STAT 1008 Assignment One
(a) The main purpose of statistics is that when the size of the population is not manageable, a sample of manageable size should be drawn from the population. While we are ultimately interested in some numerical aspect of the population, we also hope that some numerical aspect of the sample would reflect the numerical aspect of the population that we are interested in. Thus, there is no need to use census.
(b) Stratified sampling is recommended. Stratified sampling is commonly used in the scenario described in the assignment spec. It is a reasonable assumption that that the social research who undertakes this study is interested in drawing (statistical) inference about subgroup of the general population. Using stratified sampling with independent strata enables the researcher to collect such information, which would be otherwise lost in a more generalised sampling method. Given the availability of the distinct features of the population of interest, aggregating data across groups provides little benefit.
(c) Suppose that the social researcher decided to use gender as a distinct feature. Also suppose that the population X consists of m males and f females (where m+f = X), then the relative size m/X and f/X is used for proportionate allocation. Suppose Y samples must be selected for this study, and then mY/X and fY/X is to be selected within respect stratum. Random sampling can be used within each stratum.
(d) There should be 6 variables in her data set, namely, Household Residents, Full Time Workers, Number of Children, Proportion of Females, Household Income and Visits made to centre.
The features of each variable is summarized as follows:
Form 1-1 Features of the Studied Variables
Variable | Feature |
| Continuous | Count | Ordinal | Nominal |
Household Residents | N | Y | Y | N |
Full Time Workers | N | Y | Y | N |
Number of Children | N | Y | Y | N |
Proportion of Females | Y | N | Y | N |...