32 studies, standards and frameworks behind this guide. Each is rated by how much you should trust it, with a plain-English summary and a link to the original. Including the famous ones that are shakier than people think.
Replicated many times or official data. Build on it.
Solid peer-reviewed studies, but limited samples or contexts. Use with judgement.
Expert frameworks and standards. Useful rules of thumb, not lab proof.
Popular but over-stated, anecdotal or untraceable. Quote carefully.
Ratings of a web page's visual appeal after a 50 ms flash closely matched ratings after 500 ms. The gut verdict on looks is near-instant. It measured appeal only, not trust or buying.
Read the sourceSites with low visual complexity and high "prototypicality" (looking like what people expect) were rated most appealing, even at 17–50 ms exposures.
Read the sourceIn comments from 2,684 people comparing live sites, 46.1% of comments mentioned "design look" (layout, type, colour), more than any other factor. Comment analysis, not a controlled experiment.
Read the sourceWe like things more the more we see them. It's one of psychology's most replicated effects, and it's why your diet of images shapes your taste.
Read the sourceSingle features like colour or orientation "pop out" in parallel; combinations need slow serial search. Make the key thing different in one obvious way.
Read the sourceProximity, similarity, closure, continuity and figure/ground describe how we perceive wholes. Over a century of perception research builds on them.
Read the sourceDecision time rises with the number of choices (logarithmically). Strong in the lab; in real interfaces, familiarity and grouping matter too.
Read the sourceTime to hit a target depends on its distance and size. Big, near buttons are faster. Among the most robust laws in HCI.
Read the sourceMiller suggested 7±2 items; later work by Cowan points to about 4 chunks. Chunk information and don't make people remember things between screens.
Read the sourceAn item that differs from its neighbours is remembered better. That's the logic behind one highlighted plan or one red button.
Read the sourcePeople rated nicer-looking ATM layouts as easier to use. Replicated across cultures (Tractinsky, 1997), though looks don't fix real usability problems.
Read the sourceEye-tracking of 232 users showed an F-shaped scan on text-heavy pages. NN/g later stressed that it is a symptom of poor formatting, not a law.
Read the sourceVisceral (looks), behavioural (use) and reflective (meaning). A thinking framework from a leading design researcher, not an experiment.
Read the sourceAcross four studies, hue, saturation and lightness shifted brand personality: red raised "excitement", blue raised "competence", and colour affected purchase intent.
Read the sourceLogo colours were judged by how appropriate they were for the product type: functional products suit functional colours, sensory products suit expressive ones.
Read the sourceA major review: colour does influence feelings and behaviour, but the effects depend heavily on context and the field is younger than pop articles imply.
Read the sourceThe famous stat comes from a literature review, not a new experiment, and the exact range isn't traceable to one clear study. Use the honest version: colour strongly shapes first impressions.
Read the sourceAbout 1 in 12 men and 1 in 200 women have some colour-vision deficiency. Never rely on colour alone to carry meaning.
Read the sourceThe international accessibility standard asks for at least 4.5:1 contrast for body text and 3:1 for large text (AA), and 7:1 for AAA.
Read the sourcePeople chose brands more often when the font "fit" the product (sturdy font for a tough product), even with identical names.
Read the sourceReaders consistently attributed personalities to typefaces; when font and text tone clash, the message feels off.
Read the sourceExercise instructions in a hard-to-read font were estimated at 15.1 minutes instead of 8.2, and people were less willing to try. Small study (20 students), but consistent with wider "processing fluency" research.
Read the sourceThe classic typographic guideline for comfortable reading, with 66 often called ideal. Screen studies (e.g. Dyson, 2004) show speed and preference can differ, so treat it as a starting range.
Read the sourceContrast, Repetition, Alignment, Proximity: the best-known beginner framework for layout.
Read the sourceThe Swiss reference for designing with modular grids, still used by editorial and UI designers.
Read the sourceThe golden ratio (1.618) is a useful proportion, but most "this logo is built on it" diagrams are drawn on afterwards. Use it as a tool, not as proof.
Read the sourceA mathematical model suggests about 5 users uncover roughly 85% of usability problems, as long as you test in several small rounds. Debated for complex products.
Read the sourceFeatures fall into must-haves, performance needs and delighters, and delighters decay into basics over time. A widely used product-research framework.
Read the sourceA memorable case study, not a peer-reviewed experiment. Use it as a thinking tool: what "job" is your design hired for?
Read the sourceUS median pay: graphic designers $61,300, web & digital interface designers $98,090 (May 2024). Projected growth 2024–34: +2% and +7%.
Read the sourceIndustry analysis of BLS wage-survey data: counted graphic-design employment fell from about 214k (May 2024) to about 198k (May 2025) while median wages rose.
Read the sourceIra Glass's widely shared advice: beginners' taste outruns their skill, and only volume of work closes the gap. Expert opinion, and very true in practice.
Read the sourceA journal, a book or a blog? Can you find the original, not just a blog quoting a blog?
Experiment, survey, review, case study or opinion? Each answers a different question.
20 students in one university is a hint. 2,000 people across countries is a pattern.
Findings that replicate (like mere exposure or Fitts's Law) are worth building on.
A colour result from US shoppers may not hold in Mumbai. Context changes meaning.