed for replication in everyday tasks. There is also the concern that the study is suffering from intentional or unintentional bias from both the experimenter and the participants, which could lead to demand characteristics by the participants as they are behaving in a certain way which is not necessarily ‘normal’ to their behaviour.
Therefore to conclude this study into the relationship between repetition and the belief in ESP, has shown there is a significant difference between believers and non-believer in ESP. The difference was significant in the direction of the hypothesis, which stated that believers in ESP will record significantly fewer repetitions than non-believers in ESP on a random selective generation task. The null hypothesis was therefore rejected. The findings from this study are supported by previous work into this area of psychology, and can support the sheep/goat effect.
References
Blackmore, S. and Troscianko, T. (1985). Belief in the paranormal: Probability judgements, illusory control, and the ‘chance baseline shift’. British Journal of Psychology, 76 (4), 459-468.
Brugger, P., Landis, T., and Regard, M. (1990). A ‘sheep-goat effect’ in repetition avoidance: Extra-sensory perception as an effect of subjective probability. British Journal of Psychology, 81 (4), 455-468.
French, C.C. (1992). Factors underlying belief in paranormal: Do sheep and goats think differently? The Psychologist: Bulletin of the British Psychological Society, 5, 295-299.
Appendix