Chapter 3
Some Basic Statistical Measures
3.1 Purpose
The purpose of this chapter is to introduce some basic statistical measures that are
commonly used in demographic analysis. The concepts are defined in general terms
without going into theoretical details. Methods of calculation of various measures
are described. The statistical measures discussed in this chapter consist of counts,
frequencies, proportions, rates, various measures of central tendency, dispersion,
comparison, correlation and regression.
3.2 Demographic Data and Analysis
Demographic data can be classified according to their level of measurement. This is
useful because the level of measurement helps the selection of what statistical
analysis is most appropriate. Several classifications can be used. This book uses
the four-fold classification system proposed by Stevens (1946) that classifies data as
being (1) nominal, (2) ordinal, (3) interval and (4) ratio. There are other systems
such as the two-fold classification of (1) discrete and (2) continuous.
In the classification system proposed by Stevens, the nominal level is known as
the lowest level of measurement. Here, the values just name the attribute uniquely
and do not imply an ordering of cases. For example, the variable marital is
inherently nominal. In a study it might be useful to have attributes such as never
married, married, separated, divorced and widowed. These attributes are mutually
exclusive and exhaustive. They could be coded N, M, S, D and W respectively, or
coded as 1, 2, 3, 4 and 5. In the latter, the numbers are not numbers in a real sense
since they cannot be added or subtracted. Thus, numbers assigned to serve as values
for nominal level variables such as marital status cannot be added, subtracted,
multiplied or divided in a meaningful way. An exception is the dummy coding of
F. Yusuf et al., Methods of Demographic Analysis, DOI 10.1007/978-94-007-6784-3_3,
© Springer Science+Business Media Dordrecht 2014
21