📚 The key issues of these institutions in dealing with human rights | 这些机构处理人权问题的关键数学视角
Human rights institutions, such as UN treaty bodies, regional courts and national commissions, often rely on quantitative evidence to monitor abuses, rank countries and guide interventions. In A-Level Mathematics, the statistical and decision-making tools used in these processes reveal several key issues: sampling bias, measurement error, invalid inference and strategic incentives. This article examines those issues through the lens of Edexcel A-Level maths, using probability, hypothesis testing, correlation, game theory and optimisation.
处理人权问题的机构——如联合国条约机构、区域法院和国家委员会——常常依靠量化证据来监测侵权、对国家排名并指导干预。在 A-Level 数学中，这些过程中使用的统计与决策工具暴露出若干关键问题：抽样偏差、测量误差、无效推断和策略性激励。本文通过 Edexcel A-Level 数学的视角，运用概率、假设检验、相关分析、博弈论和优化方法来审视这些问题。
1. Human rights indices and composite scores | 人权指数与综合评分
Many institutions construct composite indices, such as the Human Rights Index, by combining several indicators into a weighted sum. In mathematical terms, an index I for country i may be written as follows, where xⱼᵢ are standardised scores and wⱼ are weights.
许多机构通过将若干指标加权求和来构建综合指数，例如人权指数。用数学语言表示，国家 i 的指数 I 可以写为如下形式，其中 xⱼᵢ 是标准化得分，wⱼ 是权重。
Iᵢ = w₁x₁ᵢ + w₂x₂ᵢ + ⋯ + wₙxₙᵢ
The choice of weights is subjective and can change rankings dramatically. If one institution gives torture a weight of 0.4 and another gives it 0.1, the same country can appear very different in league tables.
权重的选择是主观的，可能显著改变排名。如果一个机构赋予酷刑指标 0.4 的权重，而另一个机构赋予 0.1，那么同一个国家在排名表中的表现可能截然不同。
- Key issue: weighting uncertainty and lack of a unique mathematical solution.
- 关键问题：权重不确定且缺乏唯一的数学解。
2. Sampling bias in data collection | 数据收集中的抽样偏差
Human rights violations are rarely observed through random samples. Data often come from media reports, NGO interviews or asylum claims, which over-represent accessible urban areas and under-represent hidden abuses.
人权侵犯很少通过随机样本观察到。数据通常来自媒体报道、非政府组织访谈或庇护申请，这些数据过度代表了可接触的城市地区，而低估了隐藏的侵权行为。
In A-Level terms, this is a non-random sample, so the sample mean x̄ is a biased estimator of the true population mean μ. The bias is defined as the expected difference between the estimator and the true parameter.
用 A-Level 的话说，这是非随机样本，因此样本均值 x̄ 是真实总体均值 μ 的有偏估计量。偏差定义为估计量的期望值与真实参数之间的差。
Bias = E(X̄) − μ
When certain groups are systematically missed, this bias is unlikely to be zero, and any ranking based on such data will be distorted.
当某些群体被系统性遗漏时，这种偏差不太可能为零，任何基于此类数据的排名都会失真。
3. Measurement error and reliability | 测量误差与信度
Indicators such as “freedom from torture” are latent variables that cannot be measured directly. Institutions use proxies like reported cases, which contain random and systematic errors.
诸如“免于酷刑”等指标是无法直接测量的潜变量。机构使用报告案件数等代理变量，其中包含随机误差和系统误差。
If the observed score Y equals the true value T plus an error term e, then the reliability ratio is the proportion of variance in Y that comes from T.
若观测得分 Y 等于真值 T 加上误差项 e，则信度比是 Y 的方差中来自 T 的比例。
Y = T + e, Reliability = Var(T) ÷ Var(Y)
Low reliability weakens correlation and regression results, making comparisons between countries misleading. A measurement with high error variance can hide real differences in human rights performance.
低信度会削弱相关和回归结果，使国家之间的比较产生误导。误差方差大的测量会掩盖人权表现的真正差异。
4. Hypothesis testing in rights monitoring | 人权监测中的假设检验
A common question is whether a new policy has reduced violations. Institutions may test the null hypothesis that the mean number of violations before and after the policy is the same against the alternative that it has decreased.
一个常见问题是新政策是否减少了侵权行为。机构可以检验原假设：政策前后的平均侵权数量相同，备择假设：平均数量下降。
H₀: μ₀ = μ₁ versus H₁: μ₀ > μ₁
If the sample size is small, the test has low power, so a real improvement may not be detected. The p-value gives the probability of observing data at least as extreme as the sample, assuming H₀ is true.
如果样本量小，检验的功效就很低，因此真实改善可能无法被检测到。p 值表示在 H₀ 为真时观察到至少与样本一样极端的数据的概率。
A p-value above 0.05 does not prove no effect; it may simply reflect insufficient data. This is a key issue because institutions often treat non-significance as evidence of no violation.
p 值大于 0.05 并不能证明没有效果；它可能只反映数据不足。这是一个关键问题，因为机构常常把不显著当作没有侵权的证据。
5. Correlation and causation in abuse patterns | 侵权模式中的相关与因果
A high correlation between poverty and human rights violations does not imply that poverty causes violations. For example, Pearson’s correlation coefficient r may be 0.8, but confounding variables such as weak rule of law may drive both.
贫困与人权侵犯之间的高度相关并不意味着贫困导致了侵犯。例如，皮尔逊相关系数 r 可能为 0.8，但法治薄弱等混杂变量可能同时驱动两者。
A-Level students learn that correlation shows linear association only. The formula for r standardises the covariance between two variables.
A-Level 学生学到相关仅表示线性关联。r 的公式将两个变量的协方差标准化。
r = Sxy ÷ √(SxxSyy)
Institutions that infer causation from r risk designing ineffective interventions. A policy aimed at reducing poverty may not reduce violations if the real cause is lack of judicial independence.
从 r 推断因果关系的机构可能会设计出无效的干预措施。如果真正的原因是缺乏司法独立，那么旨在减少贫困的政策可能不会减少侵权行为。
6. Regression models for predicting violations | 预测侵权的回归模型
A simple linear regression model can be fitted, where y is the number of reported violations and x is an explanatory variable such as military expenditure.
可以拟合简单线性回归模型，其中 y 是报告侵权数量，x 是解释变量，如军费开支。
y = a + bx
The least squares estimates minimise the sum of squared residuals, where ŷᵢ is the predicted value for observation i.
最小二乘估计使残差平方和最小，其中 ŷᵢ 是第 i 个观测值的预测值。
Minimise Σ(yᵢ − ŷᵢ)²
However, extrapolation beyond the data range is dangerous: a model fitted to historical data may fail when political conditions change, and outliers can heavily influence the slope b. This makes long-term prediction of human rights violations unreliable.
然而，超出数据范围的外推是危险的：基于历史数据拟合的模型在政治条件变化时可能失效，异常值也会严重影响斜率 b。这使得对人权侵犯的长期预测不可靠。
7. Decision trees and intervention strategies | 决策树与干预策略
When an institution decides whether to intervene, it may use a decision tree. Each branch has a probability and a payoff measured in lives saved or rights protected.
当机构决定是否干预时，可以使用决策树。每个分支有概率和以挽救生命或保护权利衡量的收益。
The expected value is calculated by multiplying each outcome by its probability and summing the results.
期望值通过将每个结果乘以其概率并求和来计算。
EV = Σ pᵢ × outcomeᵢ
Choosing the branch with the highest EV is rational only if the probabilities and utilities are accurate. In human rights crises, both are uncertain, and risk aversion may lead to different decisions than expected value maximisation.
选择最高 EV 的分支只有在概率和效用准确时才是理性的。在人权危机中，两者都不确定，风险厌恶可能导致与期望值最大化不同的决策。
8. Game theory and state compliance | 博弈论与国家合规
A state may choose to comply with human rights treaties or violate them. This can be modelled as a payoff matrix. If the cost of compliance is high and the probability of sanction is low, the dominant strategy may be to violate.
国家可以选择遵守人权条约或违反条约。这可以建模为支付矩阵。如果合规成本高而制裁概率低，占优策略可能是违反。
| State A / State B | Comply | Violate |
|---|---|---|
| Comply | (2, 2) | (1, 5) |
| Violate | (5, 1) | (3, 3) |
For example, if compliance gives payoff 2 and violation gives payoff 5 with no detection, violation dominates. Institutions face the key issue that treaty enforcement resembles a repeated game, where reputation and long-term costs can alter strategies.
例如，如果合规收益为 2，违反收益为 5 且没有检测，则违反占优。机构面临的关键问题是，条约执行类似于重复博弈，声誉和长期成本可以改变策略。
9. Optimisation of limited resources | 有限资源的优化
Human rights institutions have limited budgets and must allocate resources across many crises. This is a constrained optimisation problem: maximise total impact Q subject to a budget constraint.
人权机构预算有限，必须在许多危机之间分配资源。这是一个约束优化问题：在预算约束下最大化总影响 Q。
Maximise Q = Σ qᵢ(rᵢ) subject to Σ rᵢ ≤ B
Here rᵢ is the resource allocated to country i and B is the total budget. Linear programming can find an optimal allocation, but the objective function qᵢ(rᵢ) is hard to specify because human rights outcomes are not linearly related to spending and may involve diminishing returns.
其中 rᵢ 是分配给国家 i 的资源，B 是总预算。线性规划可以找到最优分配，但目标函数 qᵢ(rᵢ) 很难确定，因为人权结果与支出并非线性相关，并且可能存在边际收益递减。
10. Uncertainty and the precautionary principle | 不确定性与预防原则
Because data are incomplete, institutions often face Type I and Type II errors. A Type I error is concluding a violation exists when it does not; a Type II error is missing a real violation.
由于数据不完整，机构常常面临第一类错误和第二类错误。第一类错误是在不存在侵权时断定存在侵权；第二类错误是遗漏了真实侵权。
In human rights work, the cost of a Type II error can be catastrophic. The precautionary principle suggests acting even when evidence is not statistically significant, which conflicts with standard hypothesis testing where the significance level α is set at 0.05.
在人权工作中，第二类错误的代价可能是灾难性的。预防原则建议即使证据没有达到统计显著也采取行动，这与标准假设检验中显著性水平 α 设为 0.05 相冲突。
This is a key issue in institutional decision making: a mathematical rule designed for scientific neutrality may be too conservative for
Published by TutorHao | A-Level Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导