Evaluating the Limitations of Non-Probability Sampling

In this assignment, the importance of sampling methods is discussed, highlighting the advantages and disadvantages of different approaches such as Probability Sampling, Accidental Sampling, Quota Sampling, and Judgmental Sampling. The limitations of non-probability research are also mentioned, concluding that Probability Sampling is more conclusive and comprehensive. The text recommends Stratified Random Sampling as a preferred method due to its high representativeness and comprehensiveness. The discussion concludes with a review of references supporting the arguments made.

QUANTITATIVE TECHNIQUES IN BUSINESS

1
Introduction
In the following paper, we will be discussing the sampling techniques, i.e., probability sampling and
non-probability sampling techniques that are being used. First of all, sampling means choosing a
specific collection from the entire population and it is basically divided into two groupings, i.e.,
probability sampling and non-probability sampling. They both are similar to each other in a way on
the other hand in actual they are different to each other as in the probability sampling method each,
and every individual gets a fair-minded chance of assortment which is not being possible in the case
of non-probability sampling. Moreover, probability sampling is an inspecting procedure, in which the
themes of the populace get an equivalent chance to be chosen as a representative test. (Tourangeau,
R., 2013)
Nonprobability testing is a technique for inspecting where it isn't realized what individual from the
populace will be elected as an example. On the other hand, probability sampling is random sampling
whereas non-probability sampling is non-unsystematic. Probability sampling research is conclusive,
and non-probability research is Exploratory. (Baker, R., 2013)
Elements of Population and the importance of sample size
There are three elements of sample size, and it is being measured in the concern that will ultimately
help in the decision making.
The three elements are explained below:
The Risk
The risk is one of the elements that is being used and helps in the decision making, and risk is of two
types, i.e., Risk of the unknown and the statistical risk. The risk of the unknown is a type of
experiment that is being used in the decision making deprived of any data. The experiment will
expose an essential data or sustain a philosophy. This danger of the obscure likewise incorporates
obscure blunders in the specimens or test process and testing itself may deliver fragmented (frequently
concealed) consequences.

2
Statistical risk includes the sampling estimate specifically. There can be a measurable possibility that
the haphazardly chose set of a unit for the specimen is by and large superior to the populace or the
other way around. The reason for measurable testing and utilizing an arbitrary samples the specimen
will speak to the populace.
We utilize the term measurable certainty that will be used in depicting the capacity to the sampling for
representing the population appropriately.
The variance
Variance is the accurate measure of inconstancy in a population of information.
It is the thing that it is, and the best way to diminish inconstancy is the change the item plan or
creation process. An extensive difference will be required for an advanced sample scale to identify a
move in outcomes of an analysis.
The precision
The precision is related to the thing that we will be going to identify. The bigger the distinction that
we want to distinguish the fewer specimens we will require. This specimen estimate changes as it is
harder to definitively recognize a little change.
We may measure one specimen from the old and new plans and see a distinction, perhaps a little
contrast.
Importance of sample size
 Confidence and margin of error: The level of our sample directs the measure of data we
have and along these lines, partially, decides our exactness or level of certainty that we have in
our sample measures. A scale dependably has a related level of vulnerability, which relies on
the fundamental fluctuation of the information and additionally the specimen measure. The
more factor the populace, the more prominent the vulnerability in our gauge. Thus, the bigger
the sample measure, the more data we have thus our vulnerability diminishes.

3
As our sample measure constructs, the confidence in measure use to expands, our vulnerability
reductions and we have more prominent accuracy. This is unmistakably shown by the
narrowing of the certainty interims in the figure above. On the off chance that we took this as
far as possible and tested our entire populace of intrigue then we would get the genuine esteem
that we are attempting to appraise – the real extent of grown-ups who claim a cell phone in the
UK, and we would have no vulnerability in our gauge. (Fassnacht, F.E., 2014)
Power and Effect size: as through increasing the size of the sample, it provides us the better
power in discovering the alterations. The statistical test that can be used in the investigation is
a binominal assessment of equivalent quantity or can also be known as two proportion Z-test.
There is deficient confirmation to build up a contrast amongst men and ladies, and the

4
outcome isn't considered measurably remarkable. The likelihood of watching a sexual
orientation impact of at least 18% if there were genuinely no distinction amongst males and
females is more noteworthy than 5%, i.e., generally likely thus the information gives no
genuine confirmation to recommend that the genuine extents of males and females with cell
phones are extraordinary. This cut-off of 5% is normally utilized and is known as the
"centrality level" of the test. It is picked ahead of time of playing out a test and is the
likelihood of a sort I mistake, i.e., of finding a measurably huge outcome, given that there is in
certainty no distinction in the populace. The Binomial test is basically taking a glimpse at how
much these sets of interims cover, and if the cover is sufficiently little, then we infer that there
truly is a distinction. (Fassnacht, F.E., 2014)
Probability sampling and non-probability sampling
Probability sampling is an inspecting procedure, in which the themes of the populace get an
equivalent chance to be chosen as a representative test.
Nonprobability testing is a technique for inspecting where it isn't realized what individual from the
populace will be elected as an example. On the other hand, probability sampling is random sampling
whereas non-probability sampling is non-unsystematic. Probability sampling research is conclusive,
and non-probability research is Exploratory.
Some of the advantages and disadvantages of probability sampling are:
Advantages of Probability sampling:
 Cluster testing: accommodation and usability.
 Simple random testing: makes tests that are exceptionally illustrative of the populace.
 Stratified random inspecting: varieties of coatings which are very illustrative of divisions or
coatings in the populace.
 Systematic sampling makes tests that are exceptionally illustrative of the populace, without
the requirement for an irregular number generator. (Rada, V.D.,2014)

5
Disadvantages of Probability sampling
 Cluster testing: won't function admirably if unit individuals are not homogeneous (i.e., on the
off chance that they are not the same as each other).
 Direct random testing: monotonous and tedious, particularly while making bigger examples.
 Stratified random examining: repetitive and tedious, particularly while making bigger
examples.
 Systematic testing: not as arbitrary as basic irregular inspecting.
Advantages and disadvantages of Non-Probability sampling
A significant advantageous position with a non-sampling method is that — as contrasted with
probability sampling — it's exceptionally cost-and time-viable. It's additionally simple to utilize and
can likewise be utilized when it's difficult to lead probability testing (e.g., when you have a little
populace to work with).
One of the remarkable problems of non-probability testing is that it's difficult to know how well you
are representing to the general population. In addition, you can't ascertain some of the certain intervals
and opportunity for the intervals of self-assurance and margin of inaccuracies.
Three sampling techniques for each the probability sampling and non-
probability sampling
Three sampling techniques of probability sampling are explained as follows:
1. Simple Random Sampling (SRS)
In this method of sampling, every individual has an equivalent chance of being chosen in an
entire population as the selection of an individual in this method is free of any biasness and
this can be calculated by the formula : n/N where n = sample size and N = population size.
Moreover, it can be done with or without replacement.as with replacement means the
possibility of selecting an item two times in a sample and without replacement means that the

6
possibility of selecting an item should be more convenient and should provide the more
detailed outcome. Some ways of selecting a simple random sampling can be done in several
ways like drawing a lottery, table of random numbers in which each item is numbered and is
selected randomly from the table, or a tippet’s numbers methods or using the grid system. (Al
Ghayab, 2016)
2. Stratified Random sampling
This sort of probability investigative strategy is a standout amongst the most usually utilized
techniques and includes the division of the entire populace into various strata.
These strata are particularly used to select and furthermore exceptionally comprehensive in
nature. From each of these strata, a straightforward irregular example is drawn, by this; the
quantity of the specimens drawn from each of the examples ends up noticeably relative to their
individual strata estimate. At the point when the populace is heterogeneous in nature, this kind
of inspecting assumes an exceptionally basic part. The stratification in this strategy is
performed in such a rich route, to the point that the difference between the strata is high and
information inside every stratum is little. (Singh, R., 2013)
Categories of Stratified Random Sampling –
Disproportionate Stratified Random Sampling – In this kind of sampling, measure up to a
number of the units are drawn from every stratum, not relying upon the span of the strata.
Proportionate Stratified Random Sampling – in this kind of examining, the quantity of
units in every stratum is proportionate to its number in the world.
3. Cluster Sampling
This sort of technique is extremely valuable in bringing down the recorded cost as this strategy
is an exceptionally functional and effectively musical show table technique. Here the populace
is separated into various gatherings and these gatherings are alluded to as clusters and further
these groups are alluded to as the essential testing units. This technique includes all else the ID
of the bunch, as indicated by which groups might be units like areas, talukas, city pieces,

7
schools and so forth. Groups ought to be homogeneous in the inside characteristics. Then
assist in this strategy, assurance of the quantity of the stages is to be finished. It might also
include single, double phase or even a multiple phase sampling.
Three non-probability sampling techniques are explained below:
1. Accidental Sampling
In this type of sampling, the researcher just connects and grabs the cases that flip in front of
the researcher, proceeding with the procedure till such time as the sample procures the desired
size. For instance, may take the initial 150 people he meets on any of the people on foot ways
of a road, who will be met or to give the sort of data he is seeking. In such a specimen, there
is, obviously, no other method for assessing the inclination ( a distinction between the normal
example esteem and the genuine populace esteem) aside from by doing a parallel report with a
probability test or by embraced a total enumeration. In any of the case that does not mean, that
unplanned samples don't have wherever in logical research. This sort of inspecting, other than
being practical and advantageous, can likewise bear the cost of a reason for incitement of bits
of knowledge and working speculations. (Albuquerque, U.P., 2014)
2. Quota Sampling
A major goal of quota sampling is the assortment from the sample that is a duplication of the
‘population' with the one that desires to take a broad view. Here, the researcher makes sure
that the equivalent or proportional issue depiction is being taken that depends on which the
behaviors are being considered on the foundation of the quota.
For instance, the basis of the quota is from the school level, and the expert wants the
representation to be equal, taking a sample size of 150, he should choose 50 form the 11th
class, another 50 from the 12th class, 50 from the 10th class. The bases of the standard taken are
generally on the basis of age, sexual orientation, instruction, race, religion and financial status
and so on. (Acharya, A.S., 2013)

8
3. Judgmental Sampling
Judgmental sampling is also known as purposive sampling. In this sort of examining, subjects
have been a piece of the sample in light of a particular reason. With judgmental testing, the
scientist trusts that a few subjects are fit for the examination contrasted with different people.
Judgment is extremely complicated, on the grounds that considerably more grounded
suppositions must be made about the public and the investigative technique than are required
while utilizing likelihood testing. Besides, reviewing the mistakes and leanings can't be
figured for this kind of tests since examining technique does not include likelihood examining
at any stage. (Heyer, R., 2014)
Critical evaluation of the three sampling techniques for probability
sampling and non-probability sampling
The Simple Random Sampling is free of any biasness, and the individuals are selected on the
random basis, which is one of the main disadvantages of this method as in some of the cases, the
random sampling is not suitable.
The Stratified Random sampling cannot be used in every kind of study. Analysts must recognize
each individual from a populace being considered and order each of them into one, and just a single,
subpopulation. Finding a thorough and complete rundown of a whole populace is the main test.
Sometimes, it is out and out unthinkable. The other drawback of this method is specifically arranging
every individual from the population into an introverted division.
In Cluster Sampling as each cluster is collected from a unit which is like each other as this may
deliver huge examining mistake and lessen the representativeness of the example.
At the point when an unequal size of a portion of the subsets is chosen, a component of test
inclination will emerge. This sort of inspecting may not be conceivable to apply its detections to
another region also. (Acharya, A.S, 2013)

9
In Accidental Sampling, the problems can occur like a possibility of being prejudiced or unfair as it
shows the view of some particular group but not of the entire population. There can be a high
possibility of the sampling error, and even the results that come out are even not comprehensive.
In Quota Sampling, there is a difficulty related to non-approachable prejudices. For instance In the
event that some individual declines to be a piece of an examination, at that point amount testing
enables the questioner to go and locate the following individual who is ready, which brings about
information that isn't completely illustrative of the populace. The purpose behind this is, non-
respondents most likely have certain qualities, and in light of the fact that the information acquired
from the example won't speak to them by any stretch of the imagination (it may be illustrative of
respondents).
Judgmental Sampling is an irrational sampling of the large population is being selected in this
sampling, and moreover, there is no logic in the selection of its sample size. (Himelein, K., 2016)
Conclusion and recommendation
It has been concluded that Probability sampling research is conclusive and non-probability research is
Exploratory. The Binomial test is basically taking a glimpse at how much these sets of interims cover,
and if the cover is sufficiently little, then we infer that there truly is a distinction. It has also analyzed
that In Accidental Sampling, the problems can occur like a possibility of being prejudiced or unfair
as it shows the view of some particular group but not of the entire population. There can be a high
possibility of the sampling error, and even the results that come out are even not comprehensive
which should not be used as it gives insignificant results.
In comparison, I will recommend using Stratified Random sampling, as this sort of probability
investigative strategy is a standout amongst the most usually utilized techniques and includes the
division of the entire populace into various strata. These strata are particularly used to select and
furthermore exceptionally comprehensive in nature.

10
References
Baker, R., Brick, J.M., Bates, N.A., Battaglia, M., Couper, M.P., Dever, J.A., Gile, K.J., and
Tourangeau, R., 2013. Summary report of the AAPOR task force on non-probability sampling.
Journal of Survey Statistics and Methodology, 1(2), pp.90-143.
Fassnacht, F.E., Hartig, F., Latifi, H., Berger, C., Hernández, J., Corvalán, P. and Koch, B., 2014.
Importance of sample size, data type and the prediction method for remote sensing-based estimations
of aboveground forest biomass. Remote Sensing of Environment, 154, pp.102-114. Denscombe, M.,
2014. The good research guide: for small-scale social research projects. McGraw-Hill Education
(UK).
Himelein, K., Eckman, S., Murray, S. and Bauer, J., 2016. Second-stage sampling for conflict areas:
methods and implications.
Al Ghayab, H.R., Li, Y., Abdulla, S., Diykh, M. and Wan, X., 2016. Classification of epileptic EEG
signals based on simple random sampling and sequential feature selection. Brain informatics, 3(2),
pp.85-91.
Singh, R., Kumar, M., Singh, R.D. and Chaudhry, M.K., 2013. Exponential ratio type estimators in
stratified random sampling. arXiv preprint arXiv:1301.5086.
Acharya, A.S., Prakash, A., Saxena, P. and Nigam, A., 2013. Sampling: Why and how of it. Indian
Journal of Medical Specialties, 4(2), pp.330-333.
Rada, V.D. and Martín, V.M., 2014. Random route and quota sampling: Do they offer any advantage
over probably sampling methods?
Heyer, R., Donnelly, M.A., Foster, M. and Mcdiarmid, R. eds., 2014. Measuring and monitoring
biological diversity: standard methods for amphibians. Smithsonian Institution.
Albuquerque, U.P., de Lucena, R.F.P. and de Freitas Lins Neto, E.M., 2014. Selection of research
participants. Methods and techniques in ethnobiology and ethnoecology, pp.1-13.

11
Acharya, A.S., Prakash, A., Saxena, P. and Nigam, A., 2013. Sampling: Why and how of it. Indian
Journal of Medical Specialties, 4(2), pp.330-333.