Research: Approach, Method, and Design

Last Updated on: 3rd February 2024, 01:30 am

One of the most confusing things about academic research is the inconsistency with terms. This is especially challenging with the concepts of research approach, method, and design.

Often, the terms are used differently by different users. The definitions vary with the source. 

The purpose of this article is to provide some semblance of standardization and consistency. I have consulted four sources commonly used in graduate research (see Reference List):

         Cooper, D. R., & Schindler, P. S. (2013).¹ 

         Creswell, J. W., & Creswell, J. D. (2022).²

         Frankfort-Nachmias, C., Nachmias, D., & DeWaard, J. (2014).³ 

         Trochim, W. (2006).⁴

The taxonomy and definitions are adapted and synthesized from these sources. 

I will focus mostly on quantitative research. But, I will provide a complete list of all research methods and encourage you to consult the references at the end for more detail. 

The Purpose Statement

Let’s start with the structure of a typical dissertation purpose statement. I’ll refer to this schematic as we proceed through the discussion.

The purpose of this is to study, employing the central phenomenon being studied (research problem).

Research Approach

Qualitative research explores the meaning that people attribute to social or human problems. The data are recorded as non-numerical units—words and phrases from interviews, questionnaires, observation, and focus groups. Research and analysis yield deep insights from the direct experience with a phenomenon.

Quantitative research deals with numerical data. We seek patterns, trends, comparisons, and correlations among the variables. We test theories, expressed as hypotheses—the postulated relationships among numerical variables.

Mixed methods research combines quantitative and qualitative approaches. Qualitive insights combined with statistical analysis yields a more complete understanding of the phenomenon under examination.

Research Method

Research methods include experimental research, quasi-experimental research, and non-experimental research. In most cases, experimental and quasi-experimental research methods are quantitative. However, some of the attributes of experimental methods apply to qualitative research:

• Random sampling
• Control over factors such as demographics
• Assignment to groups including treatment and control groups

In general, quantitative research may use any of the research methods. Qualitative research is generally non-experimental.

Over 50% of doctoral candidates don’t finish their dissertations.

A game-changing training tailored to give you a clear path to gaining your committee’s approval to graduate.

BOOK YOUR FREE CONSULTATION

Over 300+ Students Coached • 40+ Years of Experience • 90% Success Rate

Experimental

Montgomery defined an experiment as,

a test or series of tests in which purposeful changes are made to the input variables of a process or system so that we may observe and identify the reasons for changes that may be observed in the output response.⁵

I mention experimentation here as a research method because its attributes can be applied across a variety of research approaches, designs, scopes, and data collection techniques.

Experimentation is also a data collection mechanism, but it is far more than just data collection. The essential concept in experimentation is control over the input factors in the study.

Quasi-Experimental

A quasi-experimental design exhibits many of the attributes of an experimental design, but not all of them. In a quasi-experimental design, the requirement for random selection of cases is relaxed. For example, the sample from a target population may be obtained through convenience or snowball sampling. And, the sample may be stratified. 

Nonexperimental

In nonexperimental designs, the experimenter simply describes groups or cases (combinations of attributes or independent variables) as they exist. There is no control or manipulation of independent variables. Members of the groups or replications of cases are not randomly chosen.⁶ Conclusions about causal relationships cannot be drawn with the same validity as experimental designs.

Nonexperimental designs can be used in quantitative, comparative or correlational designs. Frequently, in observation, survey, and ex post facto studies, the values for both independent and dependent variables are random. We use regression and ANOVA to make comparisons between groups, and examine relationships among variables.

Research Design

Descriptive

These studies are performed to describe a phenomenon, system, activity, behavior, performance, situation, state, or condition. There is no intent to compare the findings to a standard or other phenomena. Or to understand the causes, influences, or predictors.

For example, the purpose might be to describe the demographics of a college:

• Counts of various demographic groups (e.g., race, gender)
• Mean and standard deviation of age
• Mean and standard deviation of ACT scores

Comparative

The objective is to compare one group to another (e.g., the mean for a variable). Or, to compare an attribute of a group (the mean) against a parameter.

Comparative designs employ hypothesis tests, such as t tests. Most of the time, comparative designs are intended to infer attributes of the population from a sample of that population.

For example:

• Is the mean ACT score the same between men and women at a college?
• Is the mean ACT score of entering freshmen greater than the national average (19.8)?

Relational or Correlational

Correlational studies examine the relationships among variables. More specifically, the correlations, associations, and relationships between independent variables or factors; and dependent variables. The output may be a predictive, mathematical model. In quantitative, correlational studies, the tools are often correlation, regression, and ANOVA.

A correlational relationship indicates that two entities occur in a seemingly synchronized or simultaneous way. 

Continuing with our example:

• What is the relationship between the independent variables of race, gender, age, ACT score) and a dependent variable (GPA at the end of the year for freshmen)?

Causal

Correlation is not sufficient to conclude cause-and-effect. Instead, causal studies are planned to determine if one or more factors or variables affect or cause an outcome of a phenomenon. 

Using the same question posed earlier for a correlational study, with slight modification:

• Are race, gender, and age direct causes or influences of end of year GPA? 

See my article on Causation vs. Correlation.

Qualitative Research Designs

Here are the most common qualitative research designs. I encourage you to consult the sources I have provided, which provide a detailed description of these designs:

• Narrative research
• Ethnography
• Phenomenology
• Field research
• Grounded theory
• Case study

Cumulative Nature of Research Designs

The four quantitative research designs are often cumulative. For all studies, there would likely be some descriptive statistics. A correlational design would begin with descriptive statistics, then investigate the relationships among variables. A causal design would take hypothesis testing one step further, controlling variables in an experiment, providing evidence of direct links between explanatory variables and responses, and employing subject matter expertise to explain the underlying causality. 

Time Dimension

We can describe research in terms of scope, both temporally and topically.

Regarding the time dimension, cross-sectional studies take place at a point in time. In contrast, longitudinal studies transpire over time with multiple measurements of variables.

In our example, if we examined the end of year GPA for freshmen at a college, one time (at the end of the 2023 academic year), that would be a cross-sectional study. On the other hand, if we examined this phenomenon for successive freshman classes over a period of 10 years, it would be a longitudinal study. Likewise, if we examined the GPA of the same class of students for each of their 4 years at the college, that would be longitudinal.

Regarding the topical dimension . . . . Many studies cover a spectrum of groups, attributes, activities, or events across a defined population. In contrast, a case study investigates a single situation, event, activity, process, system, individual, or organization. The intent is to focus on this one case over a limited and specific period of time. Then, let the findings speak for themselves.

Case studies are often associated with qualitative research. However, it is acceptable to employ quantitative methods in a case study. 

In our example, let’s say we are looking at the entering group of scholarship football players. We delve into their demographics, using statistics. We also explore their individual stories, families, and high school experiences. We make a deep dive into their backgrounds, experiences, and perspectives along with demographics, to more clearly understand their academic performance as freshmen.

Data Collection Techniques

There are many forms of data collection, and many kinds of instruments. See my article, Data Collection.

When multiple data collection techniques are employed, triangulation is used. First, to check for validity. But, also to produce synthesized, synergistic results. This is especially evident when qualitative analysis is used to explain statistical analysis.

Observation & Monitoring

Observation enables researchers to directly study, and collect data about, a behavior, performance, or activity. The observation normally takes place in natural settings, and often avoids researcher bias. The research does not rely on self-reporting by the participants. Data may be either qualitative or numerical. Data collection can be casual or highly structured, as in a laboratory. 

Examples include observing drivers and their behavior in traffic; children at play; or naturally occurring biological or physical phenomena.

Ex Post Facto, Secondary Data Collection & Analysis

Often, relevant data have already been collected by other researchers or reliable agencies. Assuming the data are reliable, and meet the scope, criteria, and delimitations of the study, this technique can be very useful and efficient. Often the data are from government, corporate, or medical sources; and often online.

Two examples from dissertations which I supervised: One compared the funds raised using either crowd sourcing or crypto currency, moderated by a number of demographic factors. Another explored the relationship between measures of country culture and demographics, and restaurant online ratings. In both, the data were available from several online sources.

Document Analysis

Document analysis is nothing more than deriving qualitative data from reliable government, organizational, or academic documents. As with interviews, questionnaires, and focus groups, a coding process is used to identify key words and phrases, codes, categories, and themes. 

Surveys & Questionnaires

There is a distinction between the concepts of survey and questionnaire. A questionnaire is a component of a survey, which includes the sampling mechanisms and procedures. For more information on the survey method, see my article titled, Survey vs. Questionnaire.

Interviews & Focus Groups

Interviews are either in person or via a video platform. The researcher asks participants highly structured, semi-structured, or open-ended questions depending on the research objectives. 

In focus groups, individuals are assembled, then prompted with questions by a moderator, or merely given a topic. The participants interact freely with each other and the moderator. The researcher may record the discussions. The researcher observes and records the spoken words, the body language, emotions, interactions, and other subtle behaviors.

Experimentation

I mention experimentation again—here as a data collection mechanism. But it is far more involved than just data collection. It is a holistic strategy for planning and conducting research and analyzing the data.⁵

Let’s say we are interested in an academic intervention—for example, a special, short effective writing class for a week at the beginning of the school year. We sample from the target population, carefully controlling several factors, and randomly sampling to replicate each case to achieve a minimum required sample size:

• Gender
• Race
• Age
• ACT score
• Treatment group or control group (no class)

Then, we administer a writing pre-test and a writing post-test. We accomplish two major objectives:

• Evaluating the effectiveness of the effective writing class.
• Examining the relationship between our independent variables and test performance.

Modeling & Simulation

Modeling & simulation is the development and use of models of real systems and phenomena (often digital, computer models). Models are used to simulate behavior or performance over time, generating data for analysis in practical or academic research.⁷

Models can be used to replicate the behavior of physical systems. But, also to model social behavior. For example, there are models that simulate behavior within a stock exchange, in response to various inputs and stimuli. In simulations, we can measure quantitative outcomes record qualitative behaviors.

Purpose

Your purpose is captured in a single action verb that conveys how learning is intended to take place in your research and analysis. The choice is very important. 

The words convey, succinctly, what your intent is within the context of approach, method, design, time dimension, data collection technique, and research problem—the means to accomplish your purpose. Your purpose comes down to one of these simple verbs (or perhaps others, as you deem appropriate):

• explore
• examine
• describe
• compare
• develop
• understand
• discover

Meeting Every Requirement, Yet Facing Rejection?

A game-changing training tailored to give you a clear path to gaining your committee’s approval, so you can graduate with your hard-earned doctorate.

BOOK YOUR FREE CONSULTATION

Over 300+ Students Coached • 40+ Years of Experience • 90% Success Rate

Putting It All Together

There is overlap among the approaches, methods, and designs. They’re not mutually exclusive. Some approaches may also be methods or designs. Some research may use parts of more than one approach, method, design, data collection, and purpose.

The key is to be consistent, and to define your terms.

Here are a couple of purpose statements that combine various research attributes:

The purpose of this quantitative, nonexperimental, correlational, cross-sectional study, employing a survey, is to examine the relationship between measures of employee empowerment and measures of job satisfaction, as moderated by demographic factors.

The purpose of this qualitative, case study; employing interviews, document analysis, and focus groups; is to explore the lived experiences of participants in a one-year program to increase literacy within a county in South Carolina.

Summary

Unfortunately, scholars use multiple definitions of research terms, and ways to categorize research. I covered these using the purpose statement for structure. The building blocks include research approach, method, design, time dimension, data collection technique, and purpose verb. Together, these are the essentials for creating original research.

Final Thoughts

There is no getting around the many definitions for the same terms. That said, what is important is developing your own understanding of the components and descriptors of research. Be conversant with the categories, the terms, and the definitions. Be capable of discussing and explaining them. But, be flexible enough to accommodate others’ strongly held beliefs about them—especially decision-makers, other researchers, and dissertation committee members.

References

1. Cooper, D. R., & Schindler, P. S. (2013). Business research methods (12^th ed.). McGraw-Hill.

2. Creswell, J. W., & Creswell, J. D. (2022). Research design: Qualitative, quantitative, and mixed methods approaches (6^th ed.). Sage.

3. Frankfort-Nachmias, C., Nachmias, D., & DeWaard, J. (2014). Research methods in the social sciences (8^th ed.). Worth Publishers.

4. Trochim, W. (2006). The research methods knowledge base. Atomic Dog Publishing.

5. Montgomery, D. C. (2020). Design and analysis of experiments (10^th ed.). Wiley.

6. Lobmeier, J. (Ed.) (2010). Nonexperimental designs. In Encyclopedia of research design. SAGE Publications. https://doi.org/10.4135/9781412961288

7. McAllister, B. (2023). Simulation-based performance assessment. Kindle.