𝑼𝑵𝑰𝑻 − 𝑰𝑰𝑰

𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈 𝑫𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔 𝒂𝒏𝒅 𝒕𝒆𝒔𝒕𝒊𝒏𝒈 𝑯𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔

𝑷𝒐𝒑𝒖𝒍𝒂𝒕𝒊𝒐𝒏:
𝐼𝑡 𝑖𝑠 𝑡ℎ𝑒 𝑎𝑔𝑔𝑟𝑒𝑔𝑎𝑡𝑒 𝑜𝑟 𝑡𝑜𝑡𝑎𝑙𝑖𝑡𝑦 𝑜𝑓 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐𝑎𝑙 𝑑𝑎𝑡𝑎 𝑓𝑜𝑟𝑚𝑖𝑛𝑔 𝑎 𝑠𝑢𝑏𝑗𝑒𝑐𝑡 𝑜𝑓 𝑖𝑛𝑣𝑒𝑠𝑡𝑖𝑔𝑎𝑡𝑖𝑜𝑛.
𝑺𝒂𝒎𝒑𝒍𝒆:
𝐴 𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑒𝑥𝑎𝑚𝑖𝑛𝑒𝑑 𝑤𝑖𝑡ℎ 𝑎 𝑣𝑖𝑒𝑤 𝑜𝑓 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑖𝑛𝑔 𝑡ℎ𝑒
𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠.
𝑇ℎ𝑒𝑟𝑒 𝑏𝑦 𝑠𝑡𝑢𝑑𝑦𝑖𝑛𝑔 𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠 𝑜𝑓 𝑠𝑚𝑎𝑙𝑙 𝑔𝑟𝑜𝑢𝑝 𝑎𝑛𝑑 𝑒𝑥𝑡𝑒𝑛𝑑𝑖𝑛𝑔 𝑡ℎ𝑒𝑚 𝑡𝑜 𝑡ℎ𝑒 𝑏𝑖𝑔
𝑔𝑟𝑜𝑢𝑝 𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝑆𝑎𝑚𝑝𝑙𝑖𝑛𝑔.
𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈 𝑫𝒊𝒔𝒕𝒓𝒊𝒃𝒖𝒕𝒊𝒐𝒏𝒔:
𝑇ℎ𝑒 𝑐𝑜𝑙𝑙𝑒𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝑎𝑙𝑙 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓 𝑎 𝑔𝑖𝑣𝑒𝑛 𝑠𝑖𝑧𝑒 𝑑𝑒𝑓𝑖𝑛𝑒𝑠 𝑎 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑐𝑎𝑙𝑙𝑒𝑑 𝑎𝑠 𝑆𝑎𝑚𝑝𝑙𝑖𝑛𝑔
𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛.
𝑵𝒐𝒕𝒂𝒕𝒊𝒐𝒏:
𝑷𝒂𝒓𝒂𝒎𝒆𝒕𝒆𝒓 𝑷𝒐𝒑𝒖𝒍𝒂𝒕𝒊𝒐𝒏 𝑺𝒂𝒎𝒑𝒍𝒆
𝑆𝑖𝑧𝑒 N n
𝑀𝑒𝑎𝑛 𝜇 𝑥
𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝜎2 𝒔𝟐
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝜎 S
𝑃𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 p P

𝑻𝒚𝒑𝒆𝒔 𝒐𝒇 𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈:
𝟏. 𝑹𝒂𝒏𝒅𝒐𝒎 𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈:
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑚𝑒𝑡ℎ𝑜𝑑 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔, 𝑎𝑙𝑙 𝑚𝑒𝑚𝑏𝑒𝑟𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 ℎ𝑎𝑣𝑒 𝑒𝑞𝑢𝑎𝑙 𝑐ℎ𝑎𝑛𝑐𝑒 𝑡𝑜 𝑏𝑒
𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑 𝑎𝑠 𝑎 𝑚𝑒𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒. 𝐼𝑓 𝑁 𝑖𝑠 𝑡ℎ𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛, 𝑡ℎ𝑒𝑛 𝑒𝑣𝑒𝑟𝑦 𝑖𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙
1
𝑖𝑛 𝑡ℎ𝑎𝑡 𝑔𝑟𝑜𝑢𝑝 ℎ𝑎𝑣𝑒 𝑎𝑛 𝑒𝑞𝑢𝑎𝑙 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑡𝑜 𝑏𝑒 𝑐ℎ𝑜𝑜𝑠𝑒𝑛 𝑎𝑠 𝑡ℎ𝑒 𝑚𝑒𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒.
𝑁
𝟐. 𝑺𝒕𝒓𝒂𝒕𝒊𝒇𝒊𝒆𝒅 𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈:
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑚𝑒𝑡ℎ𝑜𝑑 𝑡ℎ𝑒 𝑒𝑛𝑡𝑖𝑟𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑖𝑠 𝑑𝑖𝑣𝑖𝑑𝑒𝑑 𝑖𝑛𝑡𝑜 𝑠𝑚𝑎𝑙𝑙 𝑔𝑟𝑜𝑢𝑝𝑠 𝑤ℎ𝑖𝑐ℎ 𝑎𝑟𝑒 𝑐𝑎𝑙𝑙𝑒𝑑 𝑆𝑡𝑟𝑎𝑡𝑎
𝑇ℎ𝑒𝑠𝑒 𝑆𝑡𝑟𝑎𝑡𝑎 𝑎𝑟𝑒 ℎ𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝑖𝑛 𝑛𝑎𝑡𝑢𝑟𝑒. 𝑇ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑐𝑎𝑛 𝑏𝑒 𝑑𝑟𝑎𝑤𝑛 𝑏𝑦 𝑡𝑎𝑘𝑖𝑛𝑔 𝑡ℎ𝑒 𝑚𝑒𝑚𝑏𝑒𝑟𝑠
𝑓𝑟𝑜𝑚 𝑒𝑎𝑐ℎ 𝑜𝑓 𝑡ℎ𝑒𝑠𝑒 𝑆𝑡𝑟𝑎𝑡𝑎.
𝟑. 𝑷𝒖𝒓𝒑𝒐𝒔𝒊𝒗𝒆 𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈:
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑡𝑦𝑝𝑒 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔, 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑖𝑙𝑙 𝑏𝑒 𝑐ℎ𝑜𝑜𝑠𝑒𝑛 𝑎𝑐𝑐𝑜𝑟𝑑𝑖𝑛𝑔 𝑡𝑜 𝑡ℎ𝑒 𝑢𝑠𝑒𝑟 ′ 𝑠 𝑐ℎ𝑜𝑖𝑐𝑒.
𝑇ℎ𝑒 𝑝𝑒𝑟𝑠𝑜𝑛𝑎𝑙 𝑗𝑢𝑑𝑔𝑒𝑚𝑒𝑛𝑡 𝑝𝑙𝑎𝑦𝑠 𝑎 𝑣𝑖𝑡𝑎𝑙 𝑟𝑜𝑙𝑒 𝑖𝑛 𝑠𝑒𝑙𝑒𝑐𝑡𝑖𝑛𝑔 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒. 𝑁𝑜 𝑠𝑐𝑖𝑒𝑛𝑡𝑖𝑓𝑖𝑐 𝑚𝑒𝑡ℎ𝑜𝑑
𝑜𝑟 𝑚𝑎𝑡ℎ𝑒𝑚𝑎𝑡𝑖𝑐𝑎𝑙 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑠 𝑎𝑟𝑒 𝑖𝑛𝑐𝑙𝑢𝑑𝑒𝑑. 𝑇ℎ𝑖𝑠 𝑖𝑠 𝑝𝑢𝑟𝑒𝑙𝑦 𝑏𝑖𝑎𝑠𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔.
𝟒. 𝑺𝒆𝒒𝒖𝒆𝒏𝒕𝒊𝒂𝒍 𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈:
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑚𝑒𝑡ℎ𝑜𝑑, 𝑓𝑖𝑟𝑠𝑡 𝑠𝑎𝑚𝑝𝑙𝑒 𝑖𝑠 𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑 𝑡𝑜 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒 𝑡ℎ𝑒 𝑐ℎ𝑎𝑟𝑎𝑐𝑡𝑒𝑟𝑖𝑠𝑡𝑖𝑐𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛.
𝐼𝑓 𝑖𝑡 𝑖𝑠 𝑢𝑝𝑡𝑜 𝑡ℎ𝑒 𝑚𝑎𝑟𝑘, 𝑡ℎ𝑒 𝑝𝑟𝑜𝑐𝑒𝑠𝑠 𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑡𝑒𝑠, 𝑒𝑙𝑠𝑒 𝑡ℎ𝑒 𝑠𝑒𝑐𝑜𝑛𝑑 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑖𝑙𝑙 𝑏𝑒 𝑐ℎ𝑜𝑜𝑠𝑒𝑛
𝑒𝑖𝑡ℎ𝑒𝑟 𝑤𝑖𝑡ℎ 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑜𝑟 𝑤𝑖𝑡ℎ 𝑜𝑢𝑡 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛. 𝐻𝑒𝑛𝑐𝑒 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠
𝑎𝑟𝑒 𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑 𝑖𝑛 𝑎 𝑠𝑒𝑞𝑢𝑒𝑛𝑡𝑖𝑎𝑙 𝑚𝑎𝑛𝑛𝑒𝑟 𝑏𝑎𝑠𝑒𝑑 𝑜𝑛 𝑡ℎ𝑒 𝑟𝑒𝑠𝑢𝑙𝑡𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑎𝑠𝑡.
𝟓. 𝑺𝒚𝒔𝒕𝒆𝒎𝒂𝒕𝒊𝒄 𝑺𝒂𝒎𝒑𝒍𝒊𝒏𝒈:
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑚𝑒𝑡ℎ𝑜𝑑, 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑎𝑟𝑒 𝑐ℎ𝑜𝑜𝑠𝑒𝑛 𝑎𝑡 𝑟𝑒𝑔𝑢𝑙𝑎𝑟 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙𝑠 𝑜𝑓 𝑡𝑖𝑚𝑒 𝑖𝑛 𝑎 𝑠𝑦𝑠𝑡𝑒𝑚𝑎𝑡𝑖𝑐
𝑚𝑎𝑛𝑛𝑒𝑟. 𝐼𝑛 𝑡ℎ𝑖𝑠 𝑚𝑒𝑡ℎ𝑜𝑑 𝑡ℎ𝑒 𝑐ℎ𝑜𝑖𝑐𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑖𝑠 𝑏𝑎𝑠𝑒𝑑 𝑜𝑛 𝑡ℎ𝑒 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙 𝑜𝑓 𝑡𝑖𝑚𝑒.
1. 𝑇ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑡ℎ𝑎𝑡 𝑐𝑎𝑛 𝑏𝑒 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑖𝑧𝑒 𝑁 𝑤𝑖𝑡ℎ
𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 𝑛 𝑎𝑟𝑒 𝑤𝑖𝑡ℎ 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 = 𝑁 𝑛
2. 𝑊𝑖𝑡ℎ 𝑜𝑢𝑡 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 = 𝑁𝐶𝑛
𝑁−𝑛
3. 𝐶𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 =
𝑁−1
𝑛
1
4. 𝑆𝑎𝑚𝑝𝑙𝑒 𝑀𝑒𝑎𝑛(𝑥) = ∑ 𝑥𝑖
𝑛
𝑖=1
𝑛 𝑛
1 1
5. 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 (𝑆 2 ) = ∑(𝑥𝑖 − 𝑥)2 ≅ ∑(𝑥𝑖 − 𝑥)2
𝑛 𝑛−1
𝑖=1 𝑖=1
𝜎
6. 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐸𝑟𝑟𝑜𝑟 =
√𝑛
𝑷𝒓𝒐𝒃𝒍𝒆𝒎𝒔:
1. 𝐸𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑡ℎ𝑒 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 𝑎𝑛𝑑 𝑎𝑙𝑠𝑜, 𝑡ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑡ℎ𝑎𝑡 𝑐𝑎𝑛 𝑏𝑒 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚
𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑎𝑛𝑑 𝑤𝑖𝑡ℎ𝑜𝑢𝑡 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡, 𝑤ℎ𝑒𝑟𝑒 𝑛 = 100 𝑎𝑛𝑑 𝑛 = 5
𝑁−𝑛
𝑆𝑜𝑙: 𝐶𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 =
𝑁−1
100 − 5 95
= = = 0.9596
100 − 1 99
𝑁𝑜. 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠
𝑤𝑖𝑡ℎ 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 = 𝑁 𝑛 = 1005 = 10000000000 = 1010
𝑤𝑖𝑡ℎ 𝑜𝑢𝑡 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 = 𝑁𝐶𝑛 = 100𝐶5 = 75287520

2. 𝐴 𝑝𝑜𝑝𝑢𝑎𝑙𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑠 𝑚𝑒𝑚𝑏𝑒𝑟𝑠 2,3,6,8,11. 𝐸𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑡ℎ𝑒 𝑛𝑜. 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓 𝑠𝑖𝑧𝑒 2

𝑡ℎ𝑎𝑡 𝑐𝑎𝑛 𝑏𝑒 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑖𝑠 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛, 𝑤𝑖𝑡ℎ 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑎𝑛𝑑 ℎ𝑒𝑛𝑐𝑒 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑡ℎ𝑒
𝑓𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔 (𝑖) 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 (𝑖𝑖)𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
(𝑖𝑖𝑖) 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛𝑠 (𝑖𝑣) 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒
𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛𝑠.
𝑆𝑜𝑙: 𝐺𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝑁 = 5 𝑎𝑛𝑑 𝑛 = 2
𝑁𝑜. 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑤𝑖𝑡ℎ 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡𝑠 = 52 = 25
2 + 3 + 6 + 8 + 11 30
(𝑖) 𝑀𝑒𝑎𝑛 = 𝜇 = = =6
5 5
5
1
(𝑖𝑖) 𝜎 2 = ∑(𝑥𝑖 − 𝜇)2
5
𝑖=1

1
= [(2 − 6)2 + (3 − 6)2 + (6 − 6)2 + (8 − 6)2 + (11 − 6)2 ]
5
1
= [(−4)2 + (−3)2 + 0 + (2)2 + (5)2 ]
5
1 54
= [16 + 9 + 4 + 25] =
5 5
𝜎 2 = 10.8

𝜎 = √10.8 = 3.2863
(𝑖𝑖𝑖) 𝑆𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑖𝑠 {(2,2), (2,3), (2,6), (2,8), (2,11), (3,2), (3,3), (3,6), (3,8), (3,11),
(6,2), (6,3), (6,6), (6,8), (6,11), (8,2), (8,3), (8,6), (8,8), (8,11), (11,2), (11,3), (11,6),
(11,8), (11,11)}
𝑆𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛 𝑖𝑠
{2, 2.5, 4.5, 6.5, 2.5, 3, 4.5, 5.5, 7, 4, 4.5, 6, 7, 8.5, 5, 5.5, 7, 8, 9.5, 6.5, 7, 8.5, 9.5, 11}
𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛𝑠
1 150
= [2 + 2.5 + 4.5+ . . . . . . . . .11] = =6
25 25
(𝑖𝑣) 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛𝑠
1
= [(2 − 6)2 + (2.5 − 6)2 + (4 − 6)2 +. . . . . . +(11 − 6)2 ]
25
𝑆 2 = 5.4

𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛𝑠 = √5.4 = 2.32

3. 𝐶𝑜𝑛𝑠𝑖𝑑𝑒𝑟 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑐𝑜𝑛𝑠𝑖𝑠𝑡𝑖𝑛𝑔 𝑜𝑓 𝑡ℎ𝑒 𝑚𝑒𝑚𝑏𝑒𝑟𝑠 5,10,14,18,13,24 . 𝐶𝑜𝑛𝑠𝑖𝑑𝑒𝑟 𝑎𝑙𝑙
𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓 𝑠𝑖𝑧𝑒 2 𝑤𝑖𝑡ℎ 𝑜𝑢𝑡 𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑎𝑛𝑑 𝑡ℎ𝑒𝑛 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑡ℎ𝑒 𝑓𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔
(𝑖) 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 (𝑖𝑖)𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
(𝑖𝑖𝑖) 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛𝑠 (𝑖𝑣) 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒
𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛𝑠.
5 + 10 + 14 + 18 + 13 + 24
𝑆𝑜𝑙: (𝑖) 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝜇 = = 14
6
6
1
(𝑖𝑖) 𝜎 = ∑(𝑥𝑖 − 𝜇)2
2
6
𝑖=1

1
= [(5 − 14)2 + (10 − 14)2 + (14 − 14)2 + (18 − 14)2 + (13 − 14)2 + (24 − 14)2 ]
6
1
= [(−9)2 + (−4)2 + 0 + (4)2 + (1)2 + (10)2 ]
6
1
= [81 + 16 + 16 + 1 + 100]
6
𝜎 2 = 35.6

𝜎 = √35.6 = 5.96
(𝑖𝑖𝑖) 𝑆𝑎𝑚𝑝𝑙𝑒 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓 𝑠𝑖𝑧𝑒 2
= {(5,10), (5,14), (5,18), (5,13), (5,24), (10,14), (10,18), (10,13), (10,24), (14,18),
(14,13), (14,24), (18,13), (18,24), (13,24)}
𝑆𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝐷𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛 𝑖𝑠
{7.5, 9.5, 11.5, 9, 14.5, 12, 14, 11.5, 17, 16, 13.5, 19, 15.5, 21, 18.5}
𝑀𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑎𝑛𝑠
7.5 + 9.5 + 11.5+ . . . . . +18.5
= = 14
15
(𝑖𝑣) 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛𝑠
1
= [(7.5 − 14)2 + (9.5 − 14)2 + (11.5 − 14)2 +. . . . . . +(18.5 − 14)2 ]
15
𝑆 2 = 14.26
𝑆 = 3.77
4. 𝐶𝑜𝑛𝑠𝑖𝑑𝑒𝑟 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑖𝑧𝑒 4 𝑤𝑖𝑡ℎ 𝑡ℎ𝑒 𝑚𝑒𝑚𝑏𝑒𝑟𝑠 1,5,6,8 𝑓𝑟𝑜𝑚 𝑤ℎ𝑖𝑐ℎ 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓
𝑠𝑖𝑧𝑒 2 𝑎𝑟𝑒 𝑑𝑟𝑎𝑤𝑛. 𝐸𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑡ℎ𝑒 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑒𝑟𝑟𝑜𝑟.
1 + 5 + 6 + 8 20
𝑆𝑜𝑙: 𝜇 = = =5
4 4
 4
1
𝜎 2 = ∑(𝑥𝑖 − 𝜇)2
𝑁
𝑖=1

1
= [(1 − 5)2 + (5 − 5)2 + (6 − 5)2 + (8 − 5)2 ]
4
1 26
= [16 + 1 + 9] = = 6.5
4 4
𝜎 2.54
∴ 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑒𝑟𝑟𝑜𝑟 = = = 1.79
√𝑛 √2
5. 𝑊ℎ𝑎𝑡 𝑖𝑠 𝑡ℎ𝑒 𝑒𝑓𝑓𝑒𝑐𝑡 𝑜𝑛 𝑡ℎ𝑒 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑒𝑟𝑟𝑜𝑟𝑠 𝑖𝑓 𝑎 𝑠𝑎𝑚𝑝𝑙𝑒 𝑖𝑠 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝑤ℎ𝑒𝑟𝑒 𝑖𝑡𝑠 𝑠𝑖𝑧𝑒 𝑖𝑠 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒𝑑 𝑓𝑟𝑜𝑚 400 𝑡𝑜 900
1
𝑆𝑜𝑙: 𝑆. 𝐸 ∝
√𝑛

𝑆. 𝐸1 𝑛2
=√
𝑆. 𝐸2 𝑛1

900 30
=√ = √ = 1.5
400 20

𝑆. 𝐸1 = (1.5) 𝑆. 𝐸2
𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒊𝒐𝒏:
𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒆:
𝐴𝑛𝑦 𝑠𝑎𝑡𝑒𝑚𝑒𝑛𝑡 𝑡ℎ𝑎𝑡 𝑐𝑎𝑛 𝑏𝑒 𝑚𝑎𝑑𝑒 𝑎𝑏𝑜𝑢𝑡 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟 𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝑎𝑠 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒.
𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒐𝒓:
𝑇ℎ𝑒 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 𝑎𝑛𝑑 𝑖𝑡𝑠 𝑣𝑎𝑙𝑢𝑒 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑢𝑠𝑒𝑑 𝑡𝑜 𝑒𝑣𝑎𝑙𝑢𝑎𝑡𝑒 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟 𝑖𝑠
𝑐𝑎𝑙𝑙𝑒𝑑 𝑎𝑛 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟.
𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒊𝒐𝒏:
𝑇ℎ𝑒 𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒 𝑜𝑟 𝑡ℎ𝑒 𝑠𝑒𝑡 𝑜𝑓 𝑟𝑢𝑙𝑒𝑠 𝑤ℎ𝑖𝑐ℎ 𝑎𝑟𝑒 𝑢𝑠𝑒𝑑 𝑡𝑜 𝑚𝑒𝑎𝑠𝑢𝑟𝑒 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟 𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑖𝑜𝑛. 𝐼𝑡 𝑖𝑠 𝑜𝑓 𝑡𝑤𝑜 𝑡𝑦𝑝𝑒𝑠.
(𝒊) 𝑷𝒐𝒊𝒏𝒕 𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒊𝒐𝒏:
𝐼𝑓 𝑎 𝑠𝑖𝑛𝑔𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑖𝑠 𝑢𝑠𝑒𝑑 𝑎𝑠 𝑎𝑛 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝑡𝑜 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟, 𝑡ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠
𝑐𝑎𝑙𝑙𝑒𝑑 𝑃𝑜𝑖𝑛𝑡 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑖𝑜𝑛.
(𝒊𝒊) 𝑰𝒏𝒕𝒆𝒓𝒗𝒂𝒍 𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒊𝒐𝒏:
𝐼𝑓 𝑎 𝑟𝑎𝑛𝑔𝑒 𝑜𝑓 𝑣𝑎𝑙𝑢𝑒𝑠 𝑎𝑟𝑒 𝑠𝑝𝑒𝑐𝑖𝑓𝑦 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑤ℎ𝑖𝑐ℎ 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟 𝑙𝑖𝑒𝑠,
𝑡ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑖𝑜𝑛.
𝑷𝒓𝒐𝒑𝒆𝒓𝒕𝒊𝒆𝒔 𝒐𝒇 𝑮𝒐𝒐𝒅 𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒐𝒓:
𝐴𝑛 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝜃 𝑐 𝑜𝑓 𝜃 𝑖𝑠 𝑠𝑎𝑖𝑑 𝑡𝑜 𝑏𝑒 𝐺𝑜𝑜𝑑 𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝑖𝑓 𝑖𝑡 𝑠𝑎𝑡𝑖𝑠𝑓𝑦 𝑡ℎ𝑒 𝑓𝑜𝑙𝑙𝑜𝑤𝑖𝑛𝑔 𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑖𝑒𝑠.
→ 𝑪𝒐𝒏𝒔𝒊𝒔𝒕𝒆𝒏𝒄𝒚: 𝐴𝑛 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝜃 𝑐 𝑜𝑓 𝜃 𝑖𝑠 𝑠𝑎𝑖𝑑 𝑡𝑜 𝑏𝑒 𝑐𝑜𝑛𝑠𝑖𝑠𝑡𝑒𝑛𝑡 𝑖𝑓 𝜃 𝑐 → 𝜃 𝑎𝑠 𝑛 → ∞
→ 𝑼𝒏𝒃𝒊𝒂𝒔𝒆𝒅: 𝐴𝑛 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝜃 𝑐 𝑜𝑓 𝜃 𝑖𝑠 𝑠𝑎𝑖𝑑 𝑡𝑜 𝑏𝑒 𝑎𝑛 𝑢𝑛𝑏𝑖𝑎𝑠𝑒𝑑 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝑜𝑓 𝜃

𝐸[𝜃̂] = 𝜃

→ 𝑬𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕: 𝐿𝑒𝑡 𝜃̂1 , 𝜃̂2 𝑎𝑟𝑒 𝑡ℎ𝑒 𝑡𝑤𝑜 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟𝑠 𝑜𝑓 𝜃, 𝑡ℎ𝑒𝑛 𝜃̂1 𝑖𝑠 𝑠𝑎𝑖𝑑 𝑡𝑜 𝑏𝑒 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡

𝑡ℎ𝑎𝑛 𝜃̂2 𝑖𝑓 𝑉(𝜃̂1 ) < 𝑉(𝜃̂2 )

→ 𝑺𝒖𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕: 𝐴𝑛 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝜃̂ 𝑜𝑓 𝜃 𝑖𝑠 𝑠𝑎𝑖𝑑 𝑡𝑜 𝑏𝑒 𝑎 𝑠𝑢𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝑜𝑓 𝜃, 𝑖𝑓 𝑖𝑡

𝑐𝑜𝑛𝑡𝑎𝑖𝑛𝑠 𝑎𝑙𝑙 𝑡ℎ𝑒 𝑖𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑎𝑏𝑜𝑢𝑡 𝜃.
𝜎
𝐸𝑟𝑟𝑜𝑟 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑟𝑜𝑐𝑒𝑠𝑠 𝑜𝑓 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑖𝑜𝑛 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦 𝐸𝑚𝑎𝑥 = 𝑧𝛼⁄2
√𝑛
𝐸𝑚𝑎𝑥 . √𝑛 = 𝑧𝛼⁄2 . 𝜎
𝑧𝛼⁄2 . 𝜎
√𝑛 =
𝐸𝑚𝑎𝑥
𝑧𝛼⁄2 . 𝜎 2
𝑛=[ ]
𝐸𝑚𝑎𝑥
1. 𝐹𝑖𝑛𝑑 𝑡ℎ𝑒 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑
𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 20, 𝑧𝛼⁄2 = 1.96 𝑤ℎ𝑒𝑟𝑒 𝑡ℎ𝑒 𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑒𝑟𝑟𝑜𝑟 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒

𝑑𝑢𝑟𝑖𝑛𝑔 𝑡ℎ𝑒 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑖𝑜𝑛 𝑖𝑠 3.

𝑆𝑜𝑙: 𝑧𝛼⁄2 = 1.96 ; 𝐸𝑚𝑎𝑥 = 3 𝑎𝑛𝑑 𝜎 = 20

1.96 × 20 2
𝑛=[ ]
3
= 170.76 ≅ 171
𝑻𝒆𝒔𝒕𝒊𝒏𝒈 𝑯𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔: 𝑰
𝑯𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔: 𝐴𝑛𝑦 𝑠𝑡𝑎𝑡𝑒𝑚𝑒𝑛𝑡 𝑡ℎ𝑎𝑡 𝑐𝑎𝑛 𝑏𝑒 𝑚𝑎𝑑𝑒 𝑎𝑏𝑜𝑢𝑡 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟
𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝑎 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠.
𝐼𝑡 𝑖𝑠 𝑜𝑓 𝑡𝑤𝑜 𝑡𝑦𝑝𝑒𝑠:
(𝟏) 𝑵𝒖𝒍𝒍 𝒉𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔:
𝐼𝑡 𝑖𝑠 𝑑𝑒𝑛𝑜𝑡𝑒𝑑 𝑏𝑦 𝐻0 𝑎𝑛𝑑 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑒𝑠 𝑡ℎ𝑒 𝑠𝑡𝑎𝑡𝑒𝑚𝑒𝑛𝑡 𝑡ℎ𝑎𝑡 𝑡ℎ𝑒𝑟𝑒 𝑖𝑠 𝑛𝑜 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒
𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑎𝑛𝑑 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟𝑠.
𝐸𝑥: 𝐻0 : 𝜇 = 64
(𝟐) 𝑨𝒍𝒕𝒆𝒓𝒏𝒂𝒕𝒊𝒗𝒆 𝒉𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔:
𝐼𝑡 𝑖𝑠 𝑑𝑒𝑛𝑜𝑡𝑒𝑑 𝑏𝑦 𝐻1 𝑎𝑛𝑑 𝑖𝑠 𝑑𝑒𝑓𝑖𝑛𝑒𝑑 𝑖𝑛 𝑐𝑜𝑛𝑡𝑟𝑎𝑑𝑖𝑐𝑡𝑖𝑜𝑛 𝑡𝑜 𝑡ℎ𝑒 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠.
𝐸𝑥: 𝐻1 : 𝜇 ≠ 64(2 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
𝐻1 : 𝜇 > 64 (𝑅𝑖𝑔ℎ𝑡 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
𝐻1 : 𝜇 < 64 (𝐿𝑒𝑓𝑡 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡)
𝑇ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑐𝑎𝑛 𝑏𝑒 𝑐𝑙𝑎𝑠𝑠𝑖𝑓𝑖𝑒𝑑 𝑖𝑛𝑡𝑜 𝑡𝑤𝑜 𝑡𝑦𝑝𝑒𝑠
1. 𝑆𝑚𝑎𝑙𝑙 𝑆𝑎𝑚𝑝𝑙𝑒𝑠 (𝑛 < 30)
2. 𝐿𝑎𝑟𝑔𝑒 𝑆𝑎𝑚𝑝𝑙𝑒𝑠 (𝑛 > 30)
𝑻𝒆𝒔𝒕𝒊𝒏𝒈 𝑷𝒓𝒐𝒄𝒆𝒅𝒖𝒓𝒆:
𝟏. 𝑵𝒖𝒍𝒍 𝑯𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔
𝟐. 𝑨𝒍𝒕𝒆𝒓𝒏𝒂𝒕𝒊𝒗𝒆 𝑯𝒚𝒑𝒐𝒕𝒉𝒆𝒔𝒊𝒔
𝟑. 𝑳𝒆𝒗𝒆𝒍 𝒐𝒇 𝑺𝒊𝒈𝒏𝒊𝒇𝒊𝒄𝒂𝒏𝒄𝒆:
𝐼𝑡 𝑖𝑠 𝑑𝑒𝑛𝑜𝑡𝑒𝑑 𝑏𝑦 𝛼 𝑎𝑛𝑑 𝑐𝑎𝑛 𝑏𝑒 𝑔𝑖𝑣𝑒𝑛 𝑎 𝑣𝑎𝑙𝑢𝑒 𝑖𝑛 𝑐𝑜𝑛𝑡𝑟𝑎𝑑𝑖𝑐𝑡𝑖𝑜𝑛 𝑡𝑜 𝑡ℎ𝑒 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒
𝑙𝑒𝑣𝑒𝑙. 𝑇ℎ𝑒𝑟𝑒 𝑎𝑟𝑒 3 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡 𝑣𝑎𝑙𝑢𝑒𝑠 𝑓𝑜𝑟 𝛼. 𝑇ℎ𝑒𝑦 𝑎𝑟𝑒 1%, 5%, 10% 𝑤ℎ𝑖𝑐ℎ 𝑠𝑝𝑒𝑐𝑖𝑓𝑦
99%, 95%, 90% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑒𝑣𝑒𝑙.
𝟒. 𝑻𝒆𝒔𝒕 𝑺𝒕𝒂𝒕𝒊𝒔𝒕𝒊𝒄:
𝐼𝑡 𝑖𝑠 𝑑𝑒𝑛𝑜𝑡𝑒𝑑 𝑏𝑦 𝑧 𝑎𝑛𝑑 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑤𝑖𝑡ℎ 𝑎 𝑠𝑚𝑎𝑙𝑙 𝑓𝑜𝑟𝑚𝑢𝑙𝑎 𝑡ℎ𝑎𝑡 𝑣𝑎𝑟𝑖𝑒𝑠 𝑓𝑟𝑜𝑚 𝑡𝑒𝑠𝑡 𝑡𝑜 𝑡𝑒𝑠𝑡
𝟓. 𝑪𝒐𝒏𝒄𝒍𝒖𝒔𝒊𝒐𝒏:
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑠𝑡𝑒𝑝, 𝑎 𝑐𝑜𝑚𝑝𝑎𝑟𝑖𝑠𝑖𝑜𝑛 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑡ℎ𝑒 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑧 𝑎𝑛𝑑 𝑎 𝑡𝑎𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑧𝛼
𝐼𝑓 |𝑧| < 𝑧𝛼 𝑡ℎ𝑒𝑛 𝑡ℎ𝑒 𝑐𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑 (𝑜𝑟)
|𝑧| > 𝑧𝛼 𝑡ℎ𝑒𝑛 𝑡ℎ𝑒 𝑐𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑 𝑎𝑛𝑑 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
𝑻𝒚𝒑𝒆 − 𝑰 𝑬𝒓𝒓𝒐𝒓: 𝑅𝑒𝑗𝑒𝑐𝑡 𝐻0 𝑤ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠 𝑡𝑟𝑢𝑒.
𝑻𝒚𝒑𝒆 − 𝑰𝑰 𝑬𝒓𝒓𝒐𝒓: 𝐴𝑐𝑐𝑒𝑝𝑡 𝐻0 𝑤ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠 𝑓𝑎𝑙𝑠𝑒.
𝑨𝒄𝒄𝒆𝒑𝒕𝒆𝒏𝒄𝒆 𝒓𝒆𝒈𝒊𝒐𝒏:
𝑇ℎ𝑒 𝑟𝑒𝑔𝑖𝑜𝑛 𝑡ℎ𝑎𝑡 𝑝𝑟𝑜𝑣𝑖𝑑𝑒𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑛𝑐𝑒 𝑡𝑜 𝑡ℎ𝑒 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠 𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑛𝑐𝑒 𝑟𝑒𝑔𝑖𝑜𝑛
𝑪𝒓𝒊𝒕𝒊𝒄𝒂𝒍 𝒓𝒆𝒈𝒊𝒐𝒏 (𝒐𝒓) 𝑹𝒆𝒋𝒆𝒄𝒕𝒊𝒐𝒏 𝒓𝒆𝒈𝒊𝒐𝒏:
𝑇ℎ𝑒 𝑟𝑒𝑔𝑖𝑜𝑛 𝑡ℎ𝑎𝑡 𝑔𝑖𝑣𝑒𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑖𝑜𝑛 𝑡𝑜 𝐻0 𝑖𝑠 𝑐𝑎𝑙𝑙𝑒𝑑 𝑡ℎ𝑒 𝑐𝑟𝑖𝑡𝑖𝑐𝑎𝑙 𝑟𝑒𝑔𝑖𝑜𝑛.
𝑧𝛼 𝑣𝑎𝑙𝑢𝑒𝑠
 Level of significance
1% 5% 10%
Two-Tailed Test 2.58 1.96 1.645
Right-Tailed Test 2.33 1.645 1.28
Left-Tailed Test -2.33 -1.645 -1.28

𝑇𝑒𝑠𝑡 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑜𝑠 𝑠𝑖𝑛𝑔𝑙𝑒 𝑚𝑒𝑎𝑛:

𝑥−𝜇
𝑧=𝜎
⁄ 𝑛
√
𝑥−𝜇
𝐼𝑓 𝜎 𝑖𝑠 𝑛𝑜𝑡 𝑘𝑛𝑜𝑤𝑛, 𝑧 = 𝑠
⁄ 𝑛
√
𝑃𝑟𝑜𝑏𝑙𝑒𝑚𝑠:
1. 𝐴𝑐𝑐𝑜𝑟𝑑𝑖𝑛𝑔 𝑡𝑜 𝑡ℎ𝑒 𝑛𝑜𝑟𝑚𝑠 𝑒𝑠𝑡𝑎𝑏𝑙𝑖𝑠ℎ𝑒𝑑 𝑓𝑜𝑟 𝑎 𝑚𝑒𝑐ℎ𝑎𝑛𝑖𝑐𝑎𝑙 𝑎𝑝𝑡𝑖𝑡𝑢𝑑𝑒 𝑡𝑒𝑠𝑡 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑝𝑒𝑟𝑠𝑜𝑛𝑠
𝑤𝑖𝑡ℎ 18 𝑦𝑒𝑎𝑟𝑠 𝑜𝑓 𝑎𝑔𝑒 𝑝𝑟𝑜𝑣𝑒𝑑 𝑡ℎ𝑎𝑡, 𝑡ℎ𝑒𝑦 ℎ𝑎𝑣𝑒 𝑎𝑛 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 ℎ𝑒𝑖𝑔ℎ𝑡 𝑜𝑓 73.2 𝑤𝑖𝑡ℎ 𝑎 𝑆. 𝐷 𝑜𝑓
8.6. 𝐼𝑓 4 𝑟𝑎𝑛𝑑𝑜𝑚𝑙𝑦 𝑠𝑒𝑙𝑒𝑐𝑡𝑒𝑑 𝑝𝑒𝑟𝑠𝑜𝑛𝑠 𝑤𝑒𝑟𝑒 𝑡𝑒𝑠𝑡𝑒𝑑, 𝑡ℎ𝑒𝑟𝑒 ℎ𝑒𝑖𝑔ℎ𝑡 𝑤𝑎𝑠 𝑎𝑣𝑒𝑟𝑎𝑔𝑒𝑑 𝑎𝑠 76.7.
𝑇𝑒𝑠𝑡 𝑡ℎ𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠 𝜇 = 73.2 𝑎𝑔𝑎𝑖𝑛𝑠𝑡 𝑡ℎ𝑒 𝑎𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠, 𝜇 > 73.2 𝑎𝑡 1%
𝑙𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒.
𝑆𝑜𝑙: 𝜇 = 73.2 , 𝜎 = 8.6 , 𝑥 = 16.7; 𝛼 = 1% 𝑎𝑛𝑑 𝑛 = 4
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇 = 73.2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇 > 73.2
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 1%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑥−𝜇
𝑧=𝜎
⁄ 𝑛
√
76.7 − 73.2
= = 0.81
8.6⁄
√4
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 1% 𝑙𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒(𝐿. 𝑂. 𝑆) 𝑓𝑜𝑟 𝑎 𝑟𝑖𝑔ℎ𝑡 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 2.33
|𝑧| = 0.81
|𝑧| < 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
𝜇 = 73.2
𝑇ℎ𝑒 𝑡𝑟𝑢𝑒 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛
𝑥−𝜇
𝑎𝑟𝑒 𝑧 = 𝜎
⁄ 𝑛
√
𝑥−𝜇
𝑧=𝑠
⁄ 𝑛
√
𝜎
1) 95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑎𝑟𝑒 𝑥 ± (1.96)
√𝑛
𝜎
2) 98% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑎𝑟𝑒 𝑥 ± (2.33)
√𝑛
𝜎
3) 99% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑎𝑟𝑒 𝑥 ± (2.58)
√𝑛
2. 𝐶𝑜𝑛𝑠𝑖𝑑𝑒𝑟 𝑡ℎ𝑒 𝑟𝑎𝑛𝑑𝑜𝑚 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓 𝑠𝑖𝑧𝑒 400 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓
𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 10. 𝐼𝑡 𝑖𝑠 𝑜𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝑡ℎ𝑎𝑡 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑚𝑒𝑎𝑛 𝑖𝑠 40. 𝑇𝑒𝑠𝑡 𝑡ℎ𝑒 𝑐𝑙𝑎𝑖𝑚,
𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑡ℎ𝑖𝑠 𝑠𝑎𝑚𝑝𝑙𝑒 𝑐𝑎𝑛 𝑏𝑒 𝑟𝑒𝑔𝑎𝑟𝑑𝑒𝑑 𝑎𝑠 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝑎 𝑚𝑒𝑎𝑛 𝑜𝑓 38.
𝐴𝑙𝑠𝑜 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑏𝑡ℎ𝑒 95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛.
𝑆𝑜𝑙: 𝑛 = 400; 𝜎 = 10; 𝑥 = 40 𝑎𝑛𝑑 𝜇 = 38
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇 = 38
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇 ≠ 38[𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑥−𝜇
𝑧=𝜎
⁄ 𝑛
√
40 − 38
= = 2×2= 4
10⁄√400
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑟𝑖𝑔ℎ𝑡 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 4
|𝑧| > 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
⇒ 𝜇 ≠ 38
∴ 𝑇ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑖𝑠 𝑛𝑜𝑡 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑤𝑖𝑡ℎ 𝑚𝑒𝑎𝑛 38
𝜎
95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑓𝑜𝑟 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛 𝑥 ± (1.96)
√𝑛
𝜎
𝐿𝑜𝑤𝑒𝑟 𝑙𝑖𝑚𝑖𝑡 𝑖𝑠 𝑥 − (1.96)
√𝑛
10
= 40 − (1.96) ( )
√400
= 39.02
𝜎
𝑈𝑝𝑝𝑒𝑟 𝑙𝑖𝑚𝑖𝑡 𝑖𝑠 𝑥 + (1.96)
√𝑛
10
= 40 + (1.96) ( )
√400
= 40.98
𝑇ℎ𝑒 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙 𝑖𝑠 (39.02, 40.98)
3. 𝐴𝑛 𝑎𝑚𝑏𝑢𝑙𝑎𝑛𝑐𝑒 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑐𝑙𝑎𝑖𝑚𝑠 𝑡ℎ𝑎𝑡 𝑜𝑛 𝑎𝑛 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑖𝑡 𝑡𝑎𝑘𝑒𝑠 𝑙𝑒𝑠𝑠 𝑡ℎ𝑎𝑛 10 min 𝑡𝑜 𝑎𝑡𝑡𝑒𝑛𝑑
𝑎𝑛 𝑒𝑚𝑒𝑟𝑔𝑒𝑛𝑐𝑦 𝑐𝑎𝑙𝑙. 𝐴 𝑠𝑎𝑚𝑝𝑙𝑒 𝑜𝑓 36 𝑐𝑎𝑙𝑙𝑠 ℎ𝑎𝑠 𝑎𝑚𝑒𝑎𝑛 𝑜𝑓 11𝑚𝑖𝑛𝑠 𝑤𝑖𝑡ℎ 𝑎 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑜𝑓 16
𝑚𝑖𝑛𝑠. 𝑇𝑒𝑠𝑡 𝑎𝑡 5% 𝑙𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑡ℎ𝑒 𝑐𝑙𝑎𝑖𝑚 𝑖𝑠 𝑡𝑟𝑢𝑒.
𝑆𝑜𝑙: 𝑛 = 36 ; 𝑠 2 = 16 𝑎𝑛𝑑 𝑠 = 4 ; 𝑥 = 11 𝑎𝑛𝑑 𝜇 = 10
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇 < 10
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇 > 38[𝑟𝑖𝑔ℎ𝑡 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑥−𝜇
𝑧=𝑠
⁄ 𝑛
√
11 − 10 1 6
= = = = 1.52
4⁄√36 4⁄6 4

𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑟𝑖𝑔ℎ𝑡 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.645
|𝑧| = 1.5
|𝑧| < 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
∴ 𝑇ℎ𝑒 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑡𝑖𝑚𝑒 𝑡𝑎𝑘𝑒𝑛 𝑏𝑦 𝑎𝑚𝑏𝑢𝑙𝑎𝑛𝑐𝑒 𝑡𝑜 𝑎𝑡𝑡𝑒𝑛𝑑 𝑒𝑚𝑒𝑟𝑔𝑒𝑛𝑐𝑦 𝑐𝑎𝑙𝑙 𝑙𝑒𝑠𝑠 𝑡ℎ𝑎𝑛 10 𝑚𝑖𝑛.
4. 𝑇ℎ𝑒 𝑚𝑒𝑎𝑛 𝑙𝑖𝑓𝑒 𝑡𝑖𝑚𝑒 𝑜𝑓 100 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐 𝑏𝑢𝑙𝑏𝑠 𝑝𝑟𝑜𝑑𝑢𝑐𝑒𝑑 𝑏𝑦 𝑎 𝑐𝑜𝑚𝑝𝑎𝑛𝑦 𝑖𝑠 1560 ℎ𝑜𝑢𝑟𝑠 𝑤𝑖𝑡ℎ
𝑎 𝑆. 𝐷 𝑜𝑓 90 ℎ𝑜𝑢𝑟𝑠. 𝐶𝑎𝑛 𝑡ℎ𝑖𝑠 𝑏𝑒 𝑟𝑒𝑔𝑎𝑟𝑑𝑒𝑑 𝑎𝑠 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐
𝑏𝑢𝑙𝑏𝑠 𝑤𝑖𝑡ℎ 𝑚𝑒𝑎𝑛 𝑙𝑖𝑓𝑒 𝑡𝑖𝑚𝑒 𝑜𝑓 1580 ℎ𝑜𝑢𝑟𝑠.
𝑆𝑜𝑙: 𝐺𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝜇 = 1580; 𝑛 = 100; 𝑥 = 1560 & 𝑆 = 90
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇 = 1580
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇 ≠ 1580[𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑥−𝜇
𝑧=𝜎
⁄ 𝑛
√
1560 − 1580 1
= = −20 × = −2.2
90⁄√100 9

𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 2.2
|𝑧| > 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
⇒ 𝜇 ≠ 1580
∴ 𝑇ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑖𝑠 𝑛𝑜𝑡 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 1580
𝑻𝒆𝒔𝒕 𝒐𝒇 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒄𝒂𝒏𝒄𝒆 𝒇𝒐𝒓 𝒅𝒊𝒇𝒇𝒆𝒓𝒆𝒏𝒄𝒆 𝒐𝒇 𝒎𝒆𝒏𝒂𝒔:
𝐶𝑜𝑛𝑠𝑖𝑑𝑒𝑟 𝑡𝑤𝑜 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑤𝑖𝑡ℎ 𝑡ℎ𝑒 𝑔𝑖𝑣𝑒𝑛 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟. 𝑇ℎ𝑒 𝑡𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦
𝑥1 − 𝑥2
𝑧=
𝜎2 𝜎2
√ 1 + 2
𝑛1 𝑛2
𝑥1 − 𝑥2
𝐼𝑓 𝜎 𝑖𝑠 𝑛𝑜𝑡 𝑘𝑛𝑜𝑤𝑛, 𝑖. 𝑒. , 𝑟𝑒𝑝𝑙𝑎𝑐𝑒 𝜎1 𝑤𝑖𝑡ℎ 𝑠1 𝑎𝑛𝑑 𝜎2 𝑤𝑖𝑡ℎ 𝑠2 𝑡ℎ𝑒𝑛 𝑧 =
𝑠1 2 𝑠2 2
√
𝑛1 + 𝑛2
𝐼𝑓 𝑏𝑜𝑡ℎ 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑎𝑟𝑒 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛, 𝑡ℎ𝑒𝑛 𝜎1 2 = 𝜎2 2 𝑖. 𝑒. , 𝜎 2
𝑥1 − 𝑥2
𝑡ℎ𝑒𝑛 𝑧 =
𝜎2 𝜎2
√ +
𝑛1 𝑛2
𝑥1 − 𝑥2
=
1 1
𝜎√ +
𝑛1 𝑛2
1. 𝑇ℎ𝑒 𝑚𝑒𝑎𝑛 𝑜𝑓 𝑡𝑤𝑜 𝑙𝑎𝑟𝑔𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑜𝑓 𝑠𝑖𝑧𝑒𝑠 1000 𝑎𝑛𝑑 2000 ℎ𝑎𝑣𝑒 𝑡ℎ𝑒 𝑚𝑒𝑎𝑛 𝑣𝑎𝑙𝑢𝑒𝑠 67.5
𝑎𝑛𝑑 68 𝑟𝑒𝑠𝑝𝑒𝑐𝑡𝑖𝑣𝑒𝑙𝑦. 𝐶𝑎𝑛 𝑡ℎ𝑒𝑠𝑒 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑏𝑒 𝑟𝑒𝑔𝑎𝑟𝑑𝑒𝑑 𝑎𝑠 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛
𝑤𝑖𝑡ℎ 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 2.5
𝑆𝑜𝑙: 𝐺𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝑛1 = 1000; 𝑛2 = 2000
𝑥1 = 67.5 ; 𝑥2 = 68 𝑎𝑛𝑑 𝜎 = 2.5
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇1 = 𝜇2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇1 ≠ 𝜇2 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
 𝑥1 − 𝑥2
𝑧=
1 1
𝜎√𝑛 + 𝑛
1 2

67.5 − 68
= = −5.164
1 1
2.5 [√1000 + 2000]

𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 5.164
|𝑧| > 𝑧2 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
∴ 𝑇ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑎𝑟𝑒 𝑛𝑜𝑡 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛.
2. 𝐼𝑛 𝑎 𝑠𝑢𝑟𝑣𝑒𝑦 𝑜𝑓 𝑏𝑢𝑦𝑖𝑛𝑔 400 𝑤𝑜𝑚𝑒𝑛 𝑠ℎ𝑜𝑝𝑝𝑒𝑟𝑠 𝑎𝑟𝑒 𝑐ℎ𝑜𝑜𝑠𝑒𝑛 𝑎𝑡 𝑟𝑎𝑛𝑑𝑜𝑚 𝑎𝑡 𝑠𝑢𝑝𝑒𝑟 𝑚𝑎𝑟𝑘𝑒𝑡
𝑙𝑜𝑐𝑎𝑡𝑒𝑑 𝑎𝑡 𝑎 𝑐𝑒𝑟𝑡𝑎𝑖𝑛 𝑝𝑙𝑎𝑐𝑒 𝑖𝑛 𝑎 𝑐𝑖𝑡𝑦. 𝑇ℎ𝑒𝑖𝑟 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑤𝑒𝑒𝑘𝑙𝑦 𝑓𝑜𝑜𝑑 𝑒𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑖𝑠 250 𝑟𝑢𝑝𝑒𝑒𝑠
𝑤𝑖𝑡ℎ 𝑎 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 40 𝑟𝑢𝑝𝑒𝑒𝑠. 𝐴𝑛𝑜𝑡ℎ𝑒𝑟 𝑔𝑟𝑜𝑢𝑝 𝑜𝑓 600 𝑤𝑜𝑚𝑒𝑛 𝑠ℎ𝑜𝑝𝑝𝑒𝑟 𝑤𝑒𝑟𝑒
𝑐ℎ𝑜𝑜𝑠𝑒𝑛 𝑎𝑡 𝑟𝑎𝑛𝑑𝑜𝑚 𝑖𝑛 𝑎 𝑠𝑢𝑝𝑒𝑟 𝑏𝑎𝑧𝑎𝑟 𝑙𝑜𝑐𝑎𝑡𝑒𝑑 𝑎𝑡 𝑜𝑡ℎ𝑒𝑟 𝑝𝑎𝑟𝑡 𝑜𝑓 𝑐𝑖𝑡𝑦. 𝑇ℎ𝑒𝑖𝑟 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑤𝑒𝑒𝑘𝑒𝑙𝑦
𝑓𝑜𝑜𝑑 𝑒𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑖𝑠 220 𝑟𝑢𝑝𝑒𝑒𝑠 𝑤𝑖𝑡ℎ 𝑎 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 55 𝑟𝑢𝑝𝑒𝑒𝑠. 𝑇𝑒𝑠𝑡 𝑎𝑡 1% 𝑙𝑒𝑣𝑒𝑙
𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒, 𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑡ℎ𝑒 𝑡𝑤𝑜 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑠 ℎ𝑎𝑣𝑒 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑤𝑒𝑒𝑘𝑒𝑙𝑦 𝑓𝑜𝑜𝑑
𝑒𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒.
𝑆𝑜𝑙: 𝐺𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝑥1 = 250 ; 𝑛1 = 400 𝑎𝑛𝑑 𝑠1 = 40
𝑥2 = 220 ; 𝑛2 = 600 𝑎𝑛𝑑 𝑠2 = 55
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇1 = 𝜇2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇1 ≠ 𝜇2 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 1%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
 𝑥1 − 𝑥2
𝑧=
𝑠 2 𝑠 2
√ 1 + 2
𝑛1 𝑛2
250 − 220
= = 9.97
2 2
√(40) + (55)
400 600
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 1% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 2.58
|𝑧| = 9.97
|𝑧| > 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
∴ 𝑇ℎ𝑒 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑤𝑒𝑒𝑘𝑒𝑙𝑦 𝑓𝑜𝑜𝑑 𝑒𝑥𝑝𝑒𝑛𝑑𝑖𝑡𝑢𝑟𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑡𝑤𝑜 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑠 𝑎𝑟𝑒 𝑛𝑜𝑡 𝑖𝑑𝑒𝑛𝑡𝑖𝑐𝑎𝑙.
3. 𝑇𝑤𝑜 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑎𝑟𝑒 𝑑𝑟𝑎𝑤𝑛 𝑤𝑖𝑡ℎ 𝑡ℎ𝑒 𝑠𝑖𝑧𝑒𝑠 250 𝑎𝑛𝑑 300 𝑤𝑖𝑡ℎ 𝑚𝑒𝑎𝑛𝑠 63.2 𝑎𝑛𝑑 75.4
𝑟𝑒𝑠𝑝𝑒𝑐𝑡𝑖𝑣𝑒𝑙𝑦. 𝑇ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑖𝑠 23.7. 𝑉𝑒𝑟𝑖𝑓𝑦 𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑡ℎ𝑒
𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑎𝑟𝑒 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑠𝑎𝑚𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 10 % 𝑙𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒.
𝑆𝑜𝑙: 𝐺𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝑛1 = 250; 𝑛2 = 300
𝑥1 = 63.2 ; 𝑥2 = 75.4 𝑎𝑛𝑑 𝜎 = 23.7
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝜇1 = 𝜇2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝜇1 ≠ 𝜇2 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 10%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑥1 − 𝑥2
𝑧=
1 1
𝜎√𝑛 + 𝑛
1 2

63.2 − 75.4
= = −6.01
1 1
23.7 [√ + ]
250 300
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 10% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.6458
|𝑧| = 6.01
|𝑧| > 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
𝑻𝒆𝒔𝒕 𝒐𝒇 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒄𝒂𝒏𝒄𝒆 𝒇𝒐𝒓 𝒔𝒊𝒏𝒈𝒍𝒆 𝒑𝒓𝒐𝒑𝒐𝒓𝒕𝒊𝒐𝒏 (𝑳𝒂𝒓𝒈𝒆 𝑺𝒂𝒎𝒑𝒍𝒆𝒔)
𝐴 𝑠𝑖𝑛𝑔𝑙𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑓𝑟𝑜𝑚 𝑎 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑑𝑟𝑎𝑤𝑛 𝑤ℎ𝑒𝑟𝑒 𝑝 𝑑𝑒𝑛𝑜𝑡𝑒𝑠 𝑡ℎ𝑒 𝑠𝑎𝑚𝑝𝑙𝑒 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛
𝑃 𝑑𝑒𝑛𝑜𝑡𝑒𝑠 𝑡ℎ𝑒 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑎𝑛𝑑 𝑝 + 𝑞 = 1
𝑃+𝑄 =1
𝑝−𝑃
𝑇ℎ𝑒 𝑡𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦 𝑧 =
√𝑃𝑄
𝑛
𝑇ℎ𝑒 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑎𝑟𝑒

𝑃𝑄
1) 𝑎𝑡 95%, 𝑝 ± (1.96)√
𝑛

𝑃𝑄
2) 𝑎𝑡 98%, 𝑝 ± (2.33)√
𝑛

𝑃𝑄
3) 𝑎𝑡 99%, 𝑝 ± (2.58)√
𝑛

1. 𝐴 𝑚𝑎𝑛𝑢𝑓𝑎𝑐𝑡𝑢𝑟𝑒𝑟 𝑐𝑙𝑎𝑖𝑚𝑠 𝑡ℎ𝑎𝑡 95% 𝑜𝑓 ℎ𝑖𝑠 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑠 𝑎𝑟𝑒 𝑢𝑝𝑡𝑜 𝑡ℎ𝑒 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛𝑠. 𝐴 𝑠𝑎𝑚𝑝𝑙𝑒
𝑜𝑓 𝑡𝑤𝑜 ℎ𝑢𝑛𝑑𝑟𝑒𝑑 𝑖𝑡𝑒𝑚𝑠 𝑤𝑒𝑟𝑒 𝑝𝑖𝑐𝑘𝑒𝑑 𝑢𝑝 𝑖𝑛 𝑤ℎ𝑖𝑐ℎ 18 𝑤𝑒𝑟𝑒 𝑓𝑎𝑢𝑙𝑡𝑦. 𝑇𝑒𝑠𝑡 𝑡ℎ𝑒 𝑐𝑙𝑎𝑖𝑚 𝑎𝑡 5% 𝑙𝑒𝑣𝑒𝑙
𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒.
𝑆𝑜𝑙: 𝑃 = 95% = 0.95
𝑄 = 1 − 0.95 = 0.05
𝑛 = 200
𝐵𝑎𝑑 𝑔𝑜𝑜𝑑𝑠 = 18
𝐺𝑜𝑜𝑑 𝑜𝑛𝑒𝑠 = 200 − 18 = 182
182
𝑝= = 0.91
200
𝑞 = 1 − 0.91 = 0.09
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝑃 = 0.95
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝑃 ≠ 0.95 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑝−𝑃 0.91 − 0.95
𝑧= =
√𝑃𝑄 √(0.95)(0.05)
𝑛 200
= −2.59
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 2.59
|𝑧| > 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
∴ 𝑀𝑎𝑛𝑢𝑓𝑎𝑐𝑡𝑢𝑟𝑒𝑟 𝑐𝑙𝑎𝑖𝑚 𝑖𝑠 𝑛𝑜𝑡 𝑡𝑟𝑢𝑒.
2. 𝐼𝑛 𝑎 𝑠𝑎𝑚𝑝𝑙𝑒 𝑜𝑓 1000 𝑝𝑒𝑜𝑝𝑙𝑒 𝑓𝑟𝑜𝑚 𝐾𝑎𝑟𝑛𝑎𝑡𝑎𝑘𝑎 𝑠𝑡𝑎𝑡𝑒, 540 𝑚𝑒𝑚𝑏𝑒𝑟𝑠 𝑎𝑟𝑒 𝑟𝑖𝑐𝑒 𝑒𝑎𝑡𝑒𝑟𝑠.
𝑇ℎ𝑒 𝑟𝑒𝑚𝑎𝑖𝑛𝑖𝑛𝑔 𝑎𝑟𝑒 𝑐𝑜𝑛𝑠𝑖𝑑𝑒𝑟𝑒𝑑 𝑎𝑠 𝑤ℎ𝑒𝑎𝑡 𝑒𝑎𝑡𝑒𝑟𝑠. 𝑇𝑒𝑠𝑡 𝑡ℎ𝑒 𝑐𝑙𝑎𝑖𝑚 𝑎𝑡 1% 𝐿𝑂𝑆 𝑡ℎ𝑎𝑡
𝑡ℎ𝑒 𝑟𝑖𝑐𝑒 𝑒𝑎𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑤ℎ𝑒𝑎𝑡 𝑒𝑎𝑡𝑒𝑟𝑠 𝑎𝑟𝑒 𝑒𝑞𝑢𝑎𝑙𝑙𝑦 𝑝𝑜𝑝𝑢𝑙𝑎𝑟.
1 1
𝑆𝑜𝑙: 𝑃 = ;𝑄 =
2 2
540
𝑛 = 1000, 𝑝 = = 0.54
1000
𝑞 = 1 − 0.54 = 0.46
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
1
𝐻0 : 𝑃 =
2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
1
𝐻1 : 𝑃 ≠ [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
2
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 1%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑝−𝑃 0.54 − 0.5
𝑧= =
√𝑃𝑄 √(0.5)(0.5)
𝑛 1000
= 2.529
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 2.58
|𝑧| = 2.52
|𝑧| < 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
∴ 𝑇ℎ𝑒 𝑟𝑖𝑐𝑒 𝑒𝑎𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑤ℎ𝑒𝑎𝑡 𝑒𝑎𝑡𝑒𝑟𝑠 𝑎𝑟𝑒 𝑒𝑞𝑢𝑎𝑙𝑙𝑦 𝑝𝑜𝑝𝑢𝑙𝑎𝑟.
3. 𝐴 𝑟𝑎𝑛𝑑𝑜𝑚 𝑠𝑎𝑚𝑝𝑙𝑒 𝑜𝑓 500 𝑝𝑖𝑛𝑒𝑎𝑝𝑝𝑙𝑒 𝑤𝑒𝑟𝑒 𝑠𝑡𝑎𝑡𝑒𝑑 𝑓𝑟𝑜𝑚 𝑎 𝑙𝑎𝑟𝑔𝑒 𝑐𝑜𝑛𝑓𝑖𝑛𝑒𝑚𝑒𝑛𝑡 𝑎𝑛𝑑 65
𝑤𝑒𝑟𝑒 𝑓𝑜𝑢𝑛𝑑 𝑡𝑜 𝑏𝑒 𝑏𝑎𝑑. 𝐹𝑖𝑛𝑑 𝑡ℎ𝑒 𝑡𝑤𝑜 𝑙𝑖𝑚𝑖𝑡𝑠 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑏𝑎𝑑 𝑝𝑖𝑛𝑒𝑎𝑝𝑝𝑙𝑒𝑠 𝑖𝑛
𝑐𝑜𝑛𝑓𝑖𝑛𝑒𝑚𝑒𝑛𝑡 𝑎𝑡 95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒.
𝑆𝑜𝑙: 𝑛 = 500
65
𝑝= = 0.13
500
𝑞 = 1 − 0.13 = 0.87
𝑝𝑞
95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑎𝑟𝑒 𝑝 ± 1.96√
𝑛

𝑝𝑞 (0.13)(0.87)
𝐿𝑜𝑤𝑒𝑟 𝑙𝑖𝑚𝑖𝑡 𝑖𝑠 𝑝 − 1.96√ = 0.13 − (1.96)√ = 0.1005
𝑛 500

𝑝𝑞 (0.13)(0.87)
𝑈𝑝𝑝𝑒𝑟 𝑙𝑖𝑚𝑖𝑡 𝑖𝑠 𝑝 + 1.96√ = 0.13 + (1.96)√ = 0.1594
𝑛 500

4. 𝐼𝑛 𝑎 𝑠𝑎𝑚𝑝𝑙𝑒 𝑜𝑓 500 𝑝𝑒𝑜𝑝𝑙𝑒 𝑓𝑟𝑜𝑚 𝑎 𝑣𝑖𝑙𝑙𝑎𝑔𝑒 𝑖𝑛 𝑅𝑎𝑗𝑎𝑠𝑡ℎ𝑎𝑛, 280 𝑎𝑟𝑒 𝑓𝑜𝑢𝑛𝑑 𝑡𝑜 𝑏𝑒 𝑤ℎ𝑒𝑎𝑡
𝑒𝑎𝑡𝑒𝑟𝑠 𝑎𝑛𝑑 𝑡ℎ𝑒 𝑟𝑒𝑠𝑡 𝑎𝑟𝑒 𝑟𝑖𝑐𝑒 𝑒𝑎𝑡𝑒𝑟𝑠. 𝐶𝑎𝑛 𝑤𝑒 𝑎𝑠𝑠𝑢𝑚𝑒 𝑡ℎ𝑎𝑡 𝑏𝑜𝑡ℎ 𝑟𝑖𝑐𝑒 𝑎𝑛𝑑 𝑤ℎ𝑒𝑎𝑡 𝑒𝑎𝑡𝑒𝑟𝑠
𝑎𝑟𝑒 𝑒𝑞𝑢𝑎𝑙𝑙𝑦 𝑝𝑜𝑝𝑢𝑙𝑎𝑟 𝑖𝑛 𝑅𝑎𝑗𝑎𝑠𝑡ℎ𝑎𝑛? 𝐴𝑙𝑠𝑜 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒 𝑡ℎ𝑟𝑜𝑢𝑔ℎ 𝑡ℎ𝑒 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠 𝑓𝑜𝑟
𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑤ℎ𝑒𝑎𝑡 𝑒𝑎𝑡𝑒𝑟𝑠.
𝑆𝑜𝑙: 𝑛 = 500
65
𝑝= = 0.13
500
𝑞 = 1 − 0.13 = 0.87
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝑃 = 0.5
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝑃 ≠ 0.5 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑝−𝑃 0.56 − 0.5
𝑧= =
√𝑃𝑄 √(0.5)(0.5)
𝑛 500
= 2.68
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 2.68
|𝑧| > 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑
⇒ 𝐻1 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡𝑠

𝑃𝑄 (0.5)(0.5)
𝐿𝑜𝑤𝑒𝑟 𝑙𝑖𝑚𝑖𝑡 𝑖𝑠 𝑝 − 1.96√ = 0.56 − (1.96)√ = 0.5162
𝑛 500

𝑃𝑄 (0.5)(0.5)
𝑈𝑝𝑝𝑒𝑟 𝑙𝑖𝑚𝑖𝑡 𝑖𝑠 𝑝 + 1.96√ = 0.56 + (1.96)√ = 0.6038
𝑛 500
𝑻𝒆𝒔𝒕 𝒐𝒇 𝒔𝒊𝒈𝒏𝒊𝒇𝒊𝒄𝒂𝒏𝒄𝒆 𝒇𝒐𝒓 𝒅𝒊𝒇𝒇𝒆𝒓𝒆𝒏𝒄𝒆 𝒐𝒇 𝒑𝒓𝒐𝒑𝒐𝒓𝒕𝒊𝒐𝒏𝒔:
𝑇𝑤𝑜 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 𝑎𝑟𝑒 𝑑𝑟𝑎𝑤𝑛 𝑓𝑟𝑜𝑚 2 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛𝑠, 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦
𝑝1 − 𝑝2
𝑧=
𝑃𝑄 𝑃𝑄
√ 1 1+ 2 2
𝑛1 𝑛2
𝐼𝑓 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛𝑠 𝑃1 , 𝑄1 𝑎𝑛𝑑 𝑃2 , 𝑄2 𝑎𝑟𝑒 𝑛𝑜𝑡 𝑘𝑛𝑜𝑤𝑛, 𝑡ℎ𝑒𝑛 𝑟𝑒𝑝𝑙𝑎𝑐𝑒 𝑡ℎ𝑒𝑚
𝑤𝑖𝑡ℎ 𝑠𝑎𝑚𝑝𝑙𝑒 𝑣𝑎𝑙𝑢𝑒𝑠
𝑝1 − 𝑝2
𝑧=
𝑝1 𝑞1 𝑝2 𝑞2
√ 𝑛 + 𝑛
1 2

𝑀𝑒𝑡ℎ𝑜𝑑 𝑜𝑓 𝑝𝑜𝑜𝑙𝑖𝑛𝑔:
𝑛1 𝑝1 + 𝑛2 𝑝2
𝐼𝑛 𝑡ℎ𝑖𝑠 𝑚𝑒𝑡ℎ𝑜𝑑, 𝑙𝑒𝑡 𝑝 = 𝑡ℎ𝑒𝑛 𝑡ℎ𝑒 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦
𝑛1 + 𝑛2
𝑝1 − 𝑝2 𝑝1 − 𝑝2
𝑧= =
𝑝𝑞 𝑝𝑞 1 1
√𝑛 + 𝑛 √(𝑝𝑞) [ + ]
1 2 𝑛1 𝑛2
1. 𝐴 𝑟𝑎𝑛𝑑𝑜𝑚 𝑠𝑎𝑚𝑝𝑙𝑒 𝑜𝑓 400 𝑚𝑒𝑛 𝑎𝑛𝑑 600 𝑤𝑜𝑚𝑒𝑛 𝑤𝑒𝑟𝑒 𝑎𝑠𝑘𝑒𝑑, 𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑡ℎ𝑒𝑦 𝑤𝑜𝑢𝑙𝑑 𝑙𝑖𝑘𝑒
𝑡𝑜 ℎ𝑎𝑣𝑒 𝑓𝑙𝑦𝑜𝑣𝑒𝑟 𝑛𝑒𝑎𝑟 𝑡ℎ𝑒𝑖𝑟 𝑟𝑒𝑠𝑖𝑑𝑒𝑛𝑐𝑒. 200 𝑚𝑒𝑛 𝑎𝑛𝑑 325 𝑤𝑜𝑚𝑒𝑛 𝑤𝑒𝑟𝑒 𝑖𝑛 𝑓𝑎𝑣𝑜𝑢𝑟 𝑜𝑓 𝑡ℎ𝑒
𝑝𝑟𝑜𝑝𝑜𝑠𝑎𝑙. 𝑇𝑒𝑠𝑡 𝑡ℎ𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠 𝑡ℎ𝑎𝑡, 𝑡ℎ𝑒 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑛 𝑎𝑛𝑑 𝑤𝑜𝑚𝑒𝑛 𝑖𝑛 𝑓𝑎𝑣𝑜𝑢𝑟 𝑜𝑓
𝑡ℎ𝑒 𝑝𝑟𝑜𝑝𝑜𝑠𝑎𝑙 𝑎𝑟𝑒 𝑖𝑑𝑒𝑛𝑡𝑖𝑐𝑎𝑙 𝑎𝑡 5% 𝑙𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒.
𝑆𝑜𝑙: 𝑛1 = 400; 𝑛2 = 600
200 325
𝑝1 = = 0.5 ; 𝑝2 = = 0.5417
400 600
𝑞1 = 1 − 𝑝1 = 1 − 0.5 = 0.5
𝑞2 = 1 − 𝑝2 = 1 − 0.5417 = 0.4583
𝑛1 𝑝1 + 𝑛2 𝑝2 400(0.5) + 600(0.5417)
𝑝= = = 0.525
𝑛1 + 𝑛2 400 + 600
𝑞 = 1 − 𝑝 = 1 − 0.525 = 0.475
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝑝1 = 𝑝2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝑝1 ≠ 𝑝2 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑝1 − 𝑝2
𝑧=
1 1
√(𝑝𝑞) [
𝑛1 𝑛2 ]
+

0.5 − 0.5417
= = −1.2936
√(0.525)(0.475) [ 1 + 1 ]
400 600
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 1.2936
|𝑧| < 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑
∴ 𝑇ℎ𝑒 𝑝𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑒𝑛 𝑎𝑛𝑑 𝑤𝑜𝑚𝑒𝑛 𝑖𝑛 𝑓𝑎𝑣𝑜𝑢𝑟 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑟𝑜𝑝𝑜𝑟𝑠𝑎𝑙 𝑎𝑟𝑒 𝑖𝑑𝑒𝑛𝑡𝑖𝑐𝑎𝑙.
2. 𝑂𝑛 𝑡ℎ𝑒 𝑏𝑎𝑠𝑖𝑠 𝑜𝑓 𝑡𝑜𝑡𝑎𝑙 𝑠𝑐𝑜𝑟𝑒 200 𝑐𝑎𝑛𝑑𝑖𝑑𝑎𝑡𝑒𝑠 𝑜𝑓 𝑐𝑖𝑣𝑖𝑙 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑒𝑥𝑎𝑚 𝑎𝑟𝑒 𝑑𝑖𝑣𝑖𝑑𝑒𝑑 𝑖𝑛𝑡𝑜
𝑡𝑤𝑜 𝑔𝑟𝑜𝑢𝑝𝑠. 𝑇ℎ𝑒 𝑢𝑝𝑝𝑒𝑟 30% 𝑎𝑛𝑑 𝑡ℎ𝑒 𝑟𝑒𝑚𝑎𝑖𝑛𝑖𝑛𝑔 70% 𝑐𝑜𝑛𝑠𝑖𝑑𝑒𝑟 𝑡ℎ𝑒 1𝑠𝑡 𝑞𝑢𝑒𝑠𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒
𝑒𝑥𝑎𝑚𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑎𝑚𝑜𝑛𝑔 𝑡ℎ𝑒 1𝑠𝑡 𝑔𝑟𝑜𝑢𝑝. 40 ℎ𝑎𝑑 𝑡ℎ𝑒 𝑐𝑜𝑟𝑟𝑒𝑐𝑡 𝑎𝑛𝑠𝑤𝑒𝑟. 𝐼𝑛 𝑡ℎ𝑒 2𝑛𝑑 𝑔𝑟𝑜𝑢𝑝 80 ℎ𝑎𝑑
𝑡ℎ𝑒 𝑐𝑜𝑟𝑟𝑒𝑐𝑡 𝑎𝑛𝑠𝑤𝑒𝑟. 𝑂𝑛 𝑡ℎ𝑒 𝑏𝑎𝑠𝑖𝑠 𝑜𝑓 𝑡ℎ𝑒𝑠𝑒 𝑟𝑒𝑠𝑢𝑙𝑡𝑠, 𝑐𝑎𝑛 𝑤𝑒 𝑐𝑜𝑛𝑐𝑙𝑢𝑑𝑒 𝑡ℎ𝑎𝑡 𝑡ℎ𝑒 1𝑠𝑡 𝑞𝑢𝑒𝑠𝑡𝑖𝑜𝑛,
𝑎𝑛𝑠𝑤𝑒𝑟𝑖𝑛𝑔 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑒𝑠 𝑡ℎ𝑎𝑡, 𝑡ℎ𝑒 𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠 𝑜𝑓 𝑡ℎ𝑒 2𝑛𝑑 𝑝𝑒𝑜𝑝𝑙𝑒 𝑖𝑛𝑡𝑜 𝑡𝑤𝑜 𝑔𝑟𝑜𝑢𝑝𝑠 𝑎𝑟𝑒 𝑖𝑑𝑒𝑛𝑡𝑖𝑐𝑎𝑙.
𝑆𝑜𝑙: 𝑛1 = 60; 𝑛2 = 140
𝑥1 = 40; 𝑥2 = 80
𝑥1 40 𝑥2 80
𝑝1 = = = 0.6667 ; 𝑝2 = = = 0.5714
𝑛1 60 𝑛2 140
𝑛1 𝑝1 + 𝑛2 𝑝2 60(0.6667) + 140(0.5714)
𝑝= = = 0.6
𝑛1 + 𝑛2 60 + 140
𝑞 = 1 − 𝑝 = 1 − 0.6 = 0.4
𝑇𝑒𝑠𝑡𝑖𝑛𝑔 𝐻𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝑆𝑡𝑒𝑝1: 𝑁𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻0 : 𝑝1 = 𝑝2
𝑆𝑡𝑒𝑝2: 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠:
𝐻1 : 𝑝1 ≠ 𝑝2 [𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡]
𝑆𝑡𝑒𝑝3: 𝐿𝑒𝑣𝑒𝑙 𝑜𝑓 𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒:
𝛼 = 5%
𝑆𝑡𝑒𝑝4: 𝑇𝑒𝑠𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐:
𝑝1 − 𝑝2
𝑧=
1 1
√(𝑝𝑞) [
𝑛1 𝑛2 ]
+

0.6667 − 0.5714
= = 1.2607
√(0.6)(0.4) [ 1 + 1 ]
60 140
𝑆𝑡𝑒𝑝5: 𝐶𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛:
𝑎𝑡 5% 𝐿. 𝑂. 𝑆 𝑓𝑜𝑟 𝑎 𝑡𝑤𝑜 𝑡𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡,
𝑧𝛼 = 1.96
|𝑧| = 1.2607
|𝑧| < 𝑧𝛼 ⇒ 𝐻0 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑.