Example 1
Question: The mean and standard deviation of 80 observations were found to be 25 and 4, respectively. Later it was discovered that two observations, recorded as 30 and 14, were incorrect. Find the correct mean and standard deviation after omitting the incorrect observations.
Solution: We are given: number of observations \(n = 80\), mean \(\bar{x} = 25\), and standard deviation \(\sigma = 4\). Step 1: Find the original sum of observations. We know \(\sum x_i = n \cdot \bar{x}\), so \[\sum x_i = 80 \times 25 = 2000\] Step 2: Find the original sum of squares of observations. Using \(\sigma^2 = \frac{\sum x_i^2}{n} - \bar{x}^2\), we get \(\sum x_i^2 = n\left(\sigma^2 + \bar{x}^2\right)\). \[\sum x_i^2 = 80\left(16 + 625\right) = 80 \times 641 = 51280\] Step 3: Remove the incorrect observations (30 and 14) from the sum and sum of squares. Corrected sum: \[\sum x_i^{\text{new}} = 2000 - 30 - 14 = 1956\] Corrected sum of squares: \[\sum x_i^{2,\text{new}} = 51280 - (30)^2 - (14)^2 = 51280 - 900 - 196 = 50184\] Step 4: The new number of observations after omitting two is \(n' = 80 - 2 = 78\). Step 5: Calculate the new mean. \[\bar{x}_{\text{new}} = \frac{1956}{78} = 25.077... \approx 25.08\] Step 6: Calculate the new variance. \[\sigma_{\text{new}}^2 = \frac{\sum x_i^{2,\text{new}}}{n'} - \left(\bar{x}_{\text{new}}\right)^2\] \[\sigma_{\text{new}}^2 = \frac{50184}{78} - \left(\frac{1956}{78}\right)^2\] \[\frac{50184}{78} = 643.384...\] \[\left(\frac{1956}{78}\right)^2 = \left(\frac{326}{13}\right)^2 = \frac{106276}{169} \approx 628.854...\] Let us keep exact fractions. Note that \(\frac{1956}{78} = \frac{326}{13}\). \[\sigma_{\text{new}}^2 = \frac{50184}{78} - \frac{326^2}{13^2} = \frac{50184}{78} - \frac{106276}{169}\] Converting to a common denominator of 1014 (since \(78 = 6 \times 13\) and \(169 = 13^2\), LCM of 78 and 169 is 1014): \[\frac{50184}{78} = \frac{50184 \times 13}{1014} = \frac{652392}{1014}\] \[\frac{106276}{169} = \frac{106276 \times 6}{1014} = \frac{637656}{1014}\] \[\sigma_{\text{new}}^2 = \frac{652392 - 637656}{1014} = \frac{14736}{1014} = \frac{2456}{169}\] \[\sigma_{\text{new}}^2 \approx 14.53\] Step 7: Calculate the new standard deviation. \[\sigma_{\text{new}} = \sqrt{\frac{2456}{169}} = \frac{\sqrt{2456}}{13} \approx \frac{49.56}{13} \approx 3.81\]
Answer: New mean \(\approx 25.08\) and new standard deviation \(\approx 3.81\).
Example 2
Question: The mean and standard deviation of 50 observations are 40 and 5, respectively. It is later found that three observations, recorded as 45, 35, and 42, were incorrect and must be omitted. Find the correct mean and standard deviation of the remaining observations.
Solution: Given: \(n = 50\), \(\bar{x} = 40\), \(\sigma = 5\). Step 1: Compute the original sum of all observations. \[\sum x_i = n \cdot \bar{x} = 50 \times 40 = 2000\] Step 2: Compute the original sum of squares. Using \(\sum x_i^2 = n\left(\sigma^2 + \bar{x}^2\right)\): \[\sum x_i^2 = 50\left(25 + 1600\right) = 50 \times 1625 = 81250\] Step 3: Remove the three incorrect observations (45, 35, 42). Corrected sum: \[\sum x_i^{\text{new}} = 2000 - 45 - 35 - 42 = 1878\] Corrected sum of squares: \[\sum x_i^{2,\text{new}} = 81250 - (45)^2 - (35)^2 - (42)^2\] \[(45)^2 = 2025,\quad (35)^2 = 1225,\quad (42)^2 = 1764\] \[\sum x_i^{2,\text{new}} = 81250 - 2025 - 1225 - 1764 = 76236\] Step 4: New number of observations. \[n' = 50 - 3 = 47\] Step 5: New mean. \[\bar{x}_{\text{new}} = \frac{1878}{47} = \frac{1878}{47} = 39.957... \approx 39.96\] Let us verify: \(47 \times 40 = 1880 \ne 1878\), so \(\bar{x}_{\text{new}} = \frac{1878}{47}\). Step 6: New variance. \[\sigma_{\text{new}}^2 = \frac{\sum x_i^{2,\text{new}}}{n'} - \left(\bar{x}_{\text{new}}\right)^2\] \[\frac{76236}{47} = 1622.042...\] \[\left(\frac{1878}{47}\right)^2 = \frac{1878^2}{47^2} = \frac{3526884}{2209}\] Now, \(\frac{76236}{47} = \frac{76236 \times 47}{2209} = \frac{3583092}{2209}\). \[\sigma_{\text{new}}^2 = \frac{3583092 - 3526884}{2209} = \frac{56208}{2209}\] \[\sigma_{\text{new}}^2 \approx \frac{56208}{2209} \approx 25.45\] Step 7: New standard deviation. \[\sigma_{\text{new}} = \sqrt{\frac{56208}{2209}} = \frac{\sqrt{56208}}{47}\] Simplify: \(56208 = 16 \times 3513 = 16 \times 3 \times 1171\), so \(\sqrt{56208} = 4\sqrt{3513}\). \[\sigma_{\text{new}} = \frac{4\sqrt{3513}}{47} \approx \frac{4 \times 59.27}{47} \approx \frac{237.09}{47} \approx 5.04\]
Answer: New mean \(\approx 39.96\) and new standard deviation \(\approx 5.04\).