📚 GCSE Computer Science: Experimentation Guide | GCSE 计算机：实验操作指南
Experimentation in computer science is not just about mixing chemicals in a lab—it’s about systematically testing hypotheses, measuring performance, and drawing valid conclusions about algorithms, systems, and networks. This guide walks you through the essential steps of conducting a successful computer science experiment, from planning to analysis.
计算机科学中的实验不只是像在实验室中混合化学试剂一样——它是系统地检验假设、测量性能，并针对算法、系统和网络得出有效结论的过程。本指南将带你走过成功开展计算机科学实验的必要步骤，从规划到分析。
1. Understanding the Role of Experiments | 理解实验的作用
In GCSE Computer Science, experiments help you evaluate the efficiency of algorithms, compare hardware configurations, or test the reliability of network protocols. They provide evidence for your coursework and deepen your understanding of theoretical concepts.
在GCSE计算机科学中，实验帮助你评估算法的效率、比较硬件配置或测试网络协议的可靠性。它们为你的课程作业提供证据，并加深你对理论概念的理解。
2. Types of Computer Science Experiments | 计算机实验的类型
Common experiment types include benchmarking (measuring execution time), simulation (modelling a system), and comparative studies (testing two or more approaches). You might time a sorting algorithm, simulate a logic circuit, or compare search algorithms.
常见的实验类型包括基准测试（测量执行时间）、模拟（对系统建模）和比较研究（测试两种或多种方法）。你可能会对一个排序算法计时、模拟一个逻辑电路，或者比较搜索算法。
3. Formulating a Research Question and Hypothesis | 形成研究问题与假设
Start with a clear question, e.g., ‘Does a binary search outperform a linear search on a sorted array of 1000 items?’ Then phrase a hypothesis: ‘Binary search will be faster because it reduces the search space by half each step.’
从一个清晰的问题开始，例如：“二分搜索在1000个已排序项目的数组上是否优于线性搜索？”然后表述假设：“二分搜索会更快，因为它每一步都将搜索空间减半。”
4. Identifying Variables | 识别变量
Every experiment has independent variables (the factor you change), dependent variables (what you measure), and control variables (what you keep constant). For a sorting experiment, the independent variable could be the algorithm, the dependent variable the time taken, and controls include data size, data order, and hardware.
每个实验都有自变量（你改变的因素）、因变量（你测量的内容）和控制变量（你保持不变的因素）。对于排序实验，自变量可以是算法，因变量是所花的时间，控制变量包括数据量、数据顺序和硬件。
- Independent variable: sorting algorithm (bubble sort, merge sort)
- Dependent variable: execution time measured in milliseconds
- Control variables: number of items, initial order (random, reversed), processor speed
- 自变量：排序算法（冒泡排序、归并排序）
- 因变量：以毫秒计的执行时间
- 控制变量：项目数量、初始顺序（随机、逆序）、处理器速度
5. Choosing Tools and Setting Up | 选择工具与设置
Use a programming language like Python, an IDE that supports timing (e.g., the time module with perf_counter()), and a consistent testing environment. Disable background processes and use the same machine throughout to reduce noise.
使用如Python这样的编程语言，一个支持计时的IDE（例如time模块的perf_counter()），以及一致的测试环境。全程关闭后台进程并使用同一台机器以减少噪声。
6. Collecting Reliable Data | 收集可靠数据
Run each test multiple times (at least 5–10 repetitions) and take the average to minimise random errors. Record all raw data in a table, noting any anomalies such as sudden spikes. A well-structured data table might include columns for trial number, algorithm, input size, time, and notes.
每种测试运行多次（至少5–10次重复）并取平均值以最小化随机误差。将所有原始数据记录在表格中，注意任何异常值，比如突然的尖峰。一个结构良好的数据表可能包括试验编号、算法、输入规模、时间和备注等列。
7. Ensuring a Fair Test | 确保公平测试
Change only the independent variable. Use the same data set and hardware for all trials. If comparing algorithms, implement them in the same language with similar coding practices. Pre-load data and allow the environment to warm up if you are measuring execution time.
只改变自变量。所有试验使用相同的数据集和硬件。如果比较算法，用相同的语言和类似的编码习惯来实现它们。如果你正在测量执行时间，要预加载数据并让环境预热。
8. Analysing Results | 分析结果
Plot graphs such as line charts of time vs. input size. Look for patterns: does the time grow linearly (O(n)), quadratically (O(n²)), or logarithmically (O(log n))? Use your knowledge of Big O notation to explain trends. Calculate averages and, if appropriate, the standard deviation to assess consistency.
绘制图表，比如时间-输入规模的折线图。寻找模式：时间增长是线性的（O(n)）、平方的（O(n²)）还是对数的（O(log n)）？利用你对大O符号的知识来解释趋势。计算平均值，如果合适的话，计算标准差来评估一致性。
9. Drawing Conclusions | 得出结论
Conclude whether your hypothesis was supported by the evidence. For example, ‘Binary search was consistently faster than linear search, confirming the theoretical O(log n) advantage over O(n).’ Discuss limitations of your experiment, such as small sample sizes or unaccounted hardware variations.
得出结论说明证据是否支持你的假设。例如，“二分搜索始终比线性搜索快，证实了O(log n)相对于O(n)的理论优势。”讨论你实验的局限性，例如样本量小或未考虑的硬件差异。
10. Writing a Lab Report | 撰写实验报告
A standard report includes: title, introduction (background and hypothesis), method (variables and procedure), results (tables and graphs), discussion (interpretation and limitations), and conclusion. Use clear language and reference any sources or tools.
一份标准的报告包括：标题、引言（背景和假设）、方法（变量和步骤）、结果（表格和图表）、讨论（解释和局限性）以及结论。使用清晰的语言并引用任何来源或工具。
11. Common Pitfalls and How to Avoid Them | 常见误区与避免方法
Overlooking warm-up effects, using too few trials (leading to unreliable averages), changing multiple variables at once, and neglecting to control hardware conditions are frequent mistakes. Always pilot your experiment first and check your code for hidden inefficiencies.
忽视预热效应、试验次数太少（导致不可靠的平均值）、一次改变多个变量以及忽略硬件条件控制是常见错误。始终先进行预实验，并检查你的代码是否存在隐藏的低效问题。
12. Worked Example: Comparing Sorting Algorithms | 实例：比较排序算法
Let’s design an experiment to compare bubble sort and merge sort. Hypothesis: merge sort will be faster for large data sets because of its O(n log n) time complexity versus bubble sort’s O(n²). Independent variable: algorithm. Dependent: execution time. Controls: array sizes (1000, 5000, 10000 integers), same initial random order, same machine. Run each algorithm 10 times per size, take the mean, and plot mean time against array size. You should see the merge sort curve rising more gently, confirming the hypothesis.
我们来设计一个实验比较冒泡排序和归并排序。假设：对于大数据集归并排序会更快，因为它的时间复杂度为O(n log n)，而冒泡排序为O(n²)。自变量：算法。因变量：执行时间。控制变量：数组大小（1000, 5000, 10000个整数），相同的初始随机顺序，同一台机器。每种规模每个算法运行10次，取平均值，并将平均时间相对数组大小绘图。你应该会看到归并排序的曲线上升得更平缓，从而证实假设。
Time complexity: Bubble sort O(n²) | Merge sort O(n log n)
时间复杂度：冒泡排序 O(n²) | 归并排序 O(n log n)
Published by TutorHao | Computer Science 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