Collecting data · 数据的收集
| English | 中文 | Pinyin · 拼音 |
|---|---|---|
| population/ˌpɒpjʊˈleɪʃn/ | 总体 | zǒng tǐ |
| sample/ˈsæmpl/ | 样本 | yàng běn |
| representative/ˌreprɪˈzentətɪv/ | 有代表性的 | yǒu dài biǎo xìng de |
| qualitative data/ˈkwɒlɪteɪtɪv ˈdeɪtə/ | 定性数据 | dìng xìng shù jù |
| quantitative data/ˈkwɒntɪteɪtɪv ˈdeɪtə/ | 定量数据 | dìng liàng shù jù |
| discrete/dɪˈskriːt/ | 离散 | lí sàn |
| continuous/kənˈtɪnjuːəs/ | 连续 | lián xù |
| histogram/ˈhɪstəɡræm/ | 直方图 | zhí fāng tú |
Define whom a finding describes
- A population 总体 is the target group; a sample 样本 is the observed subset. A class survey can describe that class without establishing a school-wide percentage.
- A suitably representative 有代表性的 sample requires attention to selection, participation and the question asked. A large voluntary response does not automatically resolve nonresponse.
Distinguish labels from measured amounts
- Qualitative data 定性数据 are categories. Quantitative data 定量数据 measure amounts; discrete 离散 data use distinct countable values, while continuous 连续 data are modelled over intervals.
- A numerical route label is still a category. Recorded continuous measurements can be rounded, so printed decimal precision alone does not decide the underlying type.
Match each variable to its data type. · 把每个变量与其数据类型配对。
Counted values are discrete; measured values are continuous. The type decides which chart is legitimate. · 数出来的是离散,量出来的是连续。类型决定了哪种图表合规。
Match the display to the data
- A histogram 直方图 uses numeric class intervals. For unequal widths, frequency density is frequency divided by width so bar area represents frequency.
- Separate category bars and label units, scale and count scope. A pie chart requires mutually exclusive parts of a defined whole; missing responses must not silently become a known category.
Which chart suits continuous data such as height? · 身高这类连续数据适合用哪种图?
A histogram displays frequencies over numeric intervals. With unequal interval widths, use frequency density so area represents frequency; an empty interval can have zero height. · 直方图显示数值区间的频数。当区间宽度不等时，应使用频率密度，使面积代表频数；空区间的高度可为零。
Two sampling plans have different risks. Asking a canteen queue excludes nonusers. Selecting every tenth name can cover the full roll, but needs a random start and a check for patterns in the ordering. Neither method guarantees honest responses or removes all nonresponse; record the actual procedure and limits.
To study canteen preferences across a school, which listed selection plan best covers the full student population? · 若要研究全校学生的食堂偏好，下列抽样计划中哪一个最能覆盖全体在校学生？
Selection from the full roll can include nonusers. A systematic every-tenth-name plan needs a random start and a check for patterns; nonresponse remains a limitation, so it does not guarantee perfect representativeness. · 从完整花名册中抽样可能包含非使用者。每十人抽取一人的系统抽样法需要随机起点并检查是否存在模式；无响应仍是局限性，因此不能保证完美的代表性。
State the denominator. If 60 of 200 students reply and 45 prefer later meetings, response rate is 30% and respondent preference is 75%. Preferences of the remaining 140 are unknown. Sheet 2.5 separates preference evidence from attendance evidence.
A larger sample removes bias caused by how the sample was chosen. · 样本量更大,就能消除抽样方式带来的偏差。
Increasing sample size alone does not remove systematic selection or nonresponse bias. Examine and improve the collection method and disclose remaining limits. · 单纯增加样本量并不能消除系统性选择偏差或无响应偏差。应审查并改进收集方法，并披露剩余局限。
A truncated category-bar axis changes the visible comparison of heights. Use a zero baseline for straightforward magnitude comparisons, and make any exceptional scale choice explicit. A larger sample does not by itself fix a systematically selected sample.
A bar chart's vertical axis starts at 40 rather than 0. What must you do? · 柱状图的纵轴从 40 而不是 0 开始,你必须怎么做?
A truncated category-bar scale can exaggerate relative heights. A zero baseline supports magnitude comparisons; any exception needs a clear, justified scale rather than an unstated change. · 截断的分类条形图可能会夸大相对高度。零基线支持数量级比较；如有例外，需要有清晰、合理的刻度，而非未声明的变化。