Groundwater Sampling Design in a Semi-Arid Village
You are a Geography lecturer supervising 12 MSc students on a two-day field study in a semi-arid Rajasthan village of 180 households. The village has three open wells, one borewell, a pond, and household taps connected to a tanker supply. Local health workers report several fluorosis cases. Students have six water-testing kits (nitrate, fluoride, TDS); each kit performs 10 tests, so 60 tests total, but the budget covers only 40 tests. You must choose sampling sites to support inference about fluoride exposure. Students want to 'test everywhere'. One kit's calibration certificate is missing and its last reading was anomalous. You have two days, one vehicle, and permission to test household taps only with consent. A control village is 8 km away but adds travel time. You need to balance statistical representativeness, ethical consent, and data quality. The students must produce a field report with a map and a short policy note for the village panchayat.
1. Justify a sampling plan with numbers.
Read the worked answer
With 40 tests, do not spread them evenly across all sources. Fluoride exposure is most likely from drinking water, so prioritize household taps and the borewell, with the open wells and pond as comparison points. Proposed plan: 20 household taps selected by systematic random sampling from a numbered household list (every 9th household, with consent substitutions), 5 tests at the borewell (repeated on two days to check variability), 5 tests across the three open wells (2, 2, 1), 3 tests at the pond, and 7 tests in the control village's household taps. This yields 40. The household sample supports a village-level prevalence estimate; the borewell and wells allow source comparison; the control village provides a reference. The pond is not a drinking source but is relevant for livestock and bathing. Record GPS coordinates and consent. This is purposive-stratified, not perfectly random, because the goal is exposure assessment under field constraints.
2. How would you handle the questionable kit and missing data?
Read the worked answer
Do not use the uncalibrated kit for primary results. Assign it only to non-drinking comparison sites (e.g., pond) or use it for duplicate readings after borrowing a calibration standard from the control village's block office. If no standard is available, label its results as indicative and exclude them from the main fluoride exposure estimate. For missing data, pre-specify a rule: if a household refuses consent, move to the next household on the list and record the refusal; if a test fails, use the remaining kit capacity for a duplicate at the same site. Keep a field log of kit ID, time, and anomaly. In the report, report the number of valid tests per source and a sensitivity analysis excluding the questionable kit's readings. This protects data quality and transparency. Students should learn that missing data is not a failure if documented and handled transparently.
3. How would you present findings to the villagers without overclaiming?
Read the worked answer
Use a one-page Hindi handout and a village meeting. Start with what was measured: 40 tests from household taps, borewell, wells, pond, and a control village. Present fluoride results as ranges, not precise risk categories: e.g., 'X of 20 household taps had fluoride above the permissible limit; the borewell was lower/higher.' Do not say 'this water causes fluorosis' because the study is a small field sample and does not measure individual exposure or health outcomes. Instead say 'these results suggest that some sources should be tested further by the PHED laboratory.' Show the map with sample points and a simple table. Explain the control village comparison. Recommend that the panchayat request confirmatory testing and consider alternative sources for children. Avoid blaming any group. Invite questions and record concerns. The policy note should list three actions: confirmatory lab tests, source protection, and health screening referral.