The Ambiguous Hindi Item: 'काम' That Means Three Different Things
You are administering a household schedule in Hindi in a village of Bhagalpur district, Bihar. One printed item reads: "आपके परिवार में कितने सदस्य काम करते हैं?" Across the first 30 households, respondents interpreted it in three ways: some counted only paid work outside the home, some counted children's work at home and in the fields, and two counted elderly members who no longer work at all. When a re-check was done with a clearer wording, 14 of the 30 households gave a different number. A respondent, a retired schoolteacher, says the question is misleading and that a government form should state clearly whether it means paid employment or any work. Your supervisor asks you to keep the Hindi wording "as it is", because the schedule is printed and 96 forms remain.
1. Identify exactly what makes the Hindi item ambiguous, and rewrite it precisely with your reasoning.
Read the worked answer
The item fails on four counts:
- 'काम' is undefined. It can mean paid employment, wage labour, cultivation, or any activity including housework. Respondents legitimately read it three ways, which is why 14 of 30 answers changed on re-check.
- No reference period. Work 'ever', 'usually' or 'in the last week' gives three different numbers. A fixed recall period is essential.
- 'परिवार' is unclear. It may mean lineage or the group eating from one kitchen. The unit must be the household, not the family.
- Paid and unpaid work are merged. Children's unpaid work in the home and fields is work, but it is not employment and must be captured separately.
Rewritten as two items:
- पिछले 7 दिनों में आपके घर के कितने सदस्यों ने कोई ऐसा काम किया जिसका भुगतान नकद या अनाज के रूप में मिला? ______
- इनमें से कितनों ने केवल अपना घरेलू काम या अपने खेत का काम किया, बिना किसी मजदूरी के? ______
'घर' fixes the unit, 'पिछले 7 दिनों' fixes the period, 'भुगतान' separates paid from unpaid work, and the second item captures the unpaid work the first would otherwise hide.
2. What do you do about the 96 printed forms and the 30 responses already collected? Do you obey the supervisor?
Read the worked answer
Do not change the wording silently, and do not simply obey a 'keep it as it is' instruction that you know will produce unusable data.
The problem is already demonstrated: 14 of 30 answers changed when the question was clarified. That means the two wordings measure different things and cannot be pooled without a note.
Correct sequence:
- Stop and report. Tell the supervisor what the re-check showed, with the 14-of-30 figure - that number is the argument.
- Propose a standard probe. Keep the printed line but read aloud a fixed clarification to every respondent: paid or unpaid, last seven days, everyone who eats from this kitchen.
- Make it uniform. Apply the clarification to the remaining 96 and re-ask the first 30, with the date and the change written in the field diary.
- Document. If the supervisor still refuses, record the decision, the date and the observed effect on answers in the remarks, and submit the 30 flagged responses as a separate note.
Printing cost is a small loss; a variable that means three different things across 126 households is a much larger one.
3. How do you answer the retired schoolteacher's objection?
Read the worked answer
Acknowledge that he is right - do not defend the form because it is official.
Say plainly that the wording was ambiguous and that his point is being recorded: note his objection in the remarks with the date, the item number and the three interpretations observed. Tell him it will go to the office that drafted the schedule, since the same wording is likely in use elsewhere.
Then use him. He is a literate, respected villager; ask him to suggest the exact Hindi phrase that people in this village would understand without explanation, and test his phrasing on the next few households. A wording proposed by a local speaker and checked against real answers is better than one drafted at a desk.
Keep the register polite and non-defensive, and do not argue. A respondent who feels heard answers the remaining items - income, land, migration - far more candidly, and those answers are worth more than winning the point about a printed line.