The base rate Fallacy - Coursework Example

The base rate Fallacy For the proper understanding of psychology there needs to be thorough and complete probabilistic reasoning understanding. This is mainly because many psychological researches depend on probabilistic reasoning (Shermer, 2008). From the question, one is to assume that 1 in every 1000 employees in government and its defense contractors is a spy. As this is the information about the sample provided. Giving the test to all the employees of the government and defense contractors, it implies that 999 individuals who are not spies will be subjected to the test. This is because only one is a spy of these 1000 individuals, 50 of them will be identified as a spy by the test. The test has a 95% accuracy implying it is 95% accurate. Hence, the total number of employees that will be testing positive for spies will be 51. Of the 51 that test positive it will actually be only 1 that is truly a spy since there is one spy in 1000 employees. Hence, the chance that an individual who tests positive and is actually a spy will be tested that will be 1 in 51. Hence approximately 2% (two percent) of the individuals who test positively and identified as a spy will actually be a spy (Frederick & Bowden, 2009). From the above analysis it is evident that, from the base rate of spying, most of the population of employees do not constitute of spies. But the number of most of them gets falsely accused when combined by a positive rate that is even small. As evident from the above analysis that have proved a very small positive rate of I in 1000 leads to 50 individuals being falsely accused when a population of 1000 are tested. This is a large number of individuals when compared to the actual figure that should be 1. Shermer, (2008) argues that there exist much intertwining between techniques and empirical methods of psychological theory construction and statistics as detailed in the two case scenarios in the questions. In the second scenario relating the probability of spies it is only through statistics that one can only ascertain that actually only 2 percent will actually be spies and not 95% as suggested by the accuracy level of the test that they are being subjected to. In the First example it is only through statistics that one finds out that in a population of 1000 employees where only 5 percent are guilty and all are given a test the number of the wrongly accused will be 285. According to Frederick & Bowden (2009), probabilistic conclusions are the main conclusions that are derived from psychological research just as in the case of most sciences. These conclusions can be referred to as generalizations that usually hold in more often situations. But the generalizations are not usually applicable in every case. Like in the above question the generalization will be 2 percent. Predictions that are got from the psychological theories and findings, though not 100 percent are often of great importance and are still useful to the world today. The fact that they are not 100% true is a quality that they do share with other sciences (Shermer, 2008). Many individuals have been found to have various difficulties when it comes to reasoning in probabilistic terms. Due to this, much of the psychological research has been found to be hard to understand. Probabilistic reasoning has been found to be greatly misunderstood by most people (Shermer, 2008). Like in the question above, many individuals will fail to understand the reasoning and most probably answer as 95% the person who tests positively is actually a spy due to the notion and assumption that the test is 95% accurate. When in the real case it should be 2% since one should utilize the sample information provided for analysis and drawing conclusions. The common mistake that they normally make is the display of the gamblers fallacy from their point in situations where they tend to see linkages among some of the events that in the real sense, they are independent. The best explanation for this in the scenario of this question will be if the population of the employees of the government and its employees had a 50-50 probability that an individual chosen randomly will either be a spy or not. After the choice, the individual is returned into the population and another independent choice made. In the event the first four choices made randomly with replacement all picked spies and individuals asked what one would think the fifth choice would be. Some might argue that since the first four choices chose spies then it is likely that fifth random choice will be a spy since it is like the trend (Frederick & Bowden, 2009).The others will be of the view that, since the first four have been spies, then the time for continuing to choose spies is due hence the fifth will not be a spy. In both cases the reasoning are not correct since the random choosing are independent events that do not depend on each other. Hence in the first four choices, there is no way that it can be related or influence the fifth choice. In addition, they fail to appreciate the fact that when one uses a larger sample of population there are bound to receive more accurate results. For this example, if a smaller population could have been taken the result that could have been received in the chances of the individual who tests positive is actually a spy could be in the extremes. Hence as Shermer, (2008) suggests, one should always understand that what they are dealing with are only samples of the population and not the whole entity. If this is not clearly understood it might lead to errors in one’s judgment. As in this question they might tend to answer that, since the test is 95% accurate then the probability that the person who tests positively is actually a spy is 95 percent. This will be wrong. Finally most of them do possess vivid information, and in addition from the probabilistic information that they have they do make insufficient use of the information. It is evident that that for polygraphs the 75% accuracy rate, though it may sound good it is not actually good. From the first example it was established that if 5 percent of employees are guilty of misconduct. And all the employees are subjected to the test; the number of those wrongly accused will be 285. This is a large number of employees that will be wrongly accused going by the 70% accuracy rate for the polygraphs. Likewise in the question that has been solved, from a population of 1000 individuals, 50 will be falsely accused when the test is run on the whole population. This is a large number of falsely accused people when compared to the truth (Frederick & Bowden, 2009). References Frederick, R., & Bowden, S. (2009). The Test Validation Summary. Assessment, 16, 3, 215-236. Shermer, M. (September 01, 2008). Skeptic - Why our brains do not intuitively grasp probabilities. Scientific American, 40. Read More