等相关矩阵

考虑样本均值 $\overline{X}=\frac{1}{7}\sum_{i=1}^7 X_i$。

有

$$\mathrm{Var}(\overline{X})=\frac{1}{49}\left(\sum_{i=1}^7\mathrm{Var}(X_i)+\sum_{i\ne j}\mathrm{Cov}(X_i,X_j)\right) =\frac{1}{49}(7+42\rho)\ge 0.$$

因此 $7+42\rho\ge 0\Rightarrow \rho\ge -\frac{1}{6}$。

---

英文解析

We are going to consider $\text{Var}(\overline{X})$ here, where (\overline{X} = \frac{X_1 + \dots + X_7}{7}). By plugging this in and using properties of variance, we see that

$$\text{Var}(\overline{X}) = \frac{1}{49} \text{Var}(X_1 + \dots + X_7) = \frac{1}{49}\left[\sum_{i=1}^7 \text{Var}(X_i) + \sum_{i \neq j} \text{Cov}(X_i,X_j)\right]$$

Since we know the mean and variance of each of the random variables, we know that the first sum is just (7). Similarly, we know that (\text{Cov}(X_i,X_j) = \rho(1)(1) = \rho) and that there are (7 \cdot 6 = 42) terms in that sum. Therefore, the second sum is just (42\rho). Thus,

$$\text{Var}(\overline{X}) = \frac{7 + 42\rho}{49}$$

Our condition is that (\text{Var}(\overline{X}) \geq 0), as the variance of any random variable must be non-negative. Thus, we can disregard the denominator and find that

$$7 + 42\rho \geq 0 \iff \rho \geq -\frac{1}{6}$$

Therefore, our answer is (-\frac{1}{6}).