In the study area, how do households' drinking water practices (source choice, storage and treatment) compare with their beliefs about whether their water is safe, and what observable indicators support or challenge those beliefs?

Most families in Bangladesh believe their drinking water is safe, or hope it is. Very few have ever seen a test result. This project does not try to test water in a laboratory. Instead it studies the three things a student team can measure well in the field: where drinking water comes from, what households do to it between the source and the glass, and what people believe about its safety and why.

Your team will survey 50 or more households about their drinking water source, storage, treatment and beliefs. You will observe and photograph the path water takes in a sample of homes, from pump or tap to the cup, and record simple observable indicators such as clarity, smell, container type and cover. You will interview people who know the local water: a tube-well mechanic, a filter seller, a health worker, a jar supplier, and families. An optional part uses low-cost field indicators (TDS meter, turbidity tube, arsenic test strips where arsenic is a known concern) to add numbers, with a firm rule that none of these can show water is safe to drink.

The result is a paper that maps how drinking water quality is produced and protected, or lost, inside a real community, and how far trust and practice match each other. The awareness outcome is a one-page "Source to Glass" checklist for households. The study works in any month of the year.

Bangladesh has done more than almost any country to bring drinking water close to people: tens of millions of tube wells, growing piped supply in cities, and a large private market for filtered and bottled water. But quality is a different problem from access. Groundwater in many districts contains arsenic or iron; shallow wells near latrines or drains can carry bacteria; piped water can be clean at the plant and dirty by the time it reaches a rooftop tank; and clean water can be spoiled by a dirty jug or an uncovered pot inside the home.

Households know some of this and act on it: they boil, filter, use alum, buy jars, or choose one tube well over another for taste. These actions cost money and time, and nobody knows whether they are the right ones for that family's water. At the same time, people's trust in a source often comes from habit, taste or the seller's word, not from any test. A study that records both practice and belief in one community shows where small changes, such as covering a container or changing a filter cartridge, could matter most. That is useful to your neighbours, and it is exactly the kind of local evidence that larger studies cannot collect.

Arsenic in groundwater was identified as a national problem in Bangladesh in the 1990s, and large surveys found elevated arsenic in tube wells across many districts, with the national standard for arsenic in drinking water set at 50 micrograms per litre [verify: Bangladesh drinking water standard for arsenic and the WHO guideline value, cite an official or WHO source]. Tube wells were painted red or green in many areas after testing, though paint fades and new wells are often untested [verify: current status of well-marking programmes in your district].

Microbial contamination is the other main concern. Studies have found that water that is clean at the source can be contaminated by the time it is drunk, because of dirty containers, hands and storage practices; this is often called "recontamination at the point of use" [verify: find one peer-reviewed study on point-of-use contamination in Bangladesh, Google Scholar]. Household treatment options promoted in Bangladesh include boiling, ceramic and membrane filters, chlorine tablets and solar disinfection [verify: a WHO, UNICEF or DPHE source listing recommended household treatment methods].

Cities like Dhaka and Chittagong supply piped water through WASA systems, with quality monitored at treatment plants; quality at the tap and in building tanks depends on pipes and tank maintenance [verify: a news report or WASA statement on tap water quality or tank cleaning]. Sylhet and Noakhali have different source mixes, including deep tube wells, shallow wells and, near the coast, salinity problems [verify: a DPHE or academic source on groundwater salinity in coastal Noakhali].

Your paper needs only three or four verified sources, cited correctly. A single reliable source for each of arsenic, point-of-use contamination and household treatment is enough.

Primary research question: In [your study area], how do households' drinking water practices (source choice, storage and treatment) compare with their beliefs about whether their water is safe, and what observable indicators support or challenge those beliefs?

Secondary questions:

1. What share of households treat their drinking water, by which method, and how consistently?
2. Where along the path from source to glass do observable risks appear (uncovered storage, shared dipping cups, old filter cartridges, untested wells)?

Hypothesis: Most households (more than 60%) believe their drinking water is safe, but fewer than half use a consistent treatment method, and at least one observable risk will be found in most homes observed. Trust in the source will be based mainly on taste, habit and the seller's word, not on any test.

Your data may show the opposite, for example a community that treats carefully and distrusts its source. Report what you find.

Team size: 2 to 5. The home observation part needs two people at each visit for safety and for accurate recording, so this project needs a team of at least two.

Suited regions: All four, with different emphasis:

• Dhaka and Chittagong: piped supply, rooftop tanks, jar and filter markets.
• Sylhet: mixed tube-well and piped supply, iron and taste issues [verify].
• Noakhali: tube wells, salinity in some unions, arsenic in some areas [verify local DPHE data]; the optional arsenic strip test fits best here if a known arsenic area is nearby.

Suggested roles:

• Survey lead: builds and runs the AKTI Survey, recruits households, exports data.
• Observation lead: runs the source-to-glass observation in 15 to 20 homes with a partner, fills the observation sheet, takes photos.
• Interview lead: arranges and records 6 to 8 interviews, keeps the consent log.
• Indicator lead (optional): runs TDS, turbidity and, where relevant, arsenic strip readings and records them with the sample's source and time.
• Writer and analyst: builds the tables and charts, drafts the paper. Can be shared.

Every claim in your paper rests on evidence you collected. The three required evidence types are joined by a structured observation sheet, which is the heart of this study.

Why each fits the question: the survey measures belief and reported practice; observation measures real practice and visible risk; interviews explain the reasons; photos show readers the path from source to glass; indicators add numbers but never a verdict on safety.

Build this survey on AKTI Survey in Bangla with an English copy for the paper. Keep it under 8 minutes. Pilot it with 3 to 5 people and fix anything that confuses them.

Target respondent: one adult (18 or older) per household who usually handles drinking water, often the person who cooks. Do not survey children.

Recruitment: every second household on 3 to 5 streets, lanes or paras in your study area, in daylight. Record refusals and absences. If you include relatives or friends, mark them.

Sample size target: 50 completed responses.

Draft questions (ready to paste):

1. Which area do you live in? (short answer: ward, mahalla, para or village)
2. How many people live in your household, and how many are children under 5? (two numbers)
3. What is your main source of drinking water? (single choice: piped supply / own tube well or pump / shared or public tube well / 20-litre jar / bottled water / pond, river or rainwater / other)
4. Is this the same source you use for cooking? (yes / no, cooking source: ...)
5. If you use a tube well, is it marked with paint (red, green) or a tag? (single choice: red / green / other mark / no mark / don't know / not applicable)
6. Has your drinking water source ever been tested, as far as you know? (single choice: yes, I saw a result / yes, I heard it was tested / no / don't know)
7. Do you treat your drinking water before drinking? (single choice: always / sometimes / never)
8. If you treat it, how? (multiple choice: boiling / filter jug or candle filter / electric or membrane filter / chlorine tablets or liquid / alum / letting it settle / other)
9. If you use a filter, when did you last clean it or change the cartridge? (single choice: within a month / 1–6 months / more than 6 months / never / don't know)
10. If you boil, for how long after it starts bubbling? (single choice: less than 1 minute / 1–5 minutes / more than 5 minutes / don't know)
11. How do you store drinking water at home? (multiple choice: covered pot or kolshi / uncovered pot / plastic bottle / filter container / jar as delivered / refrigerator bottle / other)
12. How do you take water from the container? (single choice: tap on the container / pouring / dipping a cup or glass / dipping a hand / other)
13. How often is the storage container washed? (single choice: daily / weekly / monthly / rarely / don't know)
14. Do you think your drinking water is safe to drink? (single choice: yes, definitely / probably / not sure / probably not / no)
15. What makes you think so? (multiple choice: taste or smell / it looks clear / it has always been fine / a test result / the seller or supplier says so / a health worker said so / neighbours use it / I treat it / other)
16. In the last 12 months, has anyone in your household had diarrhoea or stomach illness that you connected with water? (single choice: yes / no / not sure) (sensitive: optional to answer)
17. Has your household ever changed drinking water source because of quality, taste or smell? (yes / no; if yes, from what to what)
18. How much does your household spend per month on treating or buying drinking water (filter, fuel, jars, bottles)? (number in BDT, or 0)
19. Would you like your water tested if it were free? (yes / no / don't know)
20. What is the one thing you most want to know about your drinking water? (short answer)

Questions 14 and 15 against questions 7 to 13 are the heart of the study: what people believe, and what they actually do.

Your survey and observations describe the households you reached, not your city, your district or Bangladesh. Say so in the paper.

• Who to recruit: adults who handle water, in a defined area, chosen by a rule (every second household). Aim for a mix of housing types; storage and treatment differ between a flat with a filter and a rented room with one kolshi.
• Observation homes: choose 15 to 20 homes from among survey respondents who agree to a short visit. These are volunteers, so they may be more careful than average. Say this in the limitations.
• What the sample can claim: the share of your respondents who treat water, the common storage practices, the visible risks found in observed homes, and the gap between belief and practice in your sample.
• What it cannot claim: that any water is safe or unsafe; disease rates; district-level patterns; or comparisons with areas you did not visit.
• Non-response: record refusals and absences for both the survey and the observation visits.

Record 6 to 8 interviews of 5 to 10 minutes. Interview people in these roles:

• a tube-well mechanic or the person who installs or repairs pumps locally
• a filter seller or a shop that sells filters and cartridges
• a jar supplier or refill plant worker
• a community health worker, pharmacy worker or clinic staff member
• a household that treats water carefully
• a household that does not treat water
• a building caretaker who manages a rooftop tank (cities)
• if possible, a DPHE, WASA or pourashava water staff member

Sample questions (choose 6 to 8 per interview):

1. Where does the drinking water here come from, and what happens to it before people drink it?
2. What do you see go wrong most often between the source and the glass?
3. Has the water here ever been tested? Who tested it, and what happened after?
4. What do people ask you about water quality? What do you tell them?
5. Which treatment methods do people use, and which do you think work?
6. How do people decide whether a well, a tap or a jar is safe?
7. What changes have you seen in the water in the last few years: taste, colour, iron, salt, pressure?
8. What would you tell a new family moving here about the water?
9. What do you think people misunderstand about water safety?
10. If one thing could be fixed about drinking water here, what should it be?

Consent script (read aloud before recording):

Bangla: আমি রেইনফ্লাওয়ার প্ল্যাটফর্মের একজন ছাত্র গবেষক। আমরা আমাদের এলাকার খাবার পানির মান, পানি শোধন ও মানুষের বিশ্বাস নিয়ে একটি গবেষণা করছি। এই সাক্ষাৎকারটি প্রায় দশ মিনিটের। আপনার নাম বা ঠিকানা আমরা প্রকাশ করব না। আপনি তিনটি বিষয়ে আলাদাভাবে সম্মতি দিতে পারেন: এক, সাক্ষাৎকারে অংশ নেওয়া; দুই, ভিডিও রেকর্ড করা; তিন, রেকর্ডের কোনো অংশ প্রকাশ্যে ব্যবহার করা। আপনি যেকোনো সময় থামতে পারেন। আপনি কি রাজি?

English: I am a student researcher with the Rainflower platform. We are doing a study about drinking water quality, treatment and people's beliefs in our area. This interview takes about ten minutes. We will not publish your name or address. You can agree to three things separately: one, taking part; two, being recorded on video; three, having any part of the recording shared publicly. You can stop at any time. Do you agree?

Log the three answers with date, time and place in your field diary. Use the same script, shortened, before each home observation visit.

Recording tips: phone sideways, light behind you, microphone close, 10 seconds of surroundings first, date and place said aloud at the start.

Turn on location (geo-tagging) in the camera before the first visit. Take 25 or more photos. Log each one: date, time, place, what it shows.

Photo ideas:

1. The main drinking water source: tube well, tap, jar stack or bottle shelf.
2. A tube well's paint mark or tag, or the absence of one.
3. The surroundings of a well or tap: distance to a drain, latrine or rubbish, taken from a few metres away.
4. A household storage container, showing whether it is covered.
5. The cup, jug or tap used to take water from the container.
6. A filter in use, and its cartridge or candle if the family will show it.
7. A boiling pot or the stove used for boiling water.
8. A rooftop tank and its lid, with the building caretaker's permission.
9. A jar refill point or delivery van.
10. Water in a clear glass against a white background, to show clarity (no claims about safety).
11. A filter shop shelf with cartridges and their prices.
12. Your team running an observation visit or reading an indicator, with permission.

Privacy rules: never photograph a child's face. Adults' faces only with permission for public sharing. Inside homes only with permission, and never show anything that identifies the family. Exact home locations stay in the private evidence package; the paper uses the area name only.

Source-to-glass observation sheet (one per home, 10 minutes): record source type and distance from home; well marking; visible surroundings risk (drain, latrine, animals within about 10 metres: yes/no); storage container type; covered (yes/no); how water is taken out (tap, pour, dip); container looks clean (yes/no, your judgement, say what you saw); treatment method in use; filter condition (cartridge date if known, visible dirt); water clarity in a glass (clear / slightly cloudy / cloudy); smell (none / earthy / chemical / rotten / other); any other observation. Aim for 15 to 20 homes.

These are optional. They can make your paper stronger, but the study works without them. Each reading must be tied to a source, a time and a container, and your paper must state that none of these show whether water is safe to drink.

• TDS meter (BDT 800 to 1,500): a number for dissolved solids. High readings can go with salinity or iron; low readings say nothing about bacteria or arsenic. Useful to compare a jar, a tap and a well in the same home.
• Turbidity tube (BDT 500 to 1,500, or a home-made version with a clear 1-litre bottle and a printed black cross): how cloudy the water is. Cloudy water is harder to treat; clear water is not proof of safety.
• Arsenic field test strips (BDT 1,500 to 3,000 for a kit; only in areas where arsenic is a known concern): follow the kit instructions exactly, with gloves, in open air, and let an adult supervise. A strip can show presence above a rough level; it is not a laboratory result. Never tell a family their well is safe or unsafe from a strip; say what the strip showed and recommend a DPHE test. If a reading is high, record it and inform your coordinator so Rainflower can pass it to the right authority.
• Time and distance: time the round trip to the drinking water source and count the trips per day; this is a cost of quality that surveys miss.

Plan for 13 weeks. Your AI research assistant checks each stage. Two stages have a human gate.

Decide this before data collection, and note any change in the paper.

1. Sources. From Q3 to Q6, count households by drinking source, by whether the source is marked, and by whether it was ever tested. Bar chart.
2. Treatment. From Q7 to Q10, report the share who treat always, sometimes, never; the methods used; and, for filter users, how recently the cartridge was changed; for boilers, how long they boil. Define "consistent treatment" as "always" plus a method, and report that share.
3. Storage and handling. From Q11 to Q13 and the observation sheet, count covered versus uncovered storage, dipping versus tap or pour, and washing frequency.
4. Belief. From Q14, report the share who say "definitely" or "probably" safe. From Q15, count the reasons, and report the share whose belief rests on a test result versus taste, habit or a seller's word.
5. Belief versus practice. Cross Q14 with your "consistent treatment" measure and with the observed risks. Report the share of households who believe their water is safe and use no consistent treatment and have at least one observed risk. This is your hypothesis test.
6. Observed risks. From the observation sheet, count each risk type across homes and report the median number of risks per home and the most common one.
7. Indicators (optional). Report TDS and turbidity as ranges and medians by source type. Report arsenic strips as presence counts by well, with the kit's rough level, and state clearly that these are field indicators, not safety results.
8. Interviews. Write 3 to 5 themes with one or two anonymised quotes each.
9. Health question (Q16). Report the count only, with the note that it is self-reported and not a measure of water safety. Do not link it to any single source.
10. Limitations check. Sample size, refusals, volunteer bias in observed homes, and sample-only claims.

Use a spreadsheet for all counts and keep it in the evidence package.

• Three separate permissions for every person: taking part, being recorded, and public sharing. For home observations, ask separately for photographs inside the home.
• Parental agreement for every team member before field work.
• Adults only as respondents and interviewees.
• Sensitive question: Q16 about illness is optional; say so before asking; never ask about a named child's illness.
• No verdicts. Never tell a household their water is safe or unsafe. If an optional reading looks worrying, say what the strip or meter showed, recommend a DPHE or WASA test, and report it to your coordinator the same day.
• Privacy: no names, phone numbers or addresses in the paper. Home locations and faces stay in the private evidence package, kept only as long as review and any journal process need, and deleted on request.
• Physical and chemical safety: work in daylight, in pairs; do not climb to rooftop tanks; do not lean over open wells; use gloves with arsenic kits and dispose of strips as the kit instructs; an adult supervises any chemical test.
• Respect: a family's water practices are not a test they can fail. Thank them, and leave the "Source to Glass" checklist with them at the end of the project if they want it.

Expected outcomes:

• A map of drinking water sources, marking and testing status for the households in your sample.
• Measured shares for treatment, consistent treatment, storage and handling practices.
• A belief-versus-practice table showing how many households trust untreated, untested water, and what that trust rests on.
• A count of observable risks along the source-to-glass path in 15 to 20 homes, with photos.
• 3 to 5 interview themes explaining local practice.
• An optional indicator summary, clearly labelled as field indicators.
• A one-page "Source to Glass" household checklist as the awareness outcome.

Limitations written in advance:

• Results describe one area and one sample.
• Observed homes are volunteers and may be more careful than others.
• Practices are partly self-reported and may be improved for the interviewer.
• No laboratory testing; no statement about safety is possible from this study.
• Illness reports are self-reported and cannot be linked to water.
• Conditions may change over the 13 weeks; every observation and reading carries a date.

Write 5,000 to 10,000 words in English, plus tables, charts and the appendix.

1. Title, authors, school or region, date
2. Abstract (150 to 200 words)
3. Introduction and background (3 to 4 verified sources)
4. Research question and hypothesis
5. Methods: study area, sampling, survey, observation sheet, interviews, photos, optional indicators, ethics
6. Results: sources; treatment; storage and handling; belief and its basis; belief versus practice; observed risks; indicators; interview themes
7. Discussion: what the gaps mean, surprises, comparison with the hypothesis, what a household could change
8. Limitations
9. Conclusion and the "Source to Glass" checklist
10. AI-use statement
11. Ethics statement
12. References
13. Appendix: survey questions, observation sheet, interview guide, photo log, indicator log

Cash budget: BDT 7,000 for the team. Drop the optional items if the team does not use them.

Estimated hours: about 60 hours per student across 13 weeks (planning 8, field work 27, analysis 12, writing 11, review 2). Reward: 2,500 raindrops for each student on the team.

• Our study area is one named ward, mahalla, union or para, and we wrote down why we chose it.
• Every team member has a parent's or guardian's agreement.
• Our AKTI Survey is built in Bangla with the 20 questions, tested on 3 to 5 people, and fixed.
• Our observation sheet is printed or set up as a form, and we did one practice visit.
• We have the consent script in Bangla and English and know the three separate permissions, plus the extra permission for photos inside homes.
• Location (geo-tagging) is on in every phone camera we will use.
• We have a field diary with date, time, place, what happened.
• We know who does what: survey lead, observation lead, interview lead, indicator lead (optional), writer.
• We have read the Ethics and safety section, especially "No verdicts."
• If we use indicators, we have the kit instructions, gloves, and an adult who will supervise any chemical test.
• We know who to tell the same day if a reading looks worrying: our coordinator.
• We have found what is publicly known about water sources and testing in our area and written down the source.
• We know our first-sample target for the advisor review: 10 to 20 responses, 3 observations, 1 to 2 interviews, 5 photos.

• Point of use: the place where water is actually drunk; quality can change between the source and here.
• Recontamination: clean water becoming dirty again through containers, hands or cups.
• Arsenic: a natural element in some groundwater; harmful over long exposure; invisible and tasteless.
• Microbial contamination: bacteria or other germs in water, often from human or animal waste; invisible.
• TDS: total dissolved solids, a number from a cheap meter; not a measure of safety.
• Turbidity: how cloudy water is; cloudy water is harder to treat.
• Field indicator: a quick measurement that hints at a condition; not a laboratory test.
• DPHE: Department of Public Health Engineering, the agency responsible for rural water supply and testing in Bangladesh [verify current name and role].
• WASA: Water Supply and Sewerage Authority, the city water utility in Dhaka, Chittagong and some other cities.
• Consistent treatment: treating drinking water every time with a method that is actually maintained.
• Volunteer bias: when the people who agree to be observed differ from those who do not.
• Self-reported: information a person gives about themselves, which may not match what they do.

Submit through your claim page on Rainflower, one stage at a time. Each stage opens when the previous one has passed. Your AI research assistant reviews every submission within a day and tells you what is met, what needs work, and what to do next. Stage 2 waits for your coordinator's checklist; stage 3 waits for your advisor's review; stage 7 goes to the coordinator's final checklist and then to Rainflower review and grading.

At stage 7 your evidence package manifest must list: the AKTI Survey export (CSV), every observation sheet, every interview file with its date, place and the three consent answers, every photo with its log line, the indicator log if used, the field diary, the analysis spreadsheet, and the final paper. Keep the originals; upload the manifest and the files Rainflower asks for.