Docs
Methodology and User Guide
Everything you need to go from raw data to a trustworthy report, explained in plain English.
Download the full guide (PDF)What Veridiqx does
Veridiqx is a research and data analysis platform. It takes you from collecting data to a finished report in one place, following four stages: Collect, Inspect, Analyze and Report.
Getting started
- Create a free account and sign in.
- Create a project to keep your surveys, datasets, analyses and reports together.
- Collect data with a survey, or upload a file you already have.
- Inspect the data, run an analysis, then build a report you can download or share.
1. Collect
Gather responses directly in Veridiqx or bring in data you already hold.
- Survey builder
- Write questions yourself or let AI draft them. Only the question types you select are generated.
- Editing AI questions
- Change any generated question and its answer choices (A, B, C, D). Add or remove choices; duplicates are blocked and choice questions need at least two options.
- Question types
- Single choice, multiple choice, rating and Likert scales, open text and ranking.
- Ranking questions
- Each option gets a unique rank from 1 to N, and incomplete rankings cannot be submitted.
- Saving and publishing
- Saving and then publishing update the same survey, so you never end up with a duplicate draft.
- Sharing
- Every survey has a public link and a QR code. Downloadable survey PDFs are unbranded.
- Paper form scanner
- Photograph a completed paper form and Veridiqx reads the answers into your data. Always check scanned values.
- Survey to dataset
- Turn collected responses into a dataset in one click. It stays linked to the same project.
- Offline collection
- If the connection drops, answers are saved on the respondent's device and sent automatically when it returns. The survey link must be opened once while online.
2. Inspect
Understand and clean your data before testing anything.
- File upload
- CSV, Excel (.xlsx and .xls) and JSON files are supported.
- Codebook
- Short codes and numeric values are translated into readable labels for survey data.
- Quality checks
- Missing values, duplicates and unusual values are flagged so you can decide what to do.
- Outlier detection
- Unusual numbers are found with the standard interquartile range rule and shown to you, not deleted.
3. Analyze
Pick a test and the columns to use. If a test does not fit your data, Veridiqx explains why instead of returning a misleading result.
- Descriptive statistics
- Mean, median, spread and range for numeric columns.
- Frequency tables
- Count, percent and cumulative percent with a distribution chart.
- Cross-tabulation
- Compare two categorical questions in one table.
- Likert scale analysis
- Summaries for agreement and rating scales.
- Independent t-test
- Compares the average of two groups. Reports Cohen's d as the effect size.
- Paired t-test
- Compares two measurements from the same people, such as before and after training. Reports the mean change, its 95% confidence interval and Cohen's d.
- Multiple Regression
- Tests how several predictors together explain one numeric outcome. Reports R², adjusted R², the F test, and for each predictor B, standardised β, p, 95% CI and VIF (a check that predictors don't overlap too much).
- Binary Logistic Regression
- Predicts a yes/no outcome (e.g. adopted / did not adopt). Reports odds ratios with 95% CI, Wald tests, Nagelkerke R² and how many cases the model classifies correctly.
- Two-way ANOVA
- Tests the effect of two grouping factors (e.g. gender and region) and their interaction on a numeric outcome, with partial η².
- Wilcoxon Signed-Rank
- The non-parametric version of the paired t-test, for before/after scores that are not normally distributed or are ordinal.
- Difference-in-Differences
- Economics impact evaluation: compares how a treated group and a control group changed from before to after a programme.
- Elasticity (log-log)
- Shows the % change in one variable for a 1% change in another, e.g. price elasticity of demand.
- Panel Data (Fixed & Random Effects)
- For the same countries, firms or people measured over several periods. Gives fixed and random effects estimates, an F test for entity effects and the Hausman test that tells you which model to report.
- Instrumental Variables (2SLS)
- Corrects bias when a predictor is tangled with the outcome (reverse causality, missing variables). Reports the first-stage F (instruments below 10 are weak), 2SLS and OLS side by side, and the Wu-Hausman test.
- Unit Root Test (ADF)
- Checks whether a time series such as inflation or GDP is stationary before you model it. Chooses lags automatically, compares with MacKinnon critical values, tests the first difference and adds Durbin-Watson.
- Cointegration & Error Correction
- For two trending series (e.g. income and consumption). The Engle-Granger test checks for a real long-run link, then an error correction model shows the short-run effect, the speed of adjustment back to equilibrium and its half-life.
- ARIMA Forecasting
- Picks the best ARIMA(p,d,q) model automatically (ADF for differencing, AIC for the rest), checks residuals with the Ljung-Box test, reports RMSE and MAPE and forecasts 12 periods ahead with 80% and 95% intervals. Needs at least 24 time points.
- Exploratory Factor Analysis
- Finds the hidden dimensions behind questionnaire items. Uses principal axis factoring with varimax and promax rotation, flags cross-loading and weak items, and reports KMO and Bartlett's test. Needs at least 30 complete rows.
- Mediation & Path Analysis (SEM)
- Tests whether X affects Y through a mediator M. Gives the a, b, c and c′ paths (raw and standardized), the indirect effect with a 2,000-resample bootstrap CI, the Sobel test and SEM fit indices (χ², CFI, TLI, RMSEA, SRMR).
- Gini, Lorenz & Inequality
- Measures how unequally income or wealth is shared: Gini coefficient, Lorenz curve shares, quintile shares, Theil index and Palma ratio.
- Principal Component Analysis
- Reduces many related variables into a few components, with KMO and Bartlett's test. Useful for building indices such as a wealth or wellbeing index.
- One-way ANOVA
- Compares the averages of three or more groups. Reports eta squared.
- Chi-square test
- Checks whether two categorical variables are related.
- Pearson correlation
- Measures how strongly two numeric columns move together.
- Linear regression
- Estimates how one or more columns predict an outcome.
- Text sentiment
- Summarises the tone of open text answers.
- Trend forecast
- Projects a simple trend forward for ordered numeric data.
Reading your results
- p value
- A small p value (commonly below 0.05) suggests the pattern is unlikely to be chance alone. It does not tell you how big or important the effect is.
- Effect size
- Shows how large a difference or relationship is. Veridiqx labels it small, medium or large using common guidelines.
- Confidence interval
- A range that likely contains the true value. Narrow ranges mean more precise estimates.
- AI interpretation
- Plain English summaries are drafts. Numbers are computed from your data; always review the wording before you rely on it.
Assumption checks and APA 7th results
After a test runs, Veridiqx reviews whether the test suits your data and writes the result the way journals expect.
- Assumption review
- Normality, equal variances, expected cell counts and sample size are checked and marked Met, Caution or Violated, with a suggested alternative test when needed.
- APA paragraph
- A results sentence in APA 7th style, for example t(40) = -2.27, p = .029, d = -0.70, 95% CI [-1.32, -0.08]. Copy it with one click.
- APA table
- A three-line APA table you can copy as a Word table or as LaTeX.
Power and sample size planner
Open Power Planner in the left menu to plan a study before you collect data.
- Sample size
- Find how many participants you need for a chosen effect, significance level and power.
- Achieved power
- Check how likely your study was to detect the effect with the sample you had.
- Smallest effect
- Find the smallest effect your sample can reliably detect.
- Supported tests
- Independent and paired t tests, one-way ANOVA, regression, correlation and chi-square. Results match G*Power reference values. A live power curve and an APA justification paragraph are included.
Reproducibility certificates
- Issuing
- On any result, click the certificate button. Veridiqx records digital fingerprints (SHA-256) of the test settings and the results.
- Sharing
- Download the certificate as a PDF with a QR code, or copy its link.
- Verifying
- Anyone can open the verification page, see the summary and paste results to confirm they match. Certificates do not contain your raw data.
Peer review and defense check
Prepare for a journal reviewer or thesis committee before you submit.
- Rigour score
- A score out of 100 built from sample size, assumptions, reporting and the number of comparisons, with a verdict from Submission ready to High risk.
- Likely criticisms
- The points a tough reviewer would probably raise, each with a response you can copy and supporting references.
- Defense practice
- Five likely questions with model answers. Rehearse mode hides the answers until you have tried your own.
- Summary
- Download a review summary to share with your supervisor. It is guidance, not a guarantee of acceptance.
Scale Lab
Open Scale Lab in the left menu and pick a dataset to prepare survey scales before testing.
- Assumptions
- Check normality, extreme values (more than 3 standard deviations from the mean) and sample size for many variables at once.
- Reliability
- Cronbach's alpha for a group of items, with item-total correlations and alpha if each item is deleted. An alpha of 0.70 or higher is commonly treated as acceptable.
- Recode and composites
- Reverse-code negatively worded items, combine items into a sum or mean score, or standardise a column into z-scores, then download the updated data.
- Missing data
- See how much is missing per column and run Little's MCAR test. A p value above 0.05 suggests the gaps are random.
- Correlation matrix
- Pearson or Spearman correlations between all chosen variables, with means, standard deviations and stars (* p < .05, ** p < .01, *** p < .001).
4. Report
- Report builder
- Combine findings, charts and interpretation into one document.
- Downloads
- Export reports as PDF, editable Word (.docx) with APA-style tables, or PowerPoint.
- Share links
- Send a private link to supervisors or stakeholders. Anyone with the link can view it.
- Charts and dashboards
- Every chart includes a key. Dashboards can be shared for viewing.
- Code guide
- Get code that repeats your analysis in Python, R, SQL, Stata or SPSS syntax.
Working as a team
- Admin
- Manages the project, its data and who has access.
- Analyst
- Adds data, runs analyses and creates reports.
- Viewer
- Can read everything in the project but cannot change it.
- Comments and activity
- Discuss work inside a project and see what changed and when.
Your account and your content
- Notifications
- The bell at the top shows new survey responses from the last seven days, grouped by survey. Click one to open it, or mark all as read.
- Deleting content
- You can delete your own surveys and datasets. Their responses, charts and saved analyses are removed with them.
- Deleting your account
- In Settings, type DELETE to confirm. Your projects, data, files and sign-in are permanently removed.
- Plans
- See the Pricing page for the Free, Professional and Enterprise plans.
Privacy and data handling
- Ownership
- You keep ownership of your data. It is not used to train AI models.
- Access control
- Each account can only reach its own projects and the projects it was invited to.
- What is stored
- Projects, surveys, datasets, saved reports and contact messages. Exploratory analyses run without saving results.
- Your responsibility
- Tell respondents how their answers will be used and follow the privacy laws that apply to you.
The Veridiqx logo and colours
The logo mark shows the research journey from raw data to a defensible finding. Each shape and colour has a meaning.
- Rounded square
- The frame that holds the work: research integrity and confidentiality.
- Two lower dots
- Your two data sources: surveys you collect and datasets you upload.
- Base line
- The sample your evidence rests on.
- Rising lines
- Raw data moving upward into testing. Nothing reaches the top without passing through the centre.
- Ringed centre dot
- The statistical engine: tests, effect sizes, confidence intervals and assumption checks.
- Top dot
- The finding you can stand behind: interpretation, APA results and reports.
- Faint crossing lines
- The x in Veridiqx: the variable under test.
- Sage #3F6B4F
- Main colour. Calm, grounded and trustworthy, like careful field notes.
- Coral #FF6F61
- Accent colour. Discovery and the points that need your attention.
- Paper #FAFAF7 and Surface #EEF0E8
- Soft backgrounds that feel like a notebook and are easy on the eyes.
- Ink #1F2421
- Text colour, chosen for strong contrast and easy reading.
- Type and tagline
- The Veridiqx name is set in Georgia for scholarly character; all other text is Inter for clear screen reading. Tagline: Where data becomes truth.
Install Veridiqx on your phone
Veridiqx can sit on your home screen and open like an app, with its own icon and no browser bar. No app store is needed.
- Android (Chrome)
- Open veridiqx.tech in Chrome. Tap the three-dot menu at the top right, choose Install app or Add to Home screen, then tap Install.
- iPhone and iPad (Safari)
- Open veridiqx.tech in Safari. Tap the Share button (a square with an arrow), scroll down, choose Add to Home Screen, then tap Add.
- Computer (Chrome or Edge)
- Open veridiqx.tech and click the install icon at the right of the address bar, then click Install.
- Good to know
- The installed app always shows the latest version. Survey answers still save on the device when the connection drops and send once it returns.
Common questions
- Do I need to know statistics?
- No. Veridiqx explains each test and its result in plain English, but basic understanding helps you judge the findings.
- Can I use it on my phone?
- Yes. The app works on phones, tablets and computers. Scan the QR code on the site to open it on mobile.
- Can I install it on my home screen?
- Yes. On Android, open veridiqx.tech in Chrome, tap the three-dot menu and choose Install app or Add to Home screen. On iPhone, open it in Safari, tap Share, then Add to Home Screen.
- Why did an analysis refuse to run?
- The columns chosen do not suit that test, for example text where numbers are needed. The message tells you what to change.