Sampling and households

sample size calculator

For example, if household 2 has 3 adults, interview the 2nd youngest shown in bold type. In this map, the starting point marked Start is just before the dwelling marked 1.

The interviewer finds at the randomly selected address, then follows a set of rules to work out which addresses will be chosen for interviews.

This is completely arbitrary; it's just a convention. For example, if the plan is for 12 households in a cluster, but one cluster only has 10, the closest unfinished cluster should have another 2 added.

simple random sampling

Turn left at every street corner. The birthday method Most market research books recommend asking for the person who last had a birthday or who next will have one.

Household survey project

This seldom happens, but it needs to be anticipated. What happens when a household is visited and nobody is home, or the occupants refuse to take part in the survey? Why every second household, and not every household? What is the problem with interviewing the first person the interviewer meets? This will produce a badly skewed sample, nullifying any care that has been taken in producing a representative sample of households. But the questionnaire was different from a normal one: instead of ticking boxes for "like it", "dislike it", and "not sure", the interviewer would write in the number of people giving each possible answer. The Kish Grid This is a table of numbers, named after the statistician who invented it. Interviewers always prefer to visit rich homes rather than poor ones, homes where somebody is there, and homes that are easy to reach. In theory, everybody in a household has an equal chance of being selected by this last birthday or next birthday method, but my research has found this does not produce the correct balance of sexes and age groups. Then multiple interviews can be made at larger households. A estimation problem which occurs when only one person in a household is interviewed: people in small households will be over-represented.

It's because the more time somebody spends at home, the more chance they have of being interviewed, with this method of choosing respondents. Mainly because neighbouring households tend to be more similar to each other.

Look up the column for the household number, and the row for the number of eligible people.

Introduction of household survey

Include people who sleep there, but are not there when you visit. The original registry spreadsheet listed everyone living in the camp. The actual names in the column entitled NAME have been removed to preserve confidentiality. It's because the more time somebody spends at home, the more chance they have of being interviewed, with this method of choosing respondents. If you have not, you have counted wrong or miscalculated the sampling interval. But the questionnaire was different from a normal one: instead of ticking boxes for "like it", "dislike it", and "not sure", the interviewer would write in the number of people giving each possible answer. People who spend a lot of time at home have different habits from people who are out a lot. But when a cluster includes more than about 50 dwellings, including skipped ones, it becomes too large specially in rural areas , and some of the cost savings disappear. Note that when you reach the end of the list, you should have selected your desired number of households. This is completely arbitrary; it's just a convention.

Make your first call at the next address to that: two to the right of it, when looking from across the street - or two higher in number, if there are a number of dwellings at one address.

Where everybody lives on a long road as they do in some parts of the world there are no street blocks: observing the above rules, the route will simply follow the road.

Rated 8/10 based on 85 review
Download
Sampling households and people