Skip to content

How to write a thesis (1.2) Basic terms

This article explains a concept that is good to know before you write a thesis. For example, we will cover constructs, variables, hypotheses, theories, models, etc., to lay the foundation for what it means to think and write like a researcher. For a list of articles related to the basics of writing a thesis, you can take a look at the following links:

■ Unit of Analysis

A unit of analysis refers to the object of the investigation set by the researcher. It can be a person, a group, an organization, a technology, and so on. If we examine a customer’s shopping behavior, the unit of analysis is the individual. If we examine a teamwork within an organization, the unit of analysis is a group. If the researcher’s interest is how to create a web page that is attractive to the user, then the unit of analysis is the web page (not the user).

Understanding a unit of analysis can be complex. For example, when you want to study a crime rate in a particular region, the unit of analysis is that region, not criminal or crime itself. However, if the goal of the study is to compare crimes in different regions, the unit of analysis is a crime.

The unit of analysis should be well understood because it tells you what kind of data to collect. If the unit of analysis is a web page, then the data for the web page should be examined. In this case, do not interview users for how they use web pages.

■ Construct and Variable

In order to explain something, you have to develop a concept. A concept is a trait that generalizes an attribute or is associated with an object, such as a person/event. These concepts differ in the degree of abstraction. For example, the concept of “weight” is more specific than the concept of “happiness.”

A construct is an abstract concept created or selected by a researcher to explain a phenomenon. A construct can be as simple as “weight” or it can be a collection of several complex concepts, such as “communication skills”. For example, the construct “communication skills” can be measured by words, grammar, spelling, etc. Where “weight” is a one-dimensional constuct, and “communication skills” are multi-dimensional constructs.

In scientific research, a construct must have an accurate and clear definition. For example, if you use the construct of “income,” you should clearly state whether it means a monthly salary, yearly salary, net salary, or the total income of a family member, etc.

What is often used with a construct is a variable. In general, a variable means a number that can change. The opposite word for a variable is a constant. However, variables in scientific research refer to the measurement of abstract construct. For example, the construct “intelligence” is often measured by a variable “IQ”.

A variable that describes another variable is called an independent variable, and a variable that is described by another variable is called a dependent variable. The variable that is described by an independent variable, but which describes a dependent variable is called a mediating variable or intermediate variable. Variables that affect the relationship between independent and dependent variables are called moderating variables. The variables that need to be controlled in the study are called control variables.

For example, let’s say you investigate whether “intelligence” has a positive effect on an individual’s “earning potential.” In this case, “intelligence” is an independent variable because it describes “earning potential”, and “earning potential” is a dependent variable. The parameter may have “academic achievement”. This is because “intelligence” has a positive impact on “academic achievement,” and “academic achievement” again has a positive effect on “earning potential”. The variable that affects the relationship between the two variables is “effort”, which is the moderating variable.

An image that describes the configuration concepts and variables.
Construct and Variables
■ Proposition and Hypothesis

As shown in the image above, there is a relationship between the constructs, and the researcher works to determine whether the relationship is true or false. The pattern of relationships is called a proposition. A proposition refers to a tentative and speculative relationship between the constructs that are stated in the form of a pre-declaration. In the example, the proposition could be “Students with high intelligence have high academic achievement”. A proposition does not necessarily have to be true, but it must be empirically verifiable using data.

Propositions cannot be tested directly because they represent the relationship between abstract constructs. Therefore, it is necessary to examine the units of measurement of the constructs, that is, the relationship between the variables. The empirical formulation of the relationship between variables is called a hypothesis. In the example, the variable of “intelligence” is “IQ”, and the variable of “academic achievement” can be “grade”. Therefore, the hypothesis is that “students with high IQ have high grades”.

You can set the weak hypothesis as “IQ is associated with academic achievement”. Or you can set the strong hypothesis as “IQ has a positive impact on academic achievement.” Statements such as “students are highly intelligent” cannot be hypothesized because there are no independent variables and no dependent variables.

■ Theory and Model

A theory is a set of constructs and propositions that are systematically constructed to explain and predict phenomena or behaviors of interest within a particular boundary condition and assumption. Generally, propositions connect two or three constructs, and theories mean a more complex and abstract system of propositions or hypotheses.

A model is the whole or part that makes up a system. A theory explains the phenomenon, a model represents the phenomenon. For example, a marketer can use the model to determine advertising costs based on variables such as the previous year’s advertising costs, revenue, market growth, and competitors’s status. Such a model is useful, but the aforementioned variables may not be all variables that determine the advertising cost (i.e., the advertising cost is not necessarily explained by those variables).

Bibliography

  • Social science research: principles, methods, and practices (Anol Bhattacherjee) – chapter 2

Leave a Reply

Your email address will not be published.