Research 8 - Sampling
Research 8 - Sampling
a) Population/ Universe
The entire set of objects and events which is the object of research, about
which the wants to determine some characteristics e.g. all college
students in Zimbabwe. Is the large pool of sampling units.
Target population is a specified large group of many cases from which a
researcher draws a sample and to which results from a sample are
generalized. Is a specific pool of cases the researcher wants to study.eg
students at Kushinga Polytechnic
b) SAMPLE FRAME
Is a list of all the elements of the defined target population from which
the sample is to be selected or drawn e.g. a list showing all college
students at KushingaPhikelela Poly, telephone directories, driver’s
licence records etc
c) SAMPLE
The sub—set of the whole population which is usually investigated by the
research and whose characteristics will be generalized to the entire
population. Is a small proportion of the population selected for
observation and analysis
*Despite the above weakness, non- probability sampling designs are useful
tools in the exploratory phase of research, a phase in which ideas and insights
are more important than scientific objectivity.
1. Convenience Sampling
- Also known as haphazard or accidental sampling
- This represents a sample drawn to suit the convenience of the researcher
e.g. it may be convenient to interview only employees within one
company instead of selecting employees from a number of companies or
those available during the study or readily accessible with minimum
effort and costs or friendly to the researcher
- The researcher may also selects anyone he or she happens to come across
- Saves costs and time
- Systematic errors are unpredictable and not measurable on the resultant
sample.
- The sample may not be a representative of a the defined population
2. Judgment Sampling
- The sample consists of sampling units deliberately selected from the
population on the basis of the experience, judgment and expertise used
by the researcher to select the best sampling units to include in the sample
e.g. a sample of four of the most influential economists is asked to
estimate the next year’s rate of inflation
- Researcher uses his or her judgment to select what he or she thinks to be
the best satisfactorily
- This type of sample may be subject to error which although relatively,
can still not be measured or predicted.
- Pilot studies are based on such samples
3. Quota Sampling
- Is a non random sampling in which the researcher first identifies general
categories into which cases or people will be fall and then selects cases
from each category to reach a predetermined number of cases in each
category.
- The population is divided into a number of segments and the researcher
arbitrarily select a quota e.g. a certain percentage of sampling units of
each segments eg under 10, 20 etc
- Chosen in a way that the proportion of sample elements possessing a
certain characteristics is approximate the same as the proportion of the
elements with the characteristics of the population
- Elements are assigned a quota that specifies the characteristics of
elements to be contacted believed by the researcher to be more realistic
4 Snowball sampling
Used where respondents are difficult to identify and are based on referrals
network.
An individual is initially discovered mostly using a non-probability
method, although probability can be used, and is used to locate others
who posses similar characteristics through referrals, who in turn identifies
others until the desired sample size is attained.
This method is used on issues which individuals are unwilling to come up
openly; hence incentives can be used to encourage elements to participate
Advantages
-Only cases considered judgmentally make the sample quite
representative
-Unnecessary elements are dropped out
-Purpose of the study can be fulfilled even if the sample is small
-The investigator clearly knows the objective of the research hence only
necessary elements are included
Demerits
-Selection can be biased and the resultant sample not representative of the
population
-Method cannot be used by an average investigator to select sample units
-Difficult to detect mistakes and inaccuracy till end
6) Deviant case sampling also called extreme case sampling. Is a
nonrandom sampling in which a researcher selects unusual
nonconforming cases purposely as a way to provide greater insight into
processes or a setting. Used to seek cases that differ from predominant
characteristics of other cases
1 2 3 45 = 1 4
2. Systematic Sampling
- Follows a systematic or uniformity process when selecting sample
element. Sampling begins by randomly selecting the first sampling
units, thereafter subsequent sample units are selected at a uniform
interval, relative to the first sampling unit.
3. Stratified Sampling
- Use when the population is believed to or regarded as heterogeneous
with respect to the various variables or characteristics understudy
- The population is divided into mutual exclusive segments or strata where
the sampling units in each strata are relatively homogenous
- Thereafter random sampling can be done from each stratum. The
researcher needs to estimate the average mass of a large group of people.
First divides the group into strata e.g male and female and then selects a
sample from each stratum.
4. Cluster Sampling
- The population is divided into clusters where each cluster is similar in
profile to every other cluster, which is, containing the required number of
elements and characteristics of the population. A cluster is a unit that
contains final sampling elements but can be treated temporarily as a
sampling element.
- Used when researchers lack good sampling technique for a dispersed
population and the cost to reach a sampled element is very high
- A cluster is randomly selected from other clusters for investigation. The
sampling units within this randomly selected cluster maybe then be
randomly selected to provide a more representative sample from the
population.
1. JUDGEMENT/SUBJECTIVE APPROACH
-The judgmental approach involves the researcher using informed intention
to
determine the most adequate sample size. Determination of sample size is
arbitrary, subjective and doesn’t consider the likely precision of the survey
results
or the cost of obtaining.
-Is a qualitative approach or subjective method depending on individual’s
feelings
-Is a convenient sample size determination technique, which is based on the
researcher’ judgment
-Is a cheaper and faster to obtain the required sample size and information
required
although the sample might not be a representative of the population
2. MATHEMATICAL APPROACH
- This is a quantitative approach based on statistical precision which relies on
the
traditional statistical formulae for determining sample size.
-Is objective, reliable, transmittable and measurable
-The sample size depends on 3 factors:
a) The level of confidence desired, obtained from the values of Z
scores
b) Maximum allowable error
c) The standard deviation of the population.
4. COST LILMITATION
- This cost limitation methods determine sample size on the basis of budget
allocated to
the research project. This approach involves deducting from the available
budget all
non-sampling related cost such as fixed cost of designing the survey
questionnaires
preparation, data analysis and report generation to arrive at the desired sample
size.
This approach is unsatisfactory because its emphases on cost to the exclusion
of all
other factors especially precision. Time may also be considered.
5. INDUSTRY STANDARDS
- It refers to those rules of thump developed from experience that have
become
standard industry guidelines for determining sample size. Convectional
guidelines on
sample size vary with the type of research study.
1 OMISSION DISCREPANCY
Result when the defined population is extrme larger than the sampling
frame. Some objects in the population have no chance of being in the
sample because do not appear in the sampling frame although having
the required characteristics. This result in a less representative sample
Sampling frame Population
2 COMMISSION DISCREPANCY
Occurs when a sampling frame is larger than the defined population.
Results in objects being included more that once and also including of
unnecessary objects. Arises due to lack or incomplete information
about potential study objects. This can be reduced by screening before
data collection by using qualifying questions.
3 COMMISSION-OMISSION DISCREPANCY
Is a combination of omission and commission discrepancies. Caused
by, not including all objects in the population in the sampling frame
although these essentially containing the required characteristics or
not all objects in a sampling frame are in the population. Other
elements in the population are excluded while other objects in the
sampling frame are not from the population