Report on Data Sampling

1. Introduction
Data sampling is a foundational process in research where a portion or subset of
individuals or items is chosen from a larger population. This selection aims to
create a sample that accurately represents the characteristics of the entire group.
The primary goal of this technique is to enable researchers to gather information
and make inferences about the population without the necessity of studying every
single member. Technical reports and papers often summaries information in a way
that needs to be easily accessible and concise. Data sampling contributes to this by
allowing researchers to work with manageable datasets that represent larger
phenomena. In the context of academic studies like bachelor's, master's, and
doctoral programmers in engineering, understanding effective report writing and
presentation in a professional format is crucial. Data sampling is a key part of the
methodology section in such reports, especially for engineering tasks that include
writing technical reports. The format of a guide on report writing itself can serve as
an example by following its own prescriptions. When compiling a technical report,
one of the initial planning questions is identifying the target reader. In engineering
practice, this could be a client, colleague, manager, or junior, while in academic
writing, it is typically an independent examiner who was not involved in the work
being reported. The author must provide sufficient background and detail to
convince the examiner. Effective sampling contributes to the credibility of the
results presented to the examiner.
2. The Necessity of Data Sampling
Sampling becomes necessary due to several practical and methodological reasons:
•Economy of time and money: Studying an entire population (census) can be
prohibitively time-consuming and expensive [3-5]. Sampling offers a more
efficient alternative [3, 5].
•Possibility of detailed knowledge: By focusing resources on a smaller sample,
researchers can conduct more exhaustive and intensive studies [4, 5].
•Utility in experimental studies: Sampling is essential for selecting participants or
units of observation in experiments and surveys [1, 2, 6].
•Reliability: When grounded in probability theory, sampling can provide reliable
and accurate estimates of population characteristics [5].
•In many circumstances, a census inquiry involving the entire population is not
practically possible due to the time, money, and energy required [3]

3. Types of Sampling:
Sampling designs are broadly categorized into two main types: probability
sampling and non-probability sampling
3.1 Probability Sampling In probability sampling, every element within the
population has a known probability of being selected for the sample. This
characteristic is crucial because it allows for the estimation of sampling error and
permits the generalization of findings from the sample to the larger population.
 Simple Random Sampling: In this method, every element in the population
has an equal and independent chance of being included in the sample.
Randomization techniques, such as the lottery method or using random
number tables, are employed to ensure this equal chance. A sample selected
by randomization methods is known as a simple random sample.
 Systematic Sampling: This technique involves selecting elements from an
ordered list at regular intervals. Typically, a random start is introduced to
maintain a degree of randomness. The method requires complete
information about the population, arranged systematically. The size of the
sample is decided, and then every N/n individual is selected from the list,
where N is the population size and n is the sample size. Systematic sampling
is considered an improvement over simple random sampling in some
instances.
 Stratified Sampling: When the population is not homogeneous, stratified
sampling is applied to obtain a representative sample. The population is
divided into homogeneous subgroups, or strata, based on specific
characteristics. A random sample is then drawn from each stratum. Stratified
sampling can be proportionate, where the sample size in each stratum is
proportional to the size of the unit in the population, or disproportionate,
where the sample size depends on personal judgment and convenience.
 Optimum allocation aims to select units from each stratum proportionally to
their size within the population.
 Cluster Sampling: This method involves dividing the population into intact
groups or clusters. A random selection of these clusters forms the sample,
and all elements within the selected clusters are included in the study.
Cluster sampling involves grouping the population first and then selecting
the groups rather than individual elements.
 Multi-stage Sampling: This is an extension of cluster sampling that
involves sampling at multiple levels. For example, one might select states,
then districts within those states, then towns within those districts, and
finally families within the selected towns. This sample type is more
comprehensive and representative of the population. In multi-stage
sampling, primary sample units are inclusive groups, and secondary units are
subgroups within these.
 Multiple or Double Sampling: This technique involves analyzing an initial
sample, and the information gathered is used to draw a subsequent sample
for further examination. It is sometimes used with mailed questionnaires to
obtain a more representative sample by drawing a second sample from non-
respondents and interviewing them. This helps eliminate bias if non-
respondents differ significantly from respondents.
 Sequential Sampling: In sequential sampling, the ultimate sample size is
not fixed beforehand but is determined by mathematical decisions based on
information yielded as the survey progresses. This method is often used in
statistical quality control.
3.2 Non-Probability Sampling In contrast to probability sampling, the probability
of selecting any specific element is unknown in non-probability sampling. These
techniques are frequently used in exploratory research or when probability
sampling is not practical.
 Incidental or Accidental Sampling: This involves selecting participants
who are easily accessible or convenient.
 Judgment Sampling: Also known as purposive or deliberate sampling, this
method involves selecting a group based on the researcher's belief or
 judgment that they are representative of the population or meet specific
criteria. Expert choice sampling is a form where samples are chosen because
they mirror a larger group regarding specific characteristics.
 Purposive Sampling: This is similar to judgment sampling and involves
selecting a sample with a specific purpose, often to include particular
individuals or cases relevant to the study. In some categorizations, purposive
sampling is listed under both probability and non-probability types,
indicating some overlap or different interpretations. However, it is explicitly
defined as involving the deliberate selection of particular units to constitute a
sample that represents the universe.
 Quota Sampling: This technique involves determining the proportion of
different categories within the population and then selecting a quota of
participants for each category based on the researcher's judgment. Unlike
stratified random sampling, the selection within categories is not random but
left to the interviewer's judgment.
 Snowball Sampling: This method begins by identifying a few participants
who meet the study criteria and then asking them to refer other potential
participants. This is particularly useful when the target population is difficult
to access.
4. Essentials and Characteristics of an Ideal Sample
An ideal sample should possess several key characteristics to ensure its ability to
represent the population accurately and the validity of research findings.
 Representativeness: An ideal sample must accurately reflect the
characteristics of the whole population. It should serve as a miniature picture
or cross-section of the entire group. A good sample avoids bias or the
researcher's preconceptions influencing its selection.
 Homogeneity: The units included in the sample should be similar to other
units in the population in relevant characteristics. Homogeneity amidst
complexity is an assumption underlying sampling, recognizing that despite
individual differences, similarities exist within a population.
  Adequacy: The sample size should be sufficient to allow the derivation of
conclusions that can be applied to the entire population. A sample having
10% of the whole data can be considered adequate in some contexts.
 Independence: Every unit in the population should have a free chance to be
included in the sample.
 Objectivity: The selection procedure should be objective, free from
subjective elements or the researcher's personal bias. While some non-
probability methods rely on judgment, the ideal sample strives for objective
selection.
 Accuracy: A good sample yields accurate estimates or statistics and
minimizes errors. The error of estimation should be as low as possible.
 Comprehensiveness: The sample should be comprehensive, closely linked
with representativeness and controlled by the specific purpose of the
investigation. While comprehensiveness in traits is important, it does not
automatically guarantee representativeness.
 Practicability: An ideal sample is practical to obtain and use for research
regarding time, money, and resources. Sampling itself offers advantages like
economy of time and money.
5. Important Considerations in Data Sampling
Careful planning and execution are necessary when undertaking data sampling.
Several crucial aspects must be considered:
 Defining the Target Population: Researchers must clearly identify the
entire group to which they intend to generalize the research results. The first
step in sampling is to define this target population.
 Determining the Sample Design: The most appropriate sampling method
(probability or non-probability) must be selected based on the research
objectives, available resources, and the desired level of generalizability. The
method of sampling should be specified in the research proposal. The
research plan must outline the method of sampling.
  Determining the Sample Size: The sample size should be large enough to
provide adequate statistical power for the intended analyses. The population
surveyed must be sufficiently large for appropriate statistical analyses.
Deciding the size of the sample is a necessary step in methods like
systematic sampling.
 Addressing Potential Biases: Researchers must recognize and mitigate any
systematic errors in the sampling process that could skew the results. Bias
can compromise research results. Using third parties or survey software that
prevents identification can help minimize bias and maintain confidentiality
in surveys. Awareness of potential biases within different sub-populations is
also crucial.
 Method of Drawing the Sample: The chosen sampling technique must be
implemented carefully and consistently. The research plan must clearly state
the method of sampling.
 Ensuring Representativeness: A key goal is to obtain a sample that
accurately reflects the characteristics of the target population. This might
involve defining inclusion and exclusion criteria for participants. If the
population is not homogeneous, stratified sampling can be used to help
obtain a representative sample. The possibility of representative selection is
an assumption underlying sampling, supported by probability theory and the
law of statistical regularity. However, difficulty in selecting a truly
representative sample is a limitation of sampling.
6. The Role of Sampling in Research Design and Process
Sampling is an integral part of the overall research process and research design.
The research process involves a series of steps, including formulating the research
problem, conducting a literature survey, developing hypotheses, preparing the
research design, and determining the sample design. Determining the sample
design is a distinct step in this process.
A research design serves as the framework or plan for a study, guiding data
collection and analysis. It is a blueprint for collection, measurement, and analysis
of data. Research design decisions include determining the sample design. The
overall research design can be split into parts, including the sampling design,
which deals with the method of selecting items to be observed. A good research
design should provide adequate information, including identifying the exact
research problem and the objective of the research. It also considers appropriate
statistical analysis.
The research proposal, which outlines the plan for a study, must include details
about the sample design. The proposal should specify how the units in the target
population will be selected and used and describe the method of drawing the
sample and the number of subjects to be selected. Faulty generalizations may result
if the sample is not thoroughly analyzed and precisely described in the proposal.
The sampling plan should also meet criteria such as obtaining an accurate list of
population units.
When collecting data, the researcher selects methods based on the sampling design.
After data collection, the analysis of data follows, which often involves statistical
techniques that depend on the sampling method used.
7. Data Analysis and Sampling
Data analysis encompasses activities, often employing statistical methods and
techniques, particularly in behavioral research. The analysis of data should be
sufficiently adequate to reveal its significance, and the methods of analysis used
should be appropriate.
Statistical analysis is crucial for proving that outcomes are valid "beyond
reasonable doubt". Different types of statistical analysis exist, such as one-
dimensional, two-dimensional, and multi-dimensional statistics, as well as null
hypothesis testing. Statistical support for results is necessary. When analyzing data
from surveys, statistical analysis is employed. The analysis of results from surveys
must take into account any imbalance in representation across sub-populations.
Various statistical tests, such as the paired Student's t test or unpaired t test, can be
used depending on the survey design (e.g., surveying individuals twice or
comparing different populations). The researcher should specify how the data will
be ordered and reduced and describe the statistical procedure to be used in the
research proposal. This includes whether analysis will be done hypothesis-wise. If
a complex design or obscure statistical test is planned, it should be indicated. The
analysis of data gathered during the project execution involves steps like
establishing categories, coding, tabulation, and drawing statistical inferences.
In experimental research, quantitative measurements (data) are observed to
examine the truth contained in a hypothesis. Statistical analysis is also a key part of
analyzing experimental data. The method of analysis must be appropriate and
statistically valid. Researchers should comment on the accuracy of their
experimental and modeling procedures. The quality and relevance of cited papers
and comments made on their contribution are evaluated in a literature review. A
reviewer assessing a scientific paper will consider if the method of analysis is
appropriate and statistically valid.
Data presentation techniques are also important. If more than three parameters are
measured, a clear graphical representation of all the data may not be possible, and
tables of statistical parameters are one preferred method. Charts and illustrations,
such as histograms and 2D graphs, can be used to illustrate statistical information.
The axes of graphs must be correctly named, and their units given. Tables should
have units in the column headings.
8. Reporting on Data Sampling and Methodology
The final step in the research process is the preparation of the report or thesis. The
research report provides a detailed account of the entire research experience,
including the problem definition, hypotheses, data gathering, analysis,
interpretation, and conclusions.
The main text of a research report includes the introduction, literature review, and
methodology. The introduction should clearly state the objective and explain the
methodology adopted. The methodology chapter describes and justifies the data
gathering method used and outlines how the data was analyzed. It should begin by
describing the chosen method and why it was most appropriate, citing reference
literature. This chapter should detail every step of the data gathering and analysis
process. Depending on the method, areas typically addressed include the
description of the research design, population and sample description and
justification, development of the instrument, administration of the instrument,
coding of data, and description of data analysis, including statistical analysis and
tests performed. For survey research, this involves reporting on the sampling
methods used for respondent selection. The research proposal should identify the
schedules and procedures for acquiring and recording data.
In the findings chapter, only the results from data analysis are presented, without
discussing other research literature or the implications of the findings. Tables
and/or figures should be used to illustrate and summarize all numeric information.
The discussion chapter interprets what the findings mean in relation to the
theoretical body of knowledge and the profession, answering the "So what?"
question.
When documenting experiments, sufficient information must be provided to prove
the credibility of the experimental results. In documents like theses, the theoretical
foundations of experiments must also be discussed. Processed results should
explain statistical processing and compare data sets with theory or other published
data. Conclusions involve the interpretation and discussion of results, emphasizing
important findings and pointing out unexpected results. Comparison with
theoretical or other published data is also part of the conclusions.
A scientific paper or report should clearly and unambiguously state what the
researcher wanted to say. A checklist for finalizing a report includes questions
about whether it follows the prescribed sequence, if titles and headings are clear
and concise, if the table of contents and lists correspond to the text, and if graphs
and tables are correctly labeled and formatted. For surveys, it should also be
reported if the abstract states how the project originated, what was intended, how it
was done, and what was found.
The method used for the selection of respondents and the method of reporting
should be included in summaries of research papers. When reporting research
outcomes from a survey to the respondents, it is necessary to ensure the report can
be understood by all, possibly using a journalistic style with histograms and 2D
graphs.
Reviewers of scientific papers evaluate aspects including the literature review,
experimental methods, results and analysis, and discussion and conclusion. They
assess if the experimental method is appropriate and capable of yielding correct
results to the required accuracy, if the method of analysis is appropriate and
statistically valid, and if results have been independently verified. The sampling
design and the protocol for collecting and analyzing data are key components
described in the methods section of journal-style scientific writing. Enough
quantitative detail should be provided about the experimental protocol or data
collection method so others can reproduce the work.
9. Conclusion
Data sampling is a critical technique in research, particularly in engineering and
scientific studies, driven by the necessity to make research economical, efficient,
and manageable while aiming for reliable and generalizable results. The selection
of an appropriate sampling design, whether probability or non-probability, is a
fundamental step in the research process and research design. Characteristics of an
ideal sample, such as representativeness, homogeneity, adequacy, and objectivity,
are essential for valid research outcomes. Careful consideration of potential biases,
sample size determination, and the method of drawing the sample are crucial for
ensuring the quality of the data collected. The analysis and reporting of data,
including statistical analysis and detailed descriptions of the methodology, are
directly influenced by the sampling strategy employed and must be presented
clearly and accurately in research proposals, reports, and publications. Ultimately,
a well-planned and executed sampling strategy significantly contributes to the
credibility and validity of research findings and their interpretation.