Notesale: Turn your study into money

Document Preview

Extracts from the notes are below, to see the PDF you'll receive please use the links above

---

THE RESEARCH SAMPLE
Sampling – the part of research to do with selecting
those persons or items that will be studied from the
larger population
...
Typically
too little attention is paid to this aspect and errors in the way a research sample
is chosen and accessed can have profound effects on the interpretations and
conclusions the researcher can make from the data collected
...
g
...

Evidence-based practice is a meaningless concept if the evidence is built on
unstable rocky foundations! Errors in sampling have the potential to invalidate
research findings and render the research useless and, if involving human
subjects, unethical
...

Although radiographers may collect data from many situations e
...
documents,
hospital records, images , isodose distributions (treatment plans), cassettes, all
of which can constitute the research sample, it is easier to understand the
concept of sampling if we view humans as our study units
...
e
...

e
...

all diagnostic radiographers working in a senior one grading in the UK
all therapeutic radiographers who work permanently in the pre-treatment
sections in England
...

Hopefully you can see that the target population may be:

Broadly or narrowly defined
Defined in terms of actions, characteristics, organisations or groups etc
...
However, hopefully you can also understand
that researchers rarely have access to this entire population so they have to give
some thought to who (what ) is the accessible population
...

Thus the researcher undertake various processes to select a research sample,
i
...
a sub group of the total target population
...

A representative sample has all the important characteristics of the
population from which it is drawn
...
e
...


Consider the example below
...
It is not truly representative
...
Often though we
cannot be completely certain, and there may be unknown factors that would
affect measurements that could be linked to body colour
...
When using qualitative approaches
however, the representativeness of the sample is may not be a primary concern
– researchers may try to select study units that give them the richest possible
data
...

Using the above example if we are doing qualitative study on what it is like to
have a green body we would be happy with the sample because the subjects
could be expected to be able to comment with authority on this issue
...
e we select them
on purpose because they have the necessary experience or characteristics
...
So when you are critically evaluating the work of
others this is one aspect which should receive careful scrutiny
...
We’ve shown that qualitative studies are not high
on external validity - this is not their aim
...

Reducing the opportunities for bias to impact on the findings is vital whichever
approach is used
...
It is usual for the researcher also to
define the exclusion criteria for his/ her study
E
...
using one example from above
Inclusion criteria:
Women diagnosed with T2 N0 M0 carcinoma of right breast
Women aged 50-70 years
Women receiving confirmed diagnosis between 1st September 2003 until
1st September 2004
Women whose diagnosis was established in the United Kingdom only
Exclusion criteria
Women with T2 N0 M0 carcinoma of left breast
Women with T2 carcinoma of right breast with evidence of regional nodal
involvement or suspected distant metastases
Women with right breast cancer staged T1, T3 or T4
Women aged 49 and below and 71
Women whose diagnosis was established outside UK

PROBABILITY SAMPLING
This involves some form of random selection procedures i
...
the researcher
should ensure that each study unit in the population has an equal and
independent chance of being selected – each unit of the sample is chosen on the
basis of chance
...

For example a researcher wants to compare two different imaging modalities for
knee injuries:
If the eligible study units are randomly allocated to the different modalities then
there is more chance that any uncontrollable variables, which may confound the
findings, are represented in both groups and are therefore less likely to bias the
outcome
...

Probability sampling requires that a listing of all study units exists or can be
generated
...

Examples of probability sampling methods include:
Random sampling is the purest form of probability sampling
...
May be selected using random number generator or tables
...
It is also
called an Nth name selection technique
...
As long as the list does not contain any hidden order, this
sampling method is as good as the random sampling method
...
Systematic
sampling is frequently used to select a specified number of records from a
computer file
...
A stratum is a
subset of the population that share at least one common characteristic
...
The researcher first identifies the relevant strata and their
actual representation in the population
...
"Sufficient" refers
to a sample size large enough for us to be reasonably confident that the
stratum represents the population
...

Cluster Sampling
o In cluster sampling the units sampled are chosen in clusters, close
to each other
...
The population is divided
into clusters, and some of these are then chosen at random
...

Ideally the clusters chosen should be dissimilar so that the sample
is as representative of the population as possible
...

These strategies are less likely to produce a representative sample, but despite
this are the strategies more likely to be used in social science research
...

Examples of non-probability strategies include:
Convenience samples (opportunistic)
o As the name implies, the sample is selected because they are
convenient
...
However, this convenience may
contribute to a particular bias in the sample
...
Handy but not
unbiased
...
This is
usually an extension of convenience sampling
...

When using this method, the researcher must be confident that the
chosen sample is sufficiently representative of the entire
population
...

Like stratified sampling, the researcher first identifies the strata and their
proportions as they are represented in the population
...
This differs from stratified sampling, where the strata
are filled by random sampling
...
It may be extremely difficult or cost
prohibitive to locate respondents in these situations
...

While this technique can dramatically lower search costs, it comes at the
expense of introducing bias because the technique itself reduces the
likelihood that the sample will represent a good cross section from the
population
...


BIAS
Bias in sampling is a systematic error that can lead to distortion of the results
...

Other sources of sample bias include:
self-selecting samples:
volunteers – they may be motivated, have a particular “axe to grind”
...

Qualitative approaches:
There are no fixed rules for determining the sample size in qualitative
approaches - this is driven by what the researcher is trying to find out, the
methods used to generate the data, and who the sample are
...
In exploratory studies the sample
size is usually estimated before commencing but not determined
...

This calculation will depend on whether the researcher seeks to:
measure one single variable in one group to an identified level of precision
demonstrate a significant difference between two groups
As a general rule of thumb the larger the sample size the more likely it is
to be representative of the target population
The larger the sample size the smaller the sampling errors – important for
some statistical testing
...

Attrition rates of 20% or more are not uncommon and therefore the researcher
needs to consider possible attrition when determining sample size
Number of variables:
The greater the number of independent variables and possible extraneous
variables the greater the number of subjects that will be needed

Sensitivity of the measuring device
If the researcher is using a published, recognised measuring scale which:
has been well tested previously with samples or contexts similar to
his/her intended study
has statistical reliability and validity data available
he/she could argue that this device is sensitive with respect to variables of
interest and therefore smaller sample sizes may be adequate
...

In this way we can be more confident that whatever we discover to be true for the
sample is also likely to be true for the population, and therefore our research
findings have applicability
...




---