This free curriculum covers every domain assessed in the Oxford Certified UI/UX Research Architect examination: research strategy, recruiting, interviews and field research, surveys, usability testing and metrics, information architecture, analytics and experiments, synthesis and communication, models of users, accessibility, ethics and research operations. Read each chapter, note the key points, and then sit the 70-question examination.
Generative research discovers needs and opportunities before solutions exist; evaluative research assesses how well designs work. Methods range along two dimensions: attitudinal (what people say) versus behavioural (what people do), and qualitative (why and how) versus quantitative (how many and how much). Behavioural methods — observation, usability testing, analytics — reveal actual behaviour; combining qualitative and quantitative research explains both what happens and why.
A research plan starts with the decisions the research will inform and the research questions. Research roadmaps align studies with upcoming product decisions. Researchers must guard against confirmation bias — design studies that could disprove assumptions rather than "prove" a favoured idea.
Generative vs evaluative
Attitudinal vs behavioural; qualitative vs quantitative
Start with decisions and research questions
Plan research ahead of decisions
Design studies that can disconfirm assumptions
Idea clave: The right method depends on the question, not on habit.
Capítulo 2
Recruiting and sampling
Screener questionnaires select participants who match target criteria. For qualitative usability testing, around five participants per user group typically uncover most problems, and several small iterative rounds — test, fix, retest — are more effective than one large round. Quantitative studies need larger samples to estimate metrics such as completion rates with acceptable confidence intervals.
Self-selected samples — such as newsletter subscribers — over-represent engaged users and create selection bias; representative sampling matters whenever results are generalised.
Screen for the right participants
About five users per round for qualitative usability tests
Iterate in small rounds
Larger samples for quantitative estimates
Watch for selection bias
Idea clave: Who you ask determines what you learn.
Capítulo 3
Interviews and field research
Good interviews ask about specific past behaviour ("Tell me about the last time…") rather than hypotheticals, use neutral, non-leading questions and follow up with probes such as "tell me more". Avoid explaining the product before asking for opinions. Asking users which features they want is not a reliable way to decide what to build — research should uncover needs and problems first.
Contextual inquiry observes and interviews people in their real environment while they work. Diary studies capture experiences over days or weeks. Focus groups can surface attitudes but are unreliable for measuring how individuals actually use a product.
Ask about real past behaviour
Stay neutral and probe deeper
Observe in context
Use diary studies for longitudinal behaviour
Needs first, solutions second
Idea clave: Listen for what people do and need, not just what they say they want.
Capítulo 4
Surveys
Surveys measure attitudes and behaviours at scale. Avoid double-barrelled questions that ask about two things at once, leading wording and unbalanced scales. Likert scales measure degrees of agreement on ordered scales. Representative sampling, clear neutral questions and short surveys improve data quality.
Self-selected samples — for example, only newsletter subscribers — over-represent engaged users and can make results look far more positive than reality. Combine surveys with behavioural data and qualitative research to understand the reasons behind the numbers, and report response rates and sample composition alongside results.
One concept per question
Neutral wording and balanced scales
Representative samples
Keep surveys short and purposeful
Idea clave: A biased question produces precise-looking but misleading data.
Capítulo 5
Usability testing and UX metrics
Usability tests observe participants completing realistic tasks, written without giving away interface labels. In think-aloud tests participants verbalise their thoughts; moderators stay neutral and let participants try as they normally would. Moderated remote testing allows real-time observation and follow-up in the participant's own environment.
Common metrics include task success, time on task and error rate. The System Usability Scale is a 10-item questionnaire scored 0–100, with about 68 considered average. Benchmark studies measure usability consistently over time or against competitors. Google's HEART framework covers Happiness, Engagement, Adoption, Retention and Task success. Eye tracking shows where users look and what they notice.
Realistic, unbiased tasks
Think-aloud and neutral moderation
Task success, time, errors and SUS
Benchmark over time
HEART for product-level UX metrics
Idea clave: Watching real users attempt real tasks is the fastest way to find problems.
Capítulo 6
Information architecture research
Card sorting reveals how users group and label content: in open sorts participants create their own groups; in closed sorts they use predefined categories. Tree testing checks whether users can find items in a proposed navigation structure without visual design, making it ideal before visual design begins.
Card sorting results are analysed with similarity matrices and dendrograms that show which items participants group together. Tree testing measures success rates, directness (whether users reach the answer without backtracking) and time, showing exactly where a structure fails before any visual design is built.
Card sorting for mental models
Open vs closed sorts
Tree testing for findability
Validate structure before visual design
Idea clave: Good navigation mirrors how users think about content.
Capítulo 7
Analytics, experiments and expert review
Behavioural analytics — funnels, heatmaps, session analysis — show where users drop off at scale, but not why; qualitative research explains the reasons. A/B testing measures which variant performs better on a defined metric with sufficient live traffic.
Heuristic evaluation reviews an interface against principles such as Nielsen's 10 heuristics, including visibility of system status, error prevention and recognition rather than recall (make options visible so users need not remember them).
Analytics show what; research shows why
A/B tests need traffic and a clear metric
Heuristic review finds issues quickly
Combine methods for confidence
Idea clave: Numbers locate the problem; observation explains it.
Capítulo 8
Synthesis and communicating insights
Affinity mapping clusters observations into themes. An insight is a meaningful interpretation of evidence that explains behaviour and suggests design implications — not a raw quote. Reduce bias by analysing with more than one person, seeking disconfirming evidence and separating observations from interpretations; triangulation across methods increases confidence.
Findings influence decisions when they link to business goals and are prioritised, evidenced and communicated with clear recommendations. Short highlight clips (with consent) build empathy. When findings conflict with a senior opinion, present evidence, confidence levels and limitations. Research repositories with tagged, atomic insights prevent repeated studies. Business value is shown through decisions changed and measured impact — fewer support calls, higher conversion, avoided development cost.
Cluster, interpret, prioritise
Guard against analysis bias
Link insights to goals and recommendations
Use repositories and readouts
Show research impact in business terms
Idea clave: Research only matters when it changes decisions.
Capítulo 9
Modelling users and journeys
Personas are research-based profiles of users' goals, behaviours and pain points, updated over time — not demographic fiction. Jobs-to-be-done focuses on the progress a person wants to make in a particular situation and why they "hire" a product. Journey maps show stages, actions, touchpoints, thoughts, emotions and pain points; service blueprints add the front-stage and back-stage processes and support systems behind each interaction.
These models are only valuable when they are based on evidence, shared widely and kept up to date as research continues. Teams use them to prioritise problems, align on who they are designing for and spot gaps between what customers experience and what the organisation delivers.
Research-based personas
Jobs-to-be-done explains motivation
Journey maps across touchpoints
Service blueprints reveal back-stage causes
Idea clave: Models make research reusable across teams and decisions.
Capítulo 10
Accessibility and inclusive design
WCAG organises accessibility into four principles: perceivable, operable, understandable and robust. Level AA requires a contrast ratio of at least 4.5:1 for normal body text. Automated checkers miss many barriers, so research should include people using screen readers, switch devices and magnification. Errors must not rely on colour alone; text messages linked to fields and announced to assistive technologies are needed.
Inclusive design considers the full range of human diversity — ability, language, culture and context — from the start of research, design and development.
Idea clave: Accessible design is better design for everyone.
Capítulo 11
Ethics, consent and research operations
Informed consent explains the purpose, data collected, use of data and the right to withdraw. Personal data is minimised, stored securely, access-controlled and deleted or anonymised when no longer needed. Researchers avoid harm and take special care with vulnerable groups.
Research operations provide recruitment, tools, consent processes, repositories and governance so research scales. Democratising research enables non-researchers through training, guides and review, with researchers supporting higher-risk studies. Prototype fidelity matches the research question, and design systems provide reusable, accessible components that make research-informed patterns consistent.
Informed, revocable consent
Protect participant data
Avoid harm; protect vulnerable groups
ResearchOps enables scale
Democratise with guardrails
Idea clave: Ethical, well-run research earns the trust that good insight depends on.
¿Listo? Haz el examen gratis
Haz el examen y obtén tu insignia digital gratis. Después de aprobar podrás pedir el certificado oficial.