Tuesday, 25 December 2018
W1 - Variables & Sampling
Biostatistics
Step 1: Population
• The entire cohort of people in the target group (all Australians
under 20 years)
• Usually too large to study all the subjects due to various
restraints including time, funding, and resources
• Solution: Representative sample from the population
Step 2: Descriptive statistics
• exploratory data analysis
• Summarization of data with graphs, table and other numeric
measures
Step 3: Conclusions
• Conclusions are drawn by looking at how the sample differs
from the population
Step 4: Inference
• Inference = combination of results and the formation of the
descriptive statistics and probably to draw conclusions about
the population
What is the data?
Data = piece of information about individuals which are organized
into variables
• It is obtained by collecting information from a group of patients
participating in a study
Variable = any observation that can have different values but can
be predicted
Random variable = variable that arises due to the chance and can
not be predicted
Types of Variables
Categorical (qualitative) = non-numerical labels, where
observations fall into sperate distinct categories --> Blood Type
Ordinal = observations are ordered or ranked according to a
criteria --> classifying pain on a scale on 1-10
Nominal = observations are classified in separate categories
with no logical ranking --> Blood Types
*If there are only 2 categories the data is nominal
1
Tuesday, 25 December 2018
Numerical (Quantitative) = a measurable quantity, where
observations are counted or measured--> number of people
graduated in 2018
Discrete = value are counted clearly separated from each other or
can only be taken in whole numbers --> number of people in a car
crash
Continuous = measurements that can be counted by fraction or
decimal values --> Blood pressure
Population and Sampling
• Population = the largest collection of entities for which we have
an interest at a particular time
• Target population = population of interest (often used in
scientific literature)
• Sample = representative part of a population that represents
the population as much as possible
• Sampling = each observation has an equal chance to be
chosen for the sample
• The population must be representative of the population and an
adequate sample size is required for statistical analysis
2
Tuesday, 25 December 2018
Probability Sampling: techniques based on the theory of probability
usually produce "good" sample because they are representative of
the population
Sample random sampling (SRS) - used in simple experiments
that requires a single sample to be taken from a given
population or a representative sample frame
◦ Used when target group is sufficiently large
◦ Sample creates by selecting randomly from a sample frame or
population
Systematic sampling - used when its difficult to identify items
using a simple random sampling method (w/ random numbers)
◦ Every nth item is selected
Stratified sampling - used with a smaller sub-group and used
to reduce standard error
◦ Population divided into a set of smaller non-overlapping sub-
groups (strata) and then a simple random sample is selected
from each group
Cluster random sampling - used when studied population is
spread across a wide area such that a simple random sampling
would be difficult to implement to accessing the selected
sample
◦ Population is divided up into a set of different coherent areas
and areas/clusters randomly selected to access
Non-probability sampling: uses the most easily obtained
participants or case for the study and is non-random. Population is
not entirely known and common sense is used to choose sample.
Snow Ball Sampling - one participant is chosen and then further
participants are selected/acquired through word of mouth
◦ Used when you do not have sufficient people with
characteristics you are seeking
-
