2.1. The relationship between characteristic
and variable
Let’s suppose you have the question, “Are Wollongong
teenagers obese?” You need to define your population,
your characteristic, and your variables.
Let’s say that you decide that the population
for the study is residents from 13 to 19 years of age who live in
the Wollongong Local Government area and that the units
of the study will be individual Wollongong teenagers.
That was the easy part. How are you going to define a characteristic
such as ‘obesity’; i.e. how will you define what you
mean by ‘obese’? You soon realise that there are different
concepts of what it means for different people to be obese depending
on their sex, age, and height, as well as on other variables. You
need to use a variable to define
and measure the characteristic ‘obesity’.
What variable will you use to measure obesity?
You might decide to use Body Mass Index (BMI) as an indicator of
obesity. BMI is a constructed variable because
it cannot be observed directly but must be calculated from observable
variables, namely mass (weight) and height. Body Mass Index
can be calculated by dividing a person's weight (in kilograms) by
the square of their height (in metres). You can write this as the
following equation:
If you know
your weight in kilograms and your height in metres, you can
calculate your own BMI - even if you aren't a teenager!
For your study, you decide to use the standard provided by the
World Health Organisation (WHO) in which a BMI of less than 25 kg/m2
indicates normal weight, more than 25kg/m2 indicates
a person is overweight, and more than 30kg/m2 indicates
that a person is obese.
Let’s look at an example.
SCENARIO
Consider two teenagers, both 1.5 metres tall, one of
whom weighs 45kg and the other 67.5kg. The first teenager
has a BMI of 20kg/m2 and the second a BMI
of 30kg/m2. Thus, according to the WHO specifications
for this measure, the second teenager is obese.
Is BMI necessarily the best variable for obesity?
A good researcher is flexible enough to take on board new information.
You might realise as you conduct your research that although BMI
is a convenient and simple measure of obesity, it is by no means
the only or the best measure. Two people with the same BMI are not
necessarily equally obese. Gender and age might influence BMI but,
more importantly in terms of indicating obesity, so do skeletal
and musculature factors (A big-boned, muscular, very fit athlete
may have the same BMI as a fat person!). Other variables that can
be used include:
the Ponderal Index, which is calculated by dividing a person's
height by her/his weight cubed, (height/weight3);
