Course Name- B.A. (Hons.

) Economics

Paper Name – Data Analysis

Paper No. : Skill Enhancement Course (SEC)

Module No. 1

Module Title
What is Data? and Types of Measurement
Academic Script
Hello dear students, welcome to the course Data Analysis. Topic for today's discussion is-
'What is Data and Types of Measurement'.

Objectives of the Module:

Introduction: In the present era, data analysis is the most important area of study. In every
field we are dealing with some kind of information, which is further analysed for decision
making process. This information is perceived as the vehicle for overall growth and
development of the economy because it is employed for undertaking the vital economic and
business related decisions. Generation of information or data is an integral part of data
science. To reach at the decision making process or important conclusions from this raw
information, data is processed and analysed with the help of either simple or complex
statistical techniques. In short, data is an important part of every discipline, so it becomes
imperative to learn what does the term 'data' convey? What are the different forms and
sources of data and measurement types? Which kind of statistical techniques can be
undertaken on the data? In the present module, we will delineate the term 'data', its various
types and measurement scales used for defining the form of data.

Let's define the term data.

Data: Data, which is also known as statistics, refers to 'datum', which simply means the
anthology of items or information. Information may be either qualitative or quantitative in
nature depending upon the purpose of study. Here the term statistics is broader one as
compare to term 'data'. In narrow sense, statistics means data, while in broader sense it means
statistical methods which are related to collection, tabulation, compilation, analysis and
presentation of data. In nutshell, this processing of data will assist us in deriving the major
policy implications.

Therefore, data is defined as the collection of information, which may include facts, figures,
statements and observations. In our routine life, we usually take note on data in various forms
like inflation rate based on consumer price index is found to be 6.44 per cent in February,
2023 in India, another example is as per CMIE estimates unemployment rate in India is
estimated at 7.8 per cent in March, 2023, take another example i.e. average height of students
in class was found to be 5 feet etc. So, in all these illustrations, some kind of data is presented
in meaningful way.

One important point about data is that it can be obtained from two major types of sources
namely – primary and secondary. When the researcher himself collects the required
information, then it is known as primary data for him, on the other hand when researcher use
the data collected by other persons or agencies then it will be secondary data for him. Choice
among these two sources depends on the requirements of the study. After obtaining the
factual information from either of the source, the next step is to analyse this information.
Analysis of data or information relies on the nature of data i.e. whether the information is of
qualitative nature or quantitative nature. For instance, data on caste, religion, occupation,
gender etc. implies the qualitative data, while data on age, family size, height, weight etc.,
represents the quantitative data.

Types of Data: From the earlier discussion, it is clear that there are two important types of
data, which includes quantitative data and qualitative data. Let's explain them one by one.

Qualitative Data: Under this category, qualities, attributes or features of an item are
considered. Like complexion of a person, texture of object, personality of a person, type of
dwelling etc. Further, this type of data is classified into three categories i.e.

a) Attributes
b) Nominal
c) Ordinal

Attributes: In case of attributes, there are two categories of information, one is concerned
with the presence of attribute and another with the absence of that attribute. For instance,
consider the case of gender variable, which has two categories within it i.e. male and female,
similarly information with yes or no type options is also part of qualitative information under
the attributes category.

Nominal Data: Second important category of qualitative data is the nominal data, which is
having more than two categories inside it. Like marital status of a person will include various
categories such as married, unmarried, widow, divorce, separation. We can take another
example of nominal data i.e. type of cooking fuel, which includes LPG, biogas, dungcake,
wood, oil etc.

Ordinal Data: Similar to nominal data, there exists ordinal data, which also includes more
than two categories. Difference between two is, ordinal data needs to be in order either
decreasing or increasing based on the quality of data, while for nominal data order is not
necessary. Take the case of categories of students according to their performance i.e.
intelligent, more than average, below average, poor performer etc. Another most common
example is based on the agreement i.e. strongly agree, agree, neither agree nor disagree,
disagree, strongly disagree.
Quantitative Data: Data, which can be quantified in number system, is known as
quantitative data, like data related to height of the students, weather of Punjab in last month,
number of books available in your library etc. Quantitative data is further classified into two
types i.e. discrete and continuous data. This classification is actually related to the
presentation of the numerical figures.

Under discrete quantitative data, only proper number or whole numbers are considered. It can
be illustrated with an example, like count of students in the college in particular class say BA
Honours in Economics is 50. This figure can't be 50.5 or 49.5. Therefore, under this category
only whole numbers i.e. 0, 1, 2, 3, 4,……. are included. In contrast to discrete quantitative
data, there exists continuous quantitative data, which covers all the numerical values related
to the information. Thus continuous quantitative data consists of both whole as well as real
numbers. For instance, weight of a person is 56.5 kg, or exact time of occurrence of
earthquake, marks of the students, exact age of the person in years, months and days etc.

In nutshell, both types of data have their own relevance in the analysis of research problem.
Qualitative data underline the characteristics of the data, while quantitative data assist in
quantifying the data so that important results can be obtained. Further, qualitative data leads
to deeper probe in the research thus highlighting the research gaps, which can be filled by
further research. Moreover, it helps in such research areas where mathematical computations
are unfeasible. On the other side, with the help of quantitative data, subjective aspects can be
estimated. To conclude, quantitative and qualitative data are complementary for each other.
Because to carry a research we seek quantitative as well as qualitative information from the
respondent side.

Classification of Data on the basis of Time dimension: Under this classification three
important types of datasets are considered namely time series, cross – section and pooled
data. When the information for particular variable is collected over the period, then it is
termed as time series data. Usually, data is gathered at regular intervals like on daily basis or
quarterly basis or on annual basis or after five years or after ten years etc. Examples of time
series data includes – Human Development Reports, GDP estimates, employment or
unemployment rates etc. In case of cross section data, factual information is collected for
more than one item or variable for a particular time frame like data collected by government
on census of population. Category of pooled data lies somewhere between the time series and
cross section data, i.e. when both of these data types are merged then it is termed as pooling
of data or pooled data. For instance, information or data gathered on the variables monthly
per capita expenditure, poverty rate and income for the period 1990 to 2020, will be
considered as the pooled data. In case of pooled data it is not necessary that information will
be collected from same respondent over the period. When the information is collected for the
same set of variables from the same respondents over a period, then special kind of dataset is
formed, which is popularly termed as panel or longitudinal data. It is actually used for
estimating the change.

Types of Measurement Scales: Notion of measurement is actually associated with the unit
of measurement or estimation, which can be either a number or order or label. One of the
important aspects of the research study is related to the measurement of the variables, which
means allocating numerical value or order or rank to the features of certain objects. One point
should be noted here that measurement is related to features of object not with the
measurement of object. In case of social sciences, we are not measuring persons rather their
characteristics like their behaviour, attitude, perceptions etc. The benefits of assigning the
numbers are – firstly, it allows the mathematical treatment of the information and secondly, it
assists in understanding the estimated results. For getting reliable results from the number
system, features should be assigned in uniform pattern. For serving this purpose, scaling is
employed, which is just expansion of measurement. Scaling is a technique, which consists of
the generation of continuum or range, which will indicates the measurement of the objects.
For example, for measuring the agreement level for work – life balance, scaling of 1 to 5 can
be employed. Thus, scaling implies the degree or extent of agreement i.e. 1 is for strongly
disagree and 5 for strongly agree. In simple words, scaling explains the manner in which
numerical values are assigned to different degrees of judgment, attitude, perceptions and
other related aspects. Under data science field, we consider four important types of
measurement scales, which include nominal, ordinal, interval and ratio scale.

Starting discussion with the first one i.e. Nominal scale.

Nominal Scale: As the name implies nominal is related to name, accordingly nominal data is
that kind of data under which names or labels are considered. To put it in other words,
nominal scale contains the qualitative information in the form of different options. Like, if the
question is choose your gender from the response options as male, female and transgender.
Similarly, country of your residence will be having many options say North, South, West and
East part of the country. So, these options are implying the nominal scale, when there are
only two options then it is known as dichotomous, when there is proper sequencing or
ordering of options then it is termed as nominal scale with order and lastly, when no
particular order is followed in the response options then it is defined as nominal scale without
order. Another set of queries, where nominal scale of measurement is employed are bank
account number, identification number, phone number, PAN number, Adhar card number etc.

Ordinal Scale: When order of response options matters then such kind of qualitative data is
known as ordinal measurement scale. Basically, this scale focuses on the features of the
objects in comparison terms. Take the case of rating or feedback of higher education
institutes, hotels, schools, hospitals etc. rating or order simply implies the better or worse
performance of the objects under consideration. For example, in case of question – Are you
satisfied with the services provided by your bank? Response options can be strongly
dissatisfied, partially dissatisfied, satisfied, partially satisfied and strongly satisfied. So, these
options represents the particular order of the options and the data obtained in ordinal manner
is known as ordinal data.

Interval Scale: It is actually a numeric scale of measurement, which considers the order as
well as the features of the object. Under this scale, succeeding values shows the equal value
or quantum of features which are being estimated and its base value is usually not fixed. This
additional feature of interval scale makes it more useful scale as compared to ordinal scale.
For instance, information recorded for variables as temperature, longitude, latitude, dates in
calendar etc.

Ratio Scale: Ratio scale represents the highest level of measurement and it also comprises all
the important features of a scale. This scale shows the original quantity of variables. In
contrast to interval scale, which does not have certain or fixed value, ratio scale of
measurement do have certain or fixed value, usually zero. In short, ratio scale is an expansion
of interval scale starting with fixed value i.e. zero. For instance data on salary of the workers
in a company, weight of the individuals or products, number of times one go to hospital or
bank or shopping complexes etc. shows the case of ratio scale.

So, these are four important scales of measurement, which played an important role in
description of data. Nominal scale will result in nominal data, ordinal scale will give ordinal
data, data associated with interval scale will be treated as interval data and ratio scale will
provide the ratio data. Further, operations of data science or statistics can be performed on
these scales on the basis of their respective properties. Like only mode (measure of central
tendency) can be performed on the nominal scale; median and mode can be applied on the
ordinal scale based data; mean, median, mode, standard deviation and coefficient of variation
can be performed on interval scale, while in case of data based on ratio scale of measurement
all kind of statistical methods can be employed like mean, median, mode, geometric mean,
standard deviation, coefficient of variation etc.

Conclusion: Important conclusions of this module are –

- Data simply means the anthology of items or information, which can be either
qualitative or quantitative.
- Qualitative data is related to the features, attributes or qualities of an object.
- Data which can be quantified in number system is known as quantitative data.
- Qualitative data is classified into three classes namely attributes, nominal data and
ordinal data.
- Term measurement implies the unit of measurement or estimation, which can be
either a number or order or label.
- Scaling implies the way of assigning the numbers to different degrees of judgment,
attitude, perceptions and other related aspects.
- There are four important types of measurement scales, namely - nominal, ordinal,
interval and ratio scale.

This is all about today's lecture. Hopefully, you have developed an idea about the term data,
types of data and types of measurement scales. This elementary information on these
concepts will be highly useful in the analysis of data.

Thanks a lot for listening this lecture, meet you shortly with another video lectures.