SWP2RES RESEARCH FOR SOCIAL WORK PRACTICE A

1
SWP2RES RESEARCH FOR SOCIAL WORK PRACTICE A

• MARTIN RYAN
• LECTURE NINE
• Selecting a Sample

2
The next key point to be addressed in
undertaking a piece of research is

• Who is that information to be collected from?
• What particular part of the general population is
  it going to come from?
• How is that particular part of the population to
  be selected and in what sort of numbers?
• These are the essential questions to be answered
  in sampling.

3

• Researchers tend to select a portion of that
  population for study, that is a sample, and
  either
• 1) study and describe that sample in some depth
  or
• 2) attempt to generalise from these findings from
  the sample to the population (the target
  population)

4

• The manner in which the sample is drawn
  determines to what extent we can generalise from
  the findings (Bouma Ling, 2004, p. 113)

5

• To choose a sample, it must be decided who or
  what is to be observed. This is called
  identifying the units of analysis.
• These could be, for example, newspaper articles,
  television programs, schools, groups or
  individual people.

6

• For sampling purposes, a population is the entire
  collection of people or elements that share some
  defined characteristics.
• For example
• - all residents of Victoria,
• - all social workers working in community
  health centres for less than two years
• - elderly parents (more than 65 years of age)
  of intellectually disabled adults living in the
  city of Melbourne.
• - all 2nd year Social Work students a La Trobe
  University

7

• A sample is a subset of individuals selected for
  study from persons or objects within a defined
  population.
• A case or element is the basic unit of analysis
  in a research study such as each actual person in
  a sample of individuals. Collectively, the
  persons selected for the study constitute the
  sample. Individually, each person is a case or
  element of the sample e.g. a Victorian resident,
  a La Trobe 2nd year Social Work student etc.

8

• A sampling frame is a list of all the relevant
  units from which the sample is selected e.g. the
  electoral roll or telephone book.

9

• Researchers may want to be able to generalise a
  studys findings to a target population.
• Hence the question to ask about a sample is can
  the findings from a sample be generalised to the
  population from which the sample is drawn?

10

• The answer is yes provided the sample is
  representative.
• A representative sample is one that is very
  similar to the population from which it is drawn.
  The distribution of relevant characteristics in
  the sample should be the same as those in the
  population (See Representative
  Unrepresentative Samples figure). This is
  usually ensured by random sampling whereby a
  sample is drawn from a population so that every
  member of the population has an equal chance of
  being selected in the sample.

11
TYPES of SAMPLES

• There are two general types of samples
• 1) PROBABILITY and
• 2) NON-PROBABILITY

12

• Probability sampling relies on random, or chance,
  selection procedures in which cases are selected
  into the sample only on the basis of chance.
• Random does not mean haphazard it means that
  every element in the population has the same
  known probability of being selected for inclusion
  in the sample.

13
PROBABILITY SAMPLING Types

• 1) SIMPLE RANDOM SAMPLING
• 2) SYSTEMATIC RANDOM SAMPLING
• 3) STRATIFIED RANDOM SAMPLING
• PROPORTIONAL
• NON-PROPORTIONAL OR DISPROPORTIONATE
• 4) MULTISTAGE CLUSTER SAMPLING

14
PROBABILITY SAMPLING

• Simple random sampling is a sample which involves
  randomly selecting some predetermined number of
  persons from the sampling frame.

15
5 steps in selecting a simple random sample

• 1. Obtain a complete sampling frame (list of
  elements)
• 2. Give each case a unique number starting at
  one
• 3. Decide on the required sample size
• 4. Select that many numbers by whatever method
• 5. Select the cases that correspond to the
  randomly chosen numbers.

16
PROBABILITY SAMPLING

• 2. Systematic random sampling is a simpler,
  less costly alternative to simple random
  sampling. It involves no random generation of
  numbers

17

• It involves the researcher determining the number
  in the sampling frame and the sample size, using
  the sampling fraction method (k), where the
  formula is K N/n where N the population and n
  sample..
• For example, if the sampling frame was 200 and
  the sample size 50, the required sampling
  fraction 200 divided by 50 4. Therefore every
  fourth case in the sampling frame is selected.

18

• 3. Stratified random sampling improves on a
  simple random sample and a systematic random
  sample by ensuring representation of
  particular subgroups in the population e.g.
  male students undertaking a social work course.

19

• The target population is divided into a number of
  strata and a sample drawn from each stratum.
  Existing knowledge of the population is used to
  increase the representativeness of the sample by
  dividing it into stratifying variables (a
  characteristic on which we want to ensure correct
  representation in the sample.)

20
Stratified random sampling can be further divided
into

• a) proportional stratified sampling and
• b) non-proportional or disproportionate
  stratified sampling.
• See figure on Stratified Random Sampling

21

• 4. Another form of probability sampling is
  multistage cluster sampling. This occurs when
  a comprehensive sampling frame does not
  exist and it is not possible to construct one
  e.g. where large populations are spread across
  wide geographic areas or among many different
  organisations.

22
Cluster random sampling takes place when the
population is divided into clusters or units,
then successively smaller subunits. At each
level, units and subunits are randomly
selected. So cases or a random sample of all
elements within each of the selected clusters
is studied. Cluster samples can involve
multiple stages. See Cluster Sampling figure.
23

• 2) Nonprobability sampling is done when random
  sampling would be inappropriate such as such as
  in an exploratory study when in-depth
  information is required or when a population
  has not fully described or identified or in
  qualitative studies e.g. sub-teenage homeless
  heroin addicts.

24

• Nonprobability samples are those that do not use
  random sampling and where some elements of the
  population have a greater or lesser chance of
  being selected.

25
NON PROBABILITY SAMPLING

• 1) ACCIDENTAL (or incidental or convenience or
  availability sampling )
• 2) QUOTA
• 3) PURPOSIVE

26

• 4) SNOWBALL
• 5) THEORETICAL

27
How large should my sample be? (sample size)

• Major rule is
• The greater the number the better for
  generalisation

28
Sample Size

• Depends on
• 1) resources available for getting the sample
• 2) the nature of the study and
• 3) the type of statistical analysis to be used
  in the study.

29

• Next Week no lecture (mid-semester break)
• Lecture on return (week after) will be on data
  collection methods.