Plot or Dwelling Interviews & Sample Size


Interview
: a meeting at which information is obtained (as by a reporter, television commentator or pollster) from a person or persons. (Source: Merriam Webster’s Collegiate Dictionary).


Sample size: the number of sample units to be included in the sample.


MajiData dwelling & plot interviews exercise: Similar to counting, all plots or flats or block of flats in the target area were initially the sampling frame for conducting interviews. Subsequently a dwelling or household within the sampled plot or flat or block of flat was visited as the interviewing unit. With the later modification of the interviewing design, all dwellings within the target area evolved to be the main sampling.


Justification of Sample Sizes Used:

The underlying principle during implementation of plot and/or dwelling interview was to adapt the size of the sample to the estimated size of the population. Ideally, the plot sample should accommodate between approximately 1.2 and 19% of all plots or dwellings.

In order to allow for reliable estimates regarding the average household size and the average number of people per plot or dwelling, the minimum required sample sizes used are depicted in the table below. Actual sample sizes were often higher than the minimum allowed, thus enhancing the reliability of the estimates.


Table: Population, number of plots and sample sizes

Number of plots or dwellings

(no. based on count)

Estimated population (*)

Required sample size

Percentage of plots interviewed (%)

80

500

15

19

170

1,000

20

12

500

3,000

35

7

830

5,000

50

6

1,700

10,000

65

4

2,500

15,000

75

3

3,400

20,000

80

2

5,000

30,000

90

1.8

6,700

40,000

120

1.8

8,400

50,000

150

1.8

16,700

100,000

200

1.2

25,000

150,000

300

1.2

(*): Assuming a plot or dwelling has an average population of 6 persons


Once the sample size was correctly determined, the actual plots or dwellings to be interviewed were identified. Here the goal was to ensure even distribution of the sample, in other words, all parts of the area have to be included (including the more marginal or inaccessible zones). For large areas, or areas with a large population, the routine was to divide the area up in zones (preferably making use of the streets/roads and features such as buildings, electricity transmission lines for delineation of zones). The total sample size would then be divided over the various zones taking into account the sizes and population density of each zone.

Randomization of the selected plots/dwellings was a key element. This means that in principle every plot /dwelling in the target area had the same chance to be interviewed. Realising such goal of creating a random sample was practically possible on the basis of an even geographical spread of the selected plots/dwellings. This explains the reason for assigning one or more streets, pathways or alleys to each of the Data Collectors, followed by random visitation of the plots/dwellings. Normally, while standing at the beginning of the street, pathway or alley, it would follow that a particular data collector would visit the 3rd and 7th plot on the left hand side of the road and the 4th plot on the right hand side of the road.