Your whole study, in one place
Plan, collect, analyse and write up — qualitative, quantitative, mixed methods or a systematic review with meta-analysis — without juggling separate programs.
- 40+ languages, including low-resource ones
- Statistics checked against R
- Audio deleted after transcription
Cross-sectional household survey · single proportion
428households
50% expected · 95% confidence · ±5% precision · 10% non-response
- Study brief
- Sample size
- Allocation across 4 kebeles
- Kobo/ODK form · 46 questions
One place
Everything a study needs, in one project
Each study keeps its plan, data, analysis and report together — shared with your co-researchers if you choose.
- Usually: A sample size calculator20 methods, each with its formula and referenceHow it works
- Usually: Word for the questionnaire, then a form builderQuestionnaire checks, translation, CVI — and the Kobo/ODK form built for youHow it works
- Usually: A transcriber and a translatorTranscripts in 40+ languages, translated line by lineHow it works
- Usually: NVivo or ATLAS.tiCodebook, coding, themes and findings with real quotesHow it works
- Usually: Stata, SPSS or RTests, regression, survival and survey weights — with the Stata, R and SPSS commandsHow it works
- Usually: A reference manager and WordA literature review citing only real papers, and the full report in WordHow it works
Qualitative
Interviews and focus groups — transcribed, coded, written up
Upload or record, then code where you transcribed.
40+ languages, including low-resource languages most tools handle poorly: አማርኛ Amharic · Afaan Oromoo · ትግርኛ Tigrinya · Somali · English · French · العربية Arabic · Swahili · Hausa · हिन्दी Hindi · Spanish · Portuguese · Yoruba and more
- Every speaker labelled and every line timed, in the recording’s own script — or bring transcripts from Word or PDF
- Translated into English line by line, with summaries and anonymised copies
- Codebook, coding and AI suggestions you accept or reject; a second coder with Cohen’s kappa
- Themes, a framework matrix and a findings draft with real quotes — or export to NVivo, ATLAS.ti and MAXQDA
The made-up focus group, coded. The dashed passage is an AI suggestion to accept or reject.
- 00:02M:እንኳን ደህና መጣችሁ። ዛሬ በአካባቢያችሁ ስላለው የውሃ እና የጤና አገልግሎት እንወያያለን።
- 00:15P1:አመሰግናለሁ። ስሜ አልማዝ ነው። በእኛ መንደር ውሃ ለማምጣት በቀን ሁለት ሰዓት እንጓዛለን።Time spent fetching water
- 00:31P2:አዎ፣ በተለይ በበጋ ወቅት ምንጩ ስለሚደርቅ በጣም ከባድ ነው።Seasonal scarcity
- 00:48M:ይህ በልጆች ትምህርት ላይ ምን ተጽዕኖ አለው?
- 01:02P3:ልጆች ውሃ ለመቅዳት ስለሚሄዱ ብዙ ጊዜ ትምህርት ቤት ያረፍዳሉ።Effects on schooling
- 01:25P2:[laughter] እኔ ራሴ በልጅነቴ እንደዚያ ነበርኩ።
- 02:10M:ስለ ጤና ጣቢያው የአገልግሎት ሰዓት ምን ትላላችሁ?Limited opening hours
- 02:24P1:ጠዋት ብቻ ነው የሚከፈተው፣ ከሰዓት በኋላ ዝግ ነው።
- 02:41P4:ማታ ሰው ቢታመም ወደ ደብረ ብርሃን ከተማ መሄድ አለብን፣ ትራንስፖርትም የለም።No transport at night
- 03:05M:ለመፍትሄ ምን ትጠቁማላችሁ?
- 03:18P3:የውሃ ጉድጓዱ ቢጠገን ብዙ ችግር ይቀንሳል።Repair water points
- 03:40P2:ጤና ጣቢያው ቢያንስ እስከ ማታ ድረስ ክፍት ቢሆን ጥሩ ነበር።Longer opening hours
- 05:02M:በቀበሌ ስብሰባዎች ላይ ይህን ጉዳይ አንስታችሁ ታውቃላችሁ?Requests not acted on
- 05:20P1:ብዙ ጊዜ አንስተናል፣ ግን እስካሁን ምንም ለውጥ የለም።
- 10:05P4:ሴቶች በውሃ ጉዳይ ውሳኔ ላይ መሳተፍ አለባቸው።Women in decisions · AI?
- 10:30M:በጣም አመሰግናለሁ፣ ለዛሬ በዚሁ እንጨርሳለን።
Regressionmade-up example
Water within 30 minutes: factors (binary logistic, n = 412 households)
| Variable | Category | AOR (95% CI) |
|---|---|---|
| Household wealth | Poorest | 1 |
| Richest | 2.41 (1.52, 3.83) | |
| Head’s education | None | 1 |
| Secondary or more | 1.78 (1.09, 2.90) | |
| Residence | Rural | 1 |
| Urban | 3.12 (1.96, 4.97) | |
| Household size | per person | 0.91 (0.84, 0.99) |
Also: modified Poisson (APR), multinomial, ordinal, multilevel (ICC, MOR), Poisson and negative binomial, survival with Cox — with survey weights.
Quantitative
Surveys — from sample size to the regression table
Every calculation with its formula and reference, and every result checked against R.
- Sample size with 20 methods — surveys, case-control, cohort, trials, diagnostic tests, survival, cluster designs
- A questionnaire with answer codes and skip rules, wording checks, translation with back-translation, the content validity index — and its Kobo/ODK form
- Data from KoboToolbox, ODK, SurveyCTO or Excel: checks, tables with 95% CIs and the fitting test
- Logistic, modified Poisson, multilevel and survival models and survey weights — with Stata, R and SPSS commands
- Scale validation: exploratory and confirmatory factor analysis, reliability (α, ω), convergent and discriminant validity, and structural equation models
Mixed methods and write-up
Bring it together and write it up
Both strands side by side, the papers that matter, and the whole report in Word.
- Convergent, explanatory or exploratory designs: a joint display and the integrated findings, written up with quotes
- A literature review and discussion from real papers (PubMed, Crossref) — references built from their records, never invented
- Systematic reviews: screening by two reviewers, PRISMA 2020, risk of bias and meta-analysis checked against R’s metafor
- The full report — background, methods, results, discussion, references — in Word
The report, written upmade-up example
Results · Integrated findings
Long walks for water keep children late for school
Just over half of households (52%, n = 412) had a water source within 30 minutes, and 27% of children aged 7–14 were late for school because of water (n = 624). The focus groups explain why: in the dry season the spring dries up, and children fetch water before school.
“Children are often late for school because they go to fetch water.”— P3, FGD-01 (translated from Amharic)
The strands agree on access, and the interviews add what the survey could not: why the walk is longest in the dry season.
Trust
Research you can defend
Statistics checked against R
Every model, test and meta-analysis matches R — and comes with the Stata, R and SPSS commands to reproduce it.
References never invented
Citations and the reference list are built from each paper’s record, in APA 7 or Vancouver — never written by AI.
You approve every AI step
AI codes, drafts and translations wait for you to accept or edit them; the methods text says exactly how AI was used.
Your data stays yours
Audio is deleted after transcription, projects are private unless you share them, and nothing is used to train AI.
Prices
Start free, pay for what you use
Free to start
$0
20 minutes of transcription and 2 AI runs — planning and data analysis free to use
Pay as you go
$8
per audio hour; hours never expire
Monthly plans
from $29
audio hours and the AI tools every month
Study Pass
$39
the AI tools for one study, for 12 months
Questions
Common questions
Once you’re signed in, Help and the step-by-step guides have more. Or write to us from the Contact page.
What is free?
Planning (the sample size calculator needs no account at all), the questionnaire builder with its checks and the Kobo/ODK form built from it, coding by hand, survey data analysis, finding papers, the full report and every export. To try the rest: 20 minutes of transcription and 2 AI analysis runs, no card needed.
Compare plansWho can see my recordings and data?
Only you, and any co-researchers you invite to a project. Audio is deleted after transcription — within 48 hours at the latest — survey data is read when you open it and never stored, and we use a paid AI service that doesn’t use your content to train or improve its products.
Read the Privacy PolicyWhich languages work best?
Amharic, Afaan Oromoo, Tigrinya and Somali, each in its own script, and English, French, Arabic, Swahili and dozens more. Recordings that switch between languages are handled too.
See the languagesDoes the AI do the analysis for me?
It does the heavy lifting — suggesting codes, drafting themes and findings with real quotes, drafting a questionnaire or a translation — and you stay in charge: you accept or reject each suggestion, and your methods text states exactly how AI was used, as journals and ethics boards expect. You can also do it all by hand.
How AI coding worksCan it analyse the survey data I collected?
Yes — from KoboToolbox directly, or any Excel or CSV file (ODK Central, SurveyCTO, registers). It checks the data, gives tables with 95% confidence intervals, runs comparisons with the fitting test, and models it: logistic, modified Poisson, multinomial, ordinal, multilevel, linear, Poisson and negative binomial regression, survival analysis and Cox regression, and scale validation (factor analysis, CFA and SEM) — with sampling weights, clusters and strata if your survey has them.
How quantitative analysis worksCan I do a systematic review and meta-analysis?
Yes: a PICO protocol and search string, two-reviewer screening with optional AI suggestions, the PRISMA 2020 diagram, extraction and risk of bias, the meta-analysis with its forest plot, subgroups and Egger’s test, GRADE, and the review report in Word. The statistics match R’s metafor.
How a systematic review worksWhat does it cost after the free start?
Transcription is $8 per audio hour (hours never expire), or a monthly plan from $29. The AI tools for a study — AI coding and themes, Ask your interviews, the findings draft, the questionnaire draft and translation, Kobo/ODK forms from uploaded questionnaires, the mixed-methods and literature drafts — come with a Study Pass ($39 for one study, 12 months) or a monthly plan.
See pricingStart your next study here
Free to start — no card. The sample size calculator needs no account at all.