User-centred Research Methods | A2.1

User-Centred Design (UCD) is an approach that places the needs, wants and limitations of end-users at the centre of the design process from start to finish. Rather than developing products in isolation and then hoping users will adapt to them, UCD requires designers to engage with real users continuously — before a concept exists, during development and after launch.

The foundation of UCD is empathy: the capacity to understand a user's experience as if you were that user. Empathy is not guesswork or assumption. It is built through direct contact with users — watching them work, listening to them describe their frustrations, and placing yourself in their environment. A designer who has spent an afternoon observing patients in a hospital waiting room understands something about that space that no amount of second-hand description can replicate.

Three types of understanding drive UCD:

• Needs — what the user must be able to do. Functional, often non-negotiable.
• Wants — what the user would prefer: features, aesthetics, convenience. These can be traded off.
• Limitations — the physical, cognitive, cultural and technological constraints that define what is possible for this user in this context.

When designers understand all three, they avoid two common and costly mistakes: designing products with too few features (failing to meet needs) or too many features (adding complexity that increases cost and reduces usability without meeting any real need). Research by Jacob Schmookler (1962) found that user need — not technological opportunity — is the most powerful driver of successful innovation.

以用户为中心的设计（UCD）是一种将最终用户的需求、期望和限制置于设计过程核心的方法，贯穿从开始到结束的整个设计周期。UCD要求设计师与真实用户持续互动——在概念形成之前、开发过程中以及产品发布之后——而非孤立地开发产品后期望用户去适应它。

UCD的基础是共情：以用户的视角去理解其体验的能力。共情并非猜测或假设，而是通过与用户的直接接触来建立的——观察他们工作，倾听他们描述遇到的困难，将自己置身于他们的环境之中。

UCD由三类理解驱动：

• 需求——用户必须能够做什么。功能性的，通常不可妥协。
• 期望——用户更希望拥有什么：功能、美感、便利性。这些可以进行取舍。
• 限制——界定"对这位用户在这种情境下什么是可能的"的身体、认知、文化和技术约束。

当设计师理解这三者时，便能避免两种常见且代价高昂的错误：功能太少（无法满足需求）或功能太多（增加复杂性，提高成本，却没有满足任何真实需求）。研究者Jacob Schmookler（1962年）发现，用户需求——而非技术机遇——是成功创新最有力的驱动力。

The UCD process is typically structured around five stages, each of which involves users directly rather than treating them as passive recipients at the end:

1. Research — Understand the user, the task and the environment through field research, interviews, observation and other methods. Produce personas, scenarios and specifications that represent real user needs.
2. Concept — Generate and evaluate ideas against user needs. Present concepts to users or user representatives and gather feedback before investing in detailed design.
3. Design — Develop prototypes at increasing levels of fidelity. Test with representative users; identify usability issues early, when changes are still cheap.
4. Implementation — Build or manufacture the product. Conduct formal user testing — often under the guidance of psychologists — with a broader range of end users than was possible in earlier stages.
5. Launch — Release the product and continue gathering data. Post-launch research informs the next version.

Advantages of UCD:

• Reduces costly late-stage redesigns by catching usability problems early.
• Produces products that feel intuitive — reducing training time and user error.
• Supports accessibility by identifying the needs of diverse users before production.
• Strengthens market competitiveness by aligning features with genuine user demand.

Disadvantages of UCD:

• Time-consuming and expensive to run rigorous user research at every stage.
• Users cannot always articulate needs they haven't experienced yet — research must be interpreted carefully.
• Over-reliance on user feedback can produce incremental improvement rather than breakthrough innovation.
• Coordinating diverse disciplines and continuous user involvement adds project management complexity.

UCD过程通常围绕五个阶段展开，每个阶段都直接让用户参与其中，而非将其视为最终的被动接受者：

1. 研究——通过实地调研、访谈、观察等方法了解用户、任务和环境。建立能代表真实用户需求的角色模型、场景和设计规格。
2. 概念——根据用户需求生成并评估设计方案。在投入详细设计之前，向用户或用户代表展示概念并收集反馈。
3. 设计——开发不同精度级别的原型。与代表性用户进行测试；在修改成本仍然较低的早期阶段识别可用性问题。
4. 实施——构建或制造产品。通常在心理学家的指导下，与比早期阶段更广泛的最终用户群体进行正式用户测试。
5. 发布——推出产品并持续收集数据。发布后的研究将为下一版本提供参考。

UCD的优点：

• 通过早期发现可用性问题，减少代价高昂的后期重新设计。
• 产品使用起来更加直观，减少培训时间和用户错误。
• 在生产前识别多样化用户的需求，支持无障碍设计。
• 通过使功能与真实用户需求保持一致，增强市场竞争力。

UCD的缺点：

• 在每个阶段进行严格用户研究耗时且成本高昂。
• 用户并不总能清晰表达他们尚未体验过的需求——研究结果需要谨慎解读。
• 过度依赖用户反馈可能导致渐进式改进，而非突破性创新。
• 协调多个学科并持续让用户参与，增加了项目管理的复杂性。

No single discipline can meet all the needs of a diverse user population. A psychologist understands how people perceive and make decisions; an engineer understands what is structurally possible; a sociologist understands how culture shapes behaviour; an anthropologist understands how different communities live and work. A UCD team brings these disciplines together so that design decisions are informed by the full complexity of human experience.

Typical contributors to a UCD team include:

• Psychologists and sociologists — understand how users think, feel and behave in social contexts.
• Anthropologists — conduct ethnographic field research; understand cultural variation in behaviour and expectations.
• Engineers and materials scientists — translate user requirements into manufacturable, functional solutions.
• Data analysts — identify patterns across large user datasets that qualitative research would miss.
• Marketing experts — understand how products are perceived and positioned relative to user expectations.
• Environmental scientists — consider the context in which the product will be used, including climate, noise and physical space.

Case study — Mercury spacesuit: When NASA developed the Mercury spacesuit, the engineering team was joined by experts from biology, medicine and materials science. The most effective concept was proposed not by an engineer but by a biologist. The final suit incorporated an aluminium-coated nylon outer shell for thermal regulation (developed by biologists and engineers together), a closed-loop breathing system (developed with medical experts) and a "biomed flap" connecting sensors to spacecraft telemetry. A single-discipline engineering team would have produced a very different — and likely less safe — result.

没有单一学科能够满足多样化用户群体的所有需求。心理学家了解人们如何感知和做决策；工程师了解结构上什么是可能的；社会学家了解文化如何塑造行为；人类学家了解不同社区的生活和工作方式。UCD团队将这些学科汇聚在一起，使设计决策能够基于人类体验的完整复杂性。

UCD团队的典型成员包括：

• 心理学家和社会学家——了解用户在社会情境中如何思考、感受和行动。
• 人类学家——开展民族志实地研究；了解行为和期望中的文化差异。
• 工程师和材料科学家——将用户需求转化为可制造、可运作的解决方案。
• 数据分析师——在大型用户数据集中识别定性研究难以发现的规律。
• 营销专家——了解产品如何被感知以及相对于用户期望的市场定位。
• 环境科学家——考虑产品将被使用的情境，包括气候、噪音和物理空间。

案例研究——水星太空服：NASA在开发水星太空服时，工程团队与生物学、医学和材料科学专家携手合作。最有效的方案并非由工程师提出，而是来自一位生物学家。最终的太空服采用了铝涂层尼龙外壳用于热调节（由生物学家与工程师共同开发）、封闭式呼吸系统（与医学专家共同开发），以及连接传感器与飞船遥测系统的"生物医学接口"。单一学科的工程团队将会产出截然不同——且可能安全性更低——的结果。

Each research method collects different kinds of information. Choosing the right method depends on what you need to know — not on which method is easiest to run. Research methods generate two broad categories of data:

• Qualitative data — non-numerical; describes qualities, feelings, opinions and behaviours. Answers why and how. Harder to compare across large samples but reveals the reasoning behind behaviour.
• Quantitative data — numerical; can be counted, measured and statistically analysed. Answers how many and how often. Easier to compare at scale but does not explain the reasons behind patterns.

Field research — Observing users in their authentic, real-world environment (a kitchen, a factory, a school). Uncovers behaviour influenced by cultural norms, physical habits and contextual factors that users would not think to mention in a lab or interview. Risk: observers may inadvertently change the behaviour they are watching.

Task analysis — Breaking down a complex task into a hierarchy of smaller steps to identify where errors, delays or frustration occur. Reveals the specific sub-tasks that cause the most friction. Feeds directly into redesign decisions.

User observation — Watching users perform specific tasks, either in the field or in a controlled setting. Captures hesitation, workarounds and errors that users would not self-report. Unlike field research, observation is typically structured around a defined task.

Interviews — Direct conversation with users; predominantly generates qualitative data. Structured interviews use identical questions for all participants (comparable results). Unstructured interviews follow the conversation, uncovering unexpected insights. Most practical research uses a semi-structured format: a consistent set of questions with room to follow up.

Surveys and questionnaires — Distributed to large groups; efficient for gathering quantitative data from many users quickly. A Likert scale (e.g., 1 = Strongly Disagree to 5 = Strongly Agree) converts attitudes into numbers that can be averaged and compared. Limitation: surveys cannot ask follow-up questions to understand why a user gave a particular answer.

Focus groups — A structured, moderated discussion with a carefully selected subset of the target audience (typically 6–10 people). Generates in-depth qualitative data and captures how perspectives develop through interaction between participants. Risk: a dominant participant can skew the group's responses; results may not represent the wider population.

每种研究方法收集不同类型的信息。选择正确的方法取决于你需要了解什么——而非哪种方法最容易执行。研究方法产生两大类数据：

• 定性数据——非数值型；描述特征、感受、观点和行为。回答为什么和如何。难以在大样本间进行比较，但能揭示行为背后的原因。
• 定量数据——数值型；可以统计、测量和进行统计分析。回答有多少和多频繁。易于大规模比较，但无法解释规律背后的原因。

实地研究——在用户真实的生活环境（厨房、工厂、学校）中观察他们。揭示受文化规范、身体习惯和情境因素影响的行为——这些是用户在实验室或访谈中不会主动提及的。风险：观察者可能无意间改变他们正在观察的行为。

任务分析——将复杂任务分解为层级化的小步骤，以识别错误、延迟或挫败感出现的位置。揭示引发最大摩擦的具体子任务，直接为重新设计决策提供依据。

用户观察——观察用户在现场或受控环境中执行特定任务。捕捉用户不会自我报告的犹豫、变通做法和错误。与实地研究不同，用户观察通常围绕既定任务进行结构化安排。

访谈——与用户直接对话；主要产生定性数据。结构化访谈对所有参与者使用相同的问题（结果具有可比性）；非结构化访谈跟随对话发展，能发现意外见解。大多数实际研究采用半结构化格式：有一套固定问题，但留有追问空间。

调查问卷——发放给大量人群；高效地从众多用户处快速收集定量数据。李克特量表（如1=非常不同意至5=非常同意）将态度转化为可以求平均值和比较的数字。局限性：调查无法通过追问来了解用户给出特定答案的原因。

焦点小组——与目标受众中精心挑选的代表性群体（通常6–10人）进行结构化的主持讨论。产生深度定性数据，并捕捉观点如何在参与者互动中形成和发展。风险：强势参与者可能主导小组讨论；结果可能无法代表更广泛的人群。

After research data is collected, designers use three tools to translate raw findings into a shared language that guides the whole team:

Personas — Fictional but research-based user profiles that represent a target group. A persona is not a real individual; it is a composite drawn from patterns identified across many users. A well-constructed persona includes a name, age, occupation, goals, frustrations, and relevant behavioural habits. Personas give every member of a multidisciplinary team — engineers, marketers, writers — a concrete human to design for, reducing the risk of abstract or self-referential design decisions.

Advantages: Build empathy; keep the team focused on specific user needs; make design reviews concrete ("would Maya be able to do this?"). Disadvantages: Can become quickly outdated as trends shift; risk reinforcing stereotypes if not grounded in rigorous research; may oversimplify a diverse user population into a single profile.

Scenarios — Contextual stories that describe how a persona uses a product to accomplish a goal in a specific situation. A scenario grounds design decisions in real use: "Maya is running late and tries to complete the checkout on her phone while standing in a queue." Scenarios reveal edge cases and priority conflicts that abstract specifications miss.

Population stereotypes — Culturally shared assumptions about how products should behave, built up through repeated exposure over time. These are not personality stereotypes; they are learned behavioural expectations about controls, symbols and conventions. Examples:

• Turning a knob clockwise increases volume, temperature or speed.
• In the UK, pushing a light switch down turns it on; in the USA, pushing it up turns it on. Both are population stereotypes — neither is objectively correct.
• Red means stop or danger; green means safe or go.

Population stereotypes make products feel immediately intuitive because they match expectations built up over a lifetime. Violating them — even accidentally — causes confusion and error. Under stress, users revert to ingrained stereotypes even if they have learned a different convention for a specific product. This has serious implications for safety-critical design (medical devices, emergency equipment).

在收集研究数据之后，设计师使用三种工具将原始发现转化为能够指导整个团队的共同语言：

角色模型（Persona）——基于研究数据创建的虚构用户档案，代表一个目标群体。角色模型不是真实的个人，而是从大量用户中识别出的规律综合而成的合成形象。一个构建良好的角色模型包括姓名、年龄、职业、目标、挫败点和相关行为习惯。角色模型让多学科团队的每一位成员——工程师、营销人员、文案——都有一个具体的"人"可以为之设计，降低了抽象或自我参照式设计决策的风险。

优点：培养共情；让团队专注于具体的用户需求；使设计评审更加具体（"Maya能做到这个吗？"）。缺点：随着趋势变化可能迅速过时；若不以严格研究为基础，可能强化刻板印象；可能将多元化的用户群体过度简化为单一档案。

场景（Scenario）——描述角色模型如何在特定情境下使用产品完成目标的背景故事。场景将设计决策落地于真实使用情境中。场景揭示抽象规格所忽略的边缘案例和优先级冲突。

群体刻板印象（Population Stereotypes）——通过长期反复接触而建立起来的、关于产品应如何运作的文化共识。这些不是个性刻板印象，而是关于控件、符号和惯例的习得性行为预期。例如：

• 顺时针旋转旋钮会增大音量、温度或速度。
• 在英国，将开关向下拨表示开启；在美国，向上拨表示开启。两者都是群体刻板印象——没有哪个在客观上是"正确的"。
• 红色表示停止或危险；绿色表示安全或通行。

群体刻板印象让产品使用起来立刻感觉直观，因为它们符合用户一生中积累的预期。即使是无意中违反这些预期，也会造成混乱和错误。在压力下，用户会回归根深蒂固的刻板印象，即使他们已经为某一特定产品学习了不同的惯例。这对安全关键设计（医疗设备、应急设备）具有重要意义。