SEC: IT SKILLS AND DATA

ANALYSIS
PRESENTATION ON :
1. SAMPLING AND NON SAMPLING
ERRORS
2. QUESTIONNAIRE
sampling and
non sampling
errors
 INTRODUCTION
Error generally refers to a mistake or fault. However,
in the field of data science, the term "error" refers to
the discrepancy between the estimated value of the
sample statistic and the actual value of the
population parameter.
 TYPES OF ERRORS
Generally speaking, errors fall into two categories:

1. Sampling errors
2 Non sampling errors
 Sampling Errors
Sampling or statistical errors happen when conclusions about
the population based on sampling units are incorrect or differ.
When a sample is not representative of the entire population,
sampling error occurs. When selecting the sampling units, the
sample results might not precisely match the population
parameter's actual value. The sample survey only looks at a
small portion of the universe, so some error is unavoidable.
This is called sampling errors.
Example of Sampling Error

samp
le he
POPULATION SAMPLE does re
repre not
sent
popu true
lation
 Types of Sampling Errors
There are 2 types of sampling errors :

Biased Errors Unbiased Errors

When there come a bias either in Also known as random sampling
selection or estimation of error, because these errors occur
The sample can include 220
sampling by chance. That is variations in the
students from the school
units, then it will lead to biased sampling and population units.
errors. Like in case of convenience Therefore, unbiased
sampling personal biases may error refers to the difference
affect the inferences of between the sample statistic and
study and can result in the biased its corresponding parameter.
errors.
 Reasons of Sampling
Errors
Faulty process of sample Faulty Collection of data Bias in estimation Method
selection Faulty data collection occurs Bias in the estimation method
A faulty process of sample when the data gathered is happens when the technique
selection occurs when the inaccurate, incomplete, or biased used consistently
method used to choose the due to errors in the method, tools, overestimates or
sample is biased or not or process used. This can lead to underestimates the true value.
representative of the misleading conclusions, as the This systematic error skews
population.The sample doesn't data does not accurately results, leading to conclusions
accurately reflect the broader represent the reality being that are not accurate
group being studied. studied. representations of the
population.
Methods of Reducing
Sampling Errors
Sampling errors are composed of biased
and unbiased errors. To avoid biased
errors, follow random sampling
techniques. Minimizing random sampling
errors involves increasing sample size,
which enhances efficiency and reliability.
The reduction in sampling errors is
inversely proportional to the square root
of the sampling size. .
 Non sampling Errors
Non-sampling errors are types of errors in data
collection and analysis that occur not because of the
sample size or the way the sample is selected, but due
to other factors in the survey process. They can occur
at any stage of data collection, processing, or analysis
and can significantly affect the accuracy of the results.
 Reasons of non sampling errors
Data Collection Errors: These occur when there are mistakes in how data is
gathered, such as misreporting, recording errors, or incorrect interpretation of
responses. These errors can lead to inaccurate data being collected.

Processing Errors: Errors that happen during the handling of data, such as
during data entry, coding, or analysis. These errors can result in the
misrepresentation of data in the final analysis.

Non-response Bias: This occurs when certain individuals do not respond to a

survey or study, leading to a skewed representation of the population. The
missing data can cause biased results if the non-respondents differ
significantly from respondents.
Measurement Errors: These arise when there are inaccuracies in the way data is
measured or collected, often due to poorly designed survey questions or faulty
measurement tools. This leads to data that does not accurately reflect the true
values.

Sampling Frame Errors: Occur when the sampling frame (the list from which the
sample is drawn) does not accurately represent the entire population. This can
result in certain groups being underrepresented or overrepresented in the
sample.

Interviewer Bias: Happens when the interviewer's behavior, tone, or body

language influences the respondent's answers. This can lead to data that is
not a true reflection of the respondent's actual opinions or behaviors.
 Measures of Statistical
Errors
Absolute Error : Absolute error estimates the
difference between the actual value and
estimated value of specific variable or item. In
mathematical terms, it is expressed as – AE = |a −
e| , AE denotes the absolute error, a implies the
actual value and e the estimated value.
Relative Error : It is defined as the ratio of absolute
error to the actual value,
RE = AE /Actual Value
This is used when we have to compare different
measuring units
Questionnaire
 Introduction
A questionnaire is a research tool that consists of questions or
prompts used to collect data from respondents. Questionnaires
are often used in market research, social sciences, and health
sciences. They can be used to gather both quantitative and
qualitative data, which can be analyzed to draw conclusions or
provide longer, more specific explanations.
 Importance
Researchers must design an effective questionnaire to ensure accurate
data collection. The framework and technique of asking questions are
crucial components. Exploratory and quantitative information types are
essential for understanding study objectives and hypotheses. Exploratory
information, obtained through interviews and discussions, helps frame
study objectives and hypotheses. Quantitative information, obtained
through formal questionnaires, is used to test hypotheses and analyze
results. Thus, designing of a questionnaire holds an important place in the
data collection process.
Qualities of a Good Questionnaire
• Satisfies research objective and hypotheses.
• Covers all aspects of the research study.
• Allows for complete and accurate information collection.
• Purpose of survey should justify respondent's interest in
providing data, only then the informant will provide accurate
information.
General Principles of Designing the Questionnaire
• Covering Letter: Include the purpose of the survey, benefits,
confidentiality of data, and self-addressed envelope.
• Precise Number of Questions: Limit the number of questions to
ensure respondent's interest and clarity.
• Arrangement of Questions: Ensure logical order of queries for
accurate and spontaneous responses.
• Ambiguous Questions: Avoid questions that don't clearly convey the
meaning, such as 'Do you smoke?', as they may not indicate the
frequency or type of smoking.Questions should be simple to
understand and free from ambiguities.
• Avoid personal questions: It is possible to acquire such
information through in person interviews.
• Prioritize objectivity: Questions should be objective, with multiple
options, categorized into shut questions and open questions.
• Avoid computational queries: Questions requiring computations
from informants should be avoided.
• Design cross queries: The questionnaire should provide internal
checks on data accuracy, with related questions included for
verification.
• Pre-test the questionnaire: Before mailing, the questionnaire
should be pre-tested on a small group to identify practical
problems and weaknesses.
 Editing and Pre Testing a
Questionnaire
Pre-testing a questionnaire helps identify problematic questions, reduce
error costs, and ensure that questions are easy to understand. It also
helps identify response latency and ensures accurate and reliable data. It
is recommended to test on a minimum of 12-50 people to avoid issues
related to finance, time, and energy.
 Types of Pre Tests
Expert Driven Respondent Driven
Expert-driven pre-tests involve Respondent-driven pre-tests
expert suggestions to edit involve selecting friends or
questionnaires, identify issues, colleagues or a small subsample
and improve frameworks. They of the sampled population to test
help tackle measurement issues, the questionnaire. This method
grade questions, and cross-verify helps identify potential problems
vital aspects. and allows for further editing and
improvement of the
questionnaire.
Important Issues to Identify in Pre Testing
Refusal or Inability to answer: Investigator should keep in mind that frequency of
questions with response code as ’I don’t know’ or ’Not applicable’ or ’others’
should be restricted.

“Other” Response Query-If the respondents frequently choose 'other' option in

the questionnaire, this suggests need for additional options. Researchers should
avoid 'others' and provide relevant options or leave space with open ended query

Little or No Variation in Responses- If a sub-sample is selected across

heterogeneous groups, it must reveal variations in responses, while if there are no
variations, it indicates a problem with the question or sub-sample selection.
Important Issues to Identify in Pre Testing
Sensitive Questions- Pre-testing questionnaires across cross-sectional
groups helps identify sensitive questions, which can be removed or
reframed based on the informant's hesitation or skipping of certain
questions.

Order of the Response Options-The order of response options in a

survey can influence the responses of the respondents, potentially
leading to biases. To avoid this, investigators can shuffle the order of
responses.
 Advantages of Pre Testing
Identifies ambiguities for editing, allowing removal of
irrelevant and addition of relevant questions.

Provides insight into non-response frequency, enabling

reconstruction or replacement of non-response
questions.

Enhances cooperation between respondent and

investigator.
 Thank You
Presented By-
Ananya Goel- 230219
Ishita Srivastava-230596
 Resource Page
Use these icons and illustrations in your Canva Presentation. Happy designing!
Don't forget to delete this page before presenting.
 Sampling Techniques
If every is equal to 10 students and the sample size must be 60,
inspect how each technique can be used.

CONVENIENCE SAMPLING SIMPLE RANDOM SAMPLING

The sample was selected as the

The sample is chosen randomly, providing
first listed group. The technique was
equal chances for all members.
chosen for its simplicity.
 Sampling Techniques 60
200
sample
population
STRATIFICATION
Dividing the sample into subgroups
before drawing random samples Level Total Number Proportional Number

Year 7 30 30 × 60/200 = 9

Year 8 30 9

Year 9 40 40 × 60/200 = 12

Year 10 30 9

Year 11 30 9

Year 12 40 12

POPULATION = 200 SAMPLE = 60

 Sampling Techniques
SYSTEMATIC SAMPLING

Selecting one member for every

nth of the population.
Example

The 200 students'

names may be listed
alphabetically then
every 3rd student will
be part of the sample.

Note
The starti
ng point m
be chosen ay
at random
.
 ANSWER KEY

Let's Practice!
Identify the sampling technique used in each scenario.

CONVENIENCE
STRATIFICATION SIMPLE RANDOM
SAMPLING

Every tenth person All the names of the The first 20 people who
boarding a plane was given parents were placed in a basket, passed by were chosen
a questionnaire. then 20 were selected to review the taste of
to join the activity. the chocolate bar.
Assignment
What sampling technique
must be used to
represent your school's
population?

Provide at least two reasons. Recall

the features of a good sample.
 Reference
Mathspace. “9.01 Sampling Techniques | Grade 7 Math |
Common Core 7 - 2022 Edition.” Accessed 21 June 2023,
https://mathspace.co/textbooks/syllabuses/Syllabus-
1156/topics/Topic-21902/subtopics/Subtopic-279872/
?searchString=&activeTab=theory
 Press these keys
while on Present mode!
B for blur U for unveil Q for unveil

D for a drumroll C for confetti # Any number

from 0-9 for
a timer
O for bubbles M for mic drop