What Is a Scientific Research Topic? How to Choose and Test Feasibility

What Is a Scientific Research Topic

A scientific research topic is the scope of a problem that you want to investigate through a process with a clear question, data, and testing method. A topic is usually broader than a specific research subject. For example, “student acceptance of digital banking” is a topic, while “Factors influencing students’ intention to use digital banking in Ho Chi Minh City” is a narrowed research subject.

When choosing a topic, you should be able to answer four questions: What are you studying, who are you studying, in what context, and which variables will you use to measure it? A topic suitable for a quantitative thesis must be convertible into a questionnaire and .sav or .csv data, followed by analysis in SPSS or SmartPLS.

You can also review scientific research topics to compare ways of setting the scope. If you are just starting, What is scientific research helps you distinguish scientific research from an ordinary information summary.

A basic quantitative topic can be written as follows:

Factors X influencing Y among group Z in context B.

Here, X represents the independent or explanatory variables, Y is the dependent variable, Z is the research population, and B is the context or location. This formula is only a starting framework. You still need to check the theory, measurement scale, and data-collection feasibility before finalizing the research subject title.

Why a Scientific Research Topic Matters in Quantitative Research

A good topic helps you decide on the research model, hypotheses, questionnaire, and analysis method. If the topic only describes a phenomenon without measurable variables, it will be difficult to move on to Cronbach's Alpha, EFA, regression, or PLS-SEM.

For example, “customers’ experience with shopping applications” is still broad. You need to determine which aspects measure experience, such as ease of use, information quality, trust, or satisfaction. Only then can you ask whether those factors affect repurchase intention.

The topic also sets limits for your data. If you choose “Vietnamese consumers,” the scope may exceed what a student can collect. If you change it to “final-year students at a university in Hanoi,” the questionnaire, screening criteria, and collection plan become more specific.

You should distinguish a topic from a research field. A field may be marketing, finance, education, or information technology. A topic is a specific problem within that field. The research object answers the question of which person or phenomenon you are observing.

In quantitative research, a topic usually leads to one of three types of questions. The first describes the level or current state of a variable. The second tests relationships between variables. The third compares differences between groups. You should not combine all three types before you have a theoretical reason and suitable data.

What Makes a Scientific Research Topic Feasible

There is no single number that determines whether a topic is feasible. You can use checkable criteria before finalizing it: Can the variables be measured, is the number of respondents realistic, is the number of indicators appropriate for your resources, and does the planned method match the question?

The table below focuses on thresholds commonly used when checking the feasibility of a quantitative research subject. These are sourced reference thresholds, not a promise that every topic automatically works when it reaches the stated number.

For example, if the planned model has 5 independent variables, the 50 + 8m rule gives 90 observations. If the questionnaire has 25 indicators, the rule of 5 to 10 observations per indicator gives approximately 125 to 250 responses. These two estimates do not replace consideration of the sampling design, response rate, and your supervisor’s requirements.

The key point is that Cronbach's Alpha, KMO, and total variance explained are analysis results. You cannot write in advance that your scale will definitely have an Alpha of 0.8 or that the EFA will definitely meet the criteria. Before collecting data, you can only state that the scale will be adopted, adapted, and tested.

How to Read a Scientific Research Topic in the Output

A topic does not appear as a separate line in SPSS output. It is reflected through variable names, variable codes, the Descriptive Statistics table, Reliability Statistics, KMO and Bartlett's Test, Rotated Component Matrix, and Coefficients. When you open the SPSS file and see these tables, go back and check whether they match the original research question.

For illustration, assume that the topic concerns factors influencing intention to use a digital banking application. The model has three independent variables, usefulness, ease of use, and trust, along with intention to use as the dependent variable. The figures below are illustrative output, not the results of a real study.

Illustrative output from SPSS

When reading the tables, follow the model’s order. First, check whether the variables were entered with the correct codes. Next, review scale reliability. Then, if the research design includes EFA, read KMO and Bartlett's Test and Rotated Component Matrix. Finally, read Model Summary and Coefficients to answer the hypotheses.

If the topic refers to “intention to use” but the file contains “satisfaction” as the dependent variable, stop and check the questionnaire and data coding. An output with many attractive numbers cannot rescue a model when the variables in the data do not match the research question.

What to Do When a Scientific Research Topic Is Not Feasible

If the topic is not suitable, start with the least expensive part to change, the title and scope, before changing the data. Running SPSS again cannot turn an unclear question into a good research question.

Narrow the population and context

Change “Vietnamese consumers” to a group with clear criteria, such as people who purchased through a platform within the past six months. This makes the screening question, questionnaire distribution channel, and response-exclusion criteria more specific.

Reduce the number of variables and retain justified relationships

If the model contains too many factors and you cannot explain why they affect the dependent variable, remove variables outside the main question. Each retained variable needs a definition, a reference scale, and a suitable hypothesis.

Recheck the scale before formal data collection

You need to check whether the indicators measure the same concept or mix several different ideas. A long indicator containing two ideas, or wording that respondents find difficult to understand, can create problems at the questionnaire-design stage. If you need to adjust the wording, record the before and after versions so you can explain the change.

Change the method when the question requires it

If the question only asks about frequencies, means, or sample characteristics, regression or SmartPLS may be unnecessarily heavy. If the question tests multiple relationships between latent variables, PLS-SEM may be more suitable than simple regression, but you must prepare separate measurement and structural models.

If you have already collected data but the scale does not meet the criteria, remove an indicator only when you have statistical and theoretical reasons. Do not keep deleting indicators until you obtain the result you want. Save every data file and results table, because your supervisor may ask which indicator you removed, in which round, and why.

Distinguishing a Scientific Research Topic from a Research Subject and Research Question

These three concepts are related, but they are not interchangeable. A topic is a broad problem area. A research subject is a title narrowed by population, context, and the relationship to be tested. A research question is the specific question that the data must answer.

For example:

A common confusion is using “research object” to mean the respondents. In quantitative research design, the research object may be intention to use, satisfaction, or behavior. Respondents are the units that provide data for measuring those concepts. You can review the article on the research object if these terms are getting mixed up in your proposal.

Another confusion is treating every new idea as a good topic. An idea may be new but lack a measurement scale, an accessible sample, or enough responses. For a thesis with a deadline, feasibility needs to be considered together with novelty, significance, and fit with your field of study.

Common Mistakes

The first mistake is choosing a topic that is too broad. Phrases such as “the impact of technology on society” or “improving service quality” do not tell you what you will measure or where you will measure it. Add the population, context, and outcome variable.

The second mistake is naming the research subject after the software. “Using SPSS to analyze...” is not a research topic. SPSS is data-processing software, while the topic must describe the phenomenon and relationship to be tested.

The third mistake is copying a model without checking the context. A scale used for business customers may not be suitable for students without modification. You need to review the definition, number of indicators, wording, and conditions of application.

The fourth mistake is promising the result in advance. Statements such as “X will definitely have a positive effect on Y” turn a hypothesis into a conclusion before you have data. Write “the study tests the effect of X on Y” and let the direction, magnitude, and p-value come from the output.

The final mistake is finalizing the research subject title before opening the questionnaire file. You should create a mapping table containing the research question, variable, variable code, scale, reference source, and analysis method. This table helps you detect gaps before distributing hundreds of responses.

Frequently asked questions

Is a scientific research topic the same as a research subject?

A topic is a broad problem area, while a research subject is a narrowed version that can be implemented as a specific study. The research subject usually states the population, context, and variables or relationship to be tested.

You can start with a topic and then narrow it into a research subject by adding the respondent group, location, time period, and dependent variable.

Where can you find ideas for a research subject?

You can find ideas in gaps in published studies, practical problems in your field, theoretical models you are studying, or a new context where data can be collected. When reading the literature, record the variables, scales, population, and limitations instead of only copying research subject titles.

An idea should be finalized only after you check whether you can access the sample and data. A topic that sounds attractive but has no suitable respondents will create problems during analysis.

How many variables does a quantitative scientific research topic need?

There is no fixed number of variables for every research subject. The number must come from the research question, model, and scale. Each construct is usually measured with multiple indicators, but you need to calculate the total number of indicators to estimate sample size according to (Hair et al., 2010).

If the model has too many independent variables, check whether each one has a clear theoretical role. A reasonable reduction in scope is usually better than putting every factor you can think of into one model.

Can you change the topic after collecting data?

You can adjust the title or scope if the actual data do not match the proposal, but a major change after collection will affect the questionnaire, sample, and hypotheses. Discuss the change with your supervisor before changing the dependent variable, population, or context.

Do not change the topic only because a p-value did not meet your expectation. A result that is not statistically significant can still be reported honestly and discussed in light of the limitations of the sample, scale, and model.

Does a scientific research topic need citations?

Yes. You need citations when you use a theory, model, definition, scale, or evaluation threshold. A title that you write yourself does not need a separate source, but the reasons for choosing the variables and the way you measure them should be explained with suitable literature.

In your proposal, save both the original source of the scale and the context in which it was used. This helps you answer the committee when asked why you selected those variables.

How to write this in your thesis

Open your proposal file and immediately create a table containing the topic, research question, dependent variable, independent variables, population, context, and scale source. Then compare it with the data you already have or prepare the questionnaire in M3 model, hypotheses, scales, and questionnaire.