Past performance is not an indicator of future returns. That, at least, is the advice given to investors. But can the likelihood of a person committing a crime be predicted by looking at his record? The answer, according to a team of clinical forensic psychologists, is that it cannot. Not only is risk prediction unreliable but, when applied to individuals rather than groups, the margins of error are so high as to render any result meaningless.
Making assumptions about individuals from group data is generally only reasonably safe when the variation within the group is small. Despite this, risk assessments are routinely used to help decide who should be locked up, who should undergo therapy and who should go free. Risk prediction is also set to be used to assess the threat posed by people ranging from terrorist suspects to potential delinquents.
Stephen Hart, of Simon Fraser University in British Columbia, Canada, and colleagues decided to determine how accurate the tests of risk assessment are when applied to individuals rather than groups. Typically the tests work by assigning a score to people depending on factors such as their age, the history of their relationships, their criminal past and the type of victims they have chosen. If someone's score places him in a group in which a known proportion has gone on to commit a crime on release from detention, then the risk that person will prove a recidivist is thought to be similar to the risk for the group as a whole.
The paper published by Dr Hart and his colleagues in last month's issue of the British Journal of Psychiatry focused on two popular tests that follow this logic. The first was a 12-item test designed to assess risk for general violence over periods of seven to ten years. The second was a ten-item test designed to assess risk for violence and sexual violence over periods of five to 15 years. The researchers have also assessed other tests used for predicting sexual offences and domestic violence.
They found that variations between members of the groups were very large. In one of the tests, for example, the standard estimate of the chances of members of the group sexually reoffending was put at 36% within 15 years. They calculated that the actual range was between 30% and 43% of the group, with a 95% confidence level. But calculating the average probability for a group is much easier than calculating the same probability for any individual. Thus, using standard methods to move from group inferences to individual ones, they calculated that the chance of any one person reoffending was in the range of 3% to 91%, similarly with a 95% confidence level. Clearly, the seemingly precise initial figure is misleading.。
The principle is not peculiar to psychology. It has been recognised by statisticians for decades. They call it the ecological fallacy (although this term captures broader subtleties, too). Medicine has also been confounded by statistically based procedures. Indeed, the technique is only really useful when the successes and failures are aggregated. A life-insurance company, for instance, could wrongly predict the life span of every person it insured but still get the correct result for the group.
注(1):本文选自Economist, 06/02/2007
注(2):本文习题命题模仿对象为2004年真题Text 4。
1. What do the clinical forensic psychologists think of risk prediction?
[A] Risk prediction fails in the stock market.
[B] Risk prediction is always effective when being applied to groups.
[C] Risk prediction is not dependable when it comes to individual behavior.
[D] Rist prediction should be widely used in all fields.
2. We can learn from the text that tests of risk assessment are _______.
[A] longitudinal
[B] very tricky
[C] convincing enough
[D] unreasonable
3. Traditional view and that of Dr. Hart on risk assessment are _______.
[A] identical
[B] similar
[C] complemetary
[D] opposite
4. According to Dr. Hart, using standard methods to predict individuals _______.
[A] is as easy as using them for groups
[B] yields ineffective statistics
[C] can help attain precise results
[D] might be influcenced by confidence level
5. Which of the following statements is TRUE?
[A] Ecological fallacy is a psychological phenomenon.
[B] Statisticians have been dealing with ecological fallacy for a long time and have almost succeeded in figuring out a solution.
[C] The statistics from a life insurance company on the longevity of a group of people should be reliable.
[D] Risk prediction is proved to be of zero practical value.
篇章剖析
本文主要就如何科学进行犯罪预测展开论述。第一段首先提出了风险预测能够有效预测犯罪动机的问题;第二段对风险预测做出了简要说明,指出其局限性,且这种方法有些被滥用;第三、四段主要介绍了哈特博士针对以上问题进行的实验;第五段介绍了实验结果,说明了风险预测针对集体的有效性远远超过了针对个人的情况;第六段则进一步对风险预测进行阐发性论述,提出了生态学谬论的概念。
词汇注释
margin [`mB:dVin] n. 页边的空白,差数 confidence level 置信度
assessment [E`sesmEnt] n. 评估,估价 probability [9prCbE`biliti] n. 概率
routine [ru:`ti:n] adj. 例行的;常规的 inference [`infErEns] n. 推论
therapy [`WerEpi] n. 治疗 initial [i`niFEl] adj. 最初的, 初始的
delinquent [di`liNkwEnt] n. 失职者, 违法者 fallacy [`fAlEsi] n. 谬误, 谬论
proportion [prE`pC:FEn] n. 比例, 均衡 confound [kEn`faund] vt. 使混淆, 挫败
detention [di`tenFEn] n. 拘留, 禁闭 aggregate [`Agrigeit] v. 聚集, 合计
psychiatry [sai`kaiEtri] n. 精神病学, 精神病治疗法 insure [in`FuE] vt. 给...保险
domestic [dE`mestik] adj. 家庭的, 国内的
难句突破
If someone's score places him in a group in which a known proportion has gone on to commit a crime on release from detention, then the risk that person will prove a recidivist is thought to be similar to the risk for the group as a whole.
主体句式 If someone's score places him in a group, then the rist is thought to be similar to…
结构分析 这个句子分为逗号前后的两个部分,其中的每一个部分包含一个定语从句。前半个句子中,in which引导的定语从句修饰group;而后半句中,that person will prove a recidivist 作为定语从句修饰risk。
句子译文 如果一个人的分数属于出狱后犯罪率很高的一组,那么此人是惯犯的几率就大体上和这组的整体几率相同
题目分析
1.C. 细节题。根据文章第一、二段的内容,风险预测对于个人行为的预测能力错误率极大,几乎完全不可靠。
2.A. 细节题。文章第四段中举了两个测试的例子,这两个测试都是长达多年的跟踪性试验。
3.D. 细节题。文章第二段中指出,传统上人们习惯用风险评估来对个人的行为作出判断,而哈特博士的实验结果则表明这种预测起到很大的误导作用,可见双方的意见是相反的。
4.B. 细节题。根据文章第五段,如果将群体的标准方法用于个体,最终得到的结果是具有误导性的。
5.C. 细节题。C选项的信息来自文章第六段的最后一句话,A life-insurance company, for instance, could wrongly predict the life span of every person it insured but still get the correct result for the group.即“例如一个人寿保险公司也许不能预测每个人的寿命长度,但它可以对群体做出准确预测。”。
参考译文
过去的付出并不代表未来的收获,这至少可以给投资者一些启示。但一个人的犯罪动机能通过他过去的记录来预测吗?一个临床法医心理学家小组的回答是否定的。风险预测不仅站不住脚,而且即使是只用于个体而不是群体,其误差之大也导致结果失去意义。
只有当一组数据内部变化程度相对较小时,据此做出的对个体的判断才会比较可靠。尽管如此,人们还是例行公事地用风险评估来帮助决定谁应该坐牢、谁应该接受治疗以及谁可获得自由。风险预测同样用于区分评估从来自嫌疑恐怖分子到潜在罪犯的各种威胁恐吓。
加拿大不列颠哥伦比亚省西蒙弗雷泽大学的斯蒂芬·哈特和他的同事决意要找出当这些测试用于个体而非群体时,其准确性到底有多大。这些测试先根据个人年龄、社交经历、犯罪史、以及所选择的受害者类型等因素给人们算出一个分数。如果一个人的分数属于出狱后犯罪率很高的一组,那么此人是惯犯的几率就大体上和这组的整体几率相同。
哈特博士和他的同事在上月的《英国精神病学杂志》上发表了一篇论文,主要介绍了按照这一逻辑进行的两个有名实验。第一个实验包括12项内容的测试,用来评估7至10年间的普通暴力风险;第二个实验包括10项相关内容的测试,用来评估5-15年间的暴力和性暴力风险。研究人员也评估了用来评估性侵犯和家庭暴力的其他一些实验。
他们发现每组成员间的差别很大。比如在其中一项测试中,15年内再次进行性侵犯可能性标准估计是36% 在95%的置信度下,他们得到的实际范围在30%和43%之间。但计算群体的平均概率要比计算个体的概率容易得多。因此,如果将群体的标准方法用于个体,他们计算出一个人是惯犯的几率在同样95%的置信度下为3%-91%。很明显,看似准确的原始数据却起到了误导的作用。
这个原理不仅仅用于心理学,也已统计学承认达数十年之久。他们称其为生态学谬论(尽管这个术语还有更广泛和精细的意义)。医学也被基于统计学的程序弄的一团糟。实际上,这个技术只有在成功和失败都综合起来的时候才真正有用。例如一个人寿保险公司也许不能预测每个人的寿命长度,但它可以对群体做出准确预测。
责任编辑:虫虫