Everything you need to know about using PsychScales.
Marketing scales are validated multi-item survey instruments used to measure psychological and behavioural constructs in consumer and marketing research. Because constructs such as brand loyalty, purchase intention, consumer trust, or perceived quality cannot be observed directly, researchers use marketing scales to capture them indirectly through a set of related questions or statements. Respondents rate each item — typically on a Likert-type or semantic differential scale — and their responses are combined into a composite score that represents the underlying construct. Using a validated marketing scale rather than a single question substantially improves both the reliability and validity of the measurement, and makes findings comparable across different studies and populations.
A handbook of marketing scales is a curated reference compiling validated multi-item measures used across consumer and marketing research. Traditionally, marketing scales handbooks were printed academic volumes — expensive, quickly outdated, and accessible only through university libraries. PsychScales is the modern alternative: a free, searchable, continuously updated online handbook of marketing scales drawn from leading peer-reviewed journals including the Journal of Marketing, Journal of Consumer Research, Journal of Marketing Research, Journal of Retailing, and the Journal of the Academy of Marketing Science. It is the most comprehensive open-access marketing scales handbook available — and grows continuously as new validated instruments are added.
Marketing scales cover an enormous range of constructs relevant to consumer behaviour, brand management, advertising, retailing, and service marketing. Common constructs measured by marketing scales include brand loyalty, purchase intention, consumer trust, word-of-mouth likelihood, perceived quality, customer satisfaction, brand attitude, involvement, self-congruity, price sensitivity, impulse buying, hedonic and utilitarian value, social influence, and environmental concern. Marketing scales also exist for more specific constructs such as need for cognition, regulatory focus, temporal discounting, materialism, status consumption, and compulsive buying. PsychScales organises its marketing scales by category and subcategory, making it straightforward to find the right measure for any construct.
Rigorous marketing scale development follows a multi-stage process. Researchers begin by defining the construct conceptually and generating a pool of candidate items through literature review, qualitative interviews, or expert judgment. Items are then refined through pilot testing and purified using exploratory factor analysis to retain only those that load cleanly on the intended construct. The refined marketing scale is then subjected to confirmatory factor analysis (CFA) to assess model fit, and key psychometric properties are reported: Cronbach's alpha (internal consistency reliability, with values above 0.70 generally considered acceptable), Average Variance Extracted (AVE, a measure of convergent validity), and discriminant validity tests to confirm the scale measures something distinct from related constructs. Well-validated marketing scales are then replicated across multiple studies and samples before entering the literature as reference instruments.
Choosing the right marketing scale involves three considerations. First, conceptual fit — the scale's construct definition should match your theoretical use of the variable. A marketing scale measuring brand attitude is not interchangeable with one measuring brand attachment, even though both relate to consumer-brand relationships. Second, psychometric quality — prefer marketing scales with alpha above 0.80, published AVE above 0.50, and evidence of discriminant validity. Third, contextual appropriateness — check the samples and settings in which the marketing scale was originally validated and whether they are comparable to your study. PsychScales makes this process straightforward: each entry includes the full construct definition, reliability statistics, validity evidence, the original sample context, and the source citation, so you can make an informed judgment without hunting through journal archives.
A marketing scale should be cited by referencing the paper in which it was originally published or first validated. For example, if you use a brand loyalty scale from a 2014 Journal of Marketing paper, cite that paper in your method section when you introduce the marketing scale, and include the full reference in your bibliography. If you adapted items, note which items were changed and why. PsychScales provides a ready-to-use APA 7th edition citation for every marketing scale in the database — simply open the scale detail panel and copy the citation directly. For marketing scales with complex development histories, it is good practice to cite both the original source and the adapted version you used.
The distinction is largely one of application rather than method. Both marketing scales and psychological scales are psychometric instruments built and validated using the same techniques — factor analysis, reliability testing, and validity assessment. Psychological scales are developed for clinical, personality, or cognitive constructs, typically validated in general population or clinical samples. Marketing scales are developed specifically for consumer and marketplace constructs, validated in consumer samples, and normed to the purchase, brand, and service contexts that marketing researchers study. Many marketing scales draw on psychological theories — for example, marketing scales for regulatory focus, need for cognition, or materialism are rooted in personality and social psychology — making the boundary permeable. PsychScales covers both, with particular depth in consumer behaviour and marketing research instruments.
Type any keyword into the search bar on the Browse page — scale names, abbreviations (e.g. 'PHQ-9'), constructs (e.g. 'depression', 'brand loyalty'), author names, or describe what you need in plain language (e.g. 'measure anxiety without clinical diagnosis'). Results filter instantly as you type. Press Enter to also run an AI search across the full database.
Typing filters the curated database in real time. Pressing Enter (or clicking ✦ Search) triggers the AI layer — it semantically matches your query against all scales in the database and can identify relevant scales even when exact keywords don't match. If no scales are found in the database at all, the AI will generate a full profile for the scale you're looking for.
This shows the 20 most widely cited validated scales across consumer behaviour, marketing, and psychology research, ranked by citation count. These are the scales you are most likely to encounter in published literature — a reliable starting point for high-frequency constructs like brand attitude, purchase intention, self-esteem, and perceived quality.
Yes — use the dropdowns in the toolbar when browsing all scales. You can filter by domain (e.g. Consumer Psychology & Marketing, Clinical Psychology), scale type (Clinical, Experimental, Positive, Consumer), and sort by citation count, year, number of items, or reliability (α).
It means the scale is freely available for research use — no licence or purchase required. Scales without this badge require a licence or purchase from the publisher before use in research. Always check access requirements before including a scale in a study.
An Anthropic API key is a personal credential that gives you access to Claude AI models. To get one: (1) Go to console.anthropic.com and create an account. (2) Navigate to 'API Keys' and create a new key. (3) Copy the key (it starts with sk-ant-...) and paste it into PsychScales using the ⚙ button in the header. API usage is charged per use — typical searches cost a fraction of a cent. New accounts receive free credits to get started.
Your API key is saved only in your browser's local storage on your device — it is never transmitted to any server other than the Anthropic API directly when you run a search. PsychScales has no backend server and collects no data. Clear your key at any time using the Clear button in the ⚙ settings panel.
Yes — the full curated database, all keyword search, filtering, sorting, the Top 20 list, domain browsing, the Projects page, and all scale detail cards work without any API key. The AI key is only required for the intelligent construct-matching search (pressing Enter) and for generating profiles of scales not yet in the database.
When your search finds no matches in the curated database, the AI draws on its training knowledge to generate a psychometric profile for that scale — including authors, items, psychometrics, and an APA citation. These profiles are generally accurate for well-established scales, but you should always verify details against the original publication before using the scale in research. The profile is clearly labelled as AI-generated.
A few possibilities: (1) The scale may be very new or highly specialised — try rephrasing with the full scale name or first author's surname. (2) Check your API key is entered correctly in the ⚙ settings panel. (3) If you get an error message, check your Anthropic account has available credits at console.anthropic.com.
The Design Assistant is an AI-powered tool that helps you plan your study from scratch. You can either describe your research question or hypothesis and get recommended scales, a suggested study design, and testable hypotheses — or start from a set of scales you have already chosen and get design suggestions for how to use them together.
Select 'Start from a research question', type your hypothesis or research question in as much detail as possible (include your population, key variables, and any design preferences), then click Generate Design. The AI will recommend scales from the PsychScales database, suggest a study design, and generate testable hypotheses.
Select 'Start from scales I have chosen', enter your scale names or abbreviations (e.g. PHQ-9, RSES, MAAS), and click Generate Design. You can also click the Import button to automatically pull in all scales from your active project. The AI will suggest the best study design for those scales and recommend any complementary scales you may have missed.
Yes — each recommended scale card has an Add button that stars it directly into your active project. You can also click 'Add all recommended to project' at the bottom to add all database-matched scales at once, then switch to the Projects view to review your full battery.
Click the ☆ star icon on any scale card to add it to your active project — the star turns gold (★) when it's been added. Alternatively, open any scale's detail panel and click the 'Add to Project' button at the top. You can create multiple projects for different studies and switch between them using the tabs on the Projects page.
It is a rough estimate based on approximately 0.4 minutes per item — a common rule of thumb from survey methodology research. Actual completion times vary depending on item complexity, response format, and population. Use this as a planning guide only.
In the Projects page, scroll to the Reference List section and click '📋 Copy All References'. This copies all APA 7th edition citations for your selected scales to your clipboard, ready to paste into a Word document, reference manager, or manuscript.
It depends on the scale's access status. For freely available scales, PsychScales shows the complete set of items — you can use these as a reference, and the green Download button links to the official instrument. For licensed or proprietary scales, representative items are shown to illustrate the scale's content and style; you must obtain the full instrument from the publisher before use in research. In both cases, always cite the original publication in your method section.
Cronbach's alpha is the most common measure of internal consistency reliability — how well the items in a scale measure the same underlying construct. As a rough guide: α ≥ 0.90 = excellent, 0.80–0.89 = good, 0.70–0.79 = acceptable, below 0.70 = questionable for most research purposes. Note that alpha is sensitive to the number of items — very short scales (2–3 items) often have lower alpha even when valid. PsychScales applies a stricter standard than this general field convention: scales are generally required to report α ≥ 0.80 (or comparable composite reliability) to be included in the database at all, with limited, clearly flagged exceptions for seminal, highly-cited legacy instruments.
The 0.70 figure is widely treated as a universal recommendation from Nunnally (1978), but that is a partial reading of his original guidance. Nunnally distinguished between different stages and purposes of measurement: a reliability of 0.70 was offered as a floor for early-stage, exploratory scale development, where the priority is efficiency rather than precision. For what he termed basic research — work involving genuine comparisons between measures, groups, or treatment effects — he specified 0.80 as the appropriate standard, and for applied settings where individual scores inform consequential decisions, he argued for 0.90 or higher. PsychScales exists to support serious, reusable measurement rather than preliminary scale development, so it adopts Nunnally's basic-research standard of 0.80 rather than the exploratory-stage floor that has, somewhat by convention rather than justification, become the field's default citation. An alpha of 0.70 still admits a meaningful share of measurement error into a study's conclusions; requiring 0.80 meaningfully tightens that margin. This threshold reflects PsychScales' broader curation philosophy: the aim is not to catalogue every instrument that has appeared in print, but to surface the subset of measures precise and well-validated enough to be trusted as a genuine starting point for research — not merely a publishable one.
Average Variance Extracted (AVE) is a measure of convergent validity used in structural equation modelling (SEM). It reflects how much variance is captured by the construct relative to measurement error. AVE ≥ 0.50 is generally considered acceptable. Many classic scales predate SEM and were validated with other methods — hence 'not reported' for older instruments.
In a reflective scale, the underlying construct causes the item responses — indicators are interchangeable, expected to correlate highly, and removing any single item shouldn't fundamentally change what's being measured. This is the standard model behind most Likert-type scales, and it's the assumption Cronbach's alpha itself depends on. A formative scale runs in the opposite direction: the items are causes rather than effects, jointly forming the construct from distinct facets rather than reflecting a single underlying variable. Formative items are not expected to correlate with one another, which means alpha and similar internal-consistency statistics are not meaningful quality signals for them — a composite index of socioeconomic status built from income, education, and occupation is a classic formative example. Because a high alpha reported for a scale explicitly labelled 'formative' is a red flag for measurement-model mislabelling rather than a mark of quality, PsychScales treats such cases as ones requiring closer review before inclusion.
Try the AI search: enter the full scale name or first author's surname and press Enter (API key required). If it's still not found, try the PsycTESTS database (APA), the Mental Measurement Yearbook (Buros), or MIDSS (Measurement Instrument Database for the Social Sciences at midss.org). You can also email the scale's authors directly — most academic researchers are happy to share their instruments for research use.
Free access to 2383+ psychometric scales at PsychScales