Research design is a comprehensive plan that outlines how a research study will be conducted. It serves as a blueprint for the entire research process, from the formulation of the research question to the analysis and interpretation of the data. A well-designed research study is crucial for ensuring the validity, reliability, and generalizability of the findings.
Creating a research design involves making key choices regarding:
- The approach and research objectives
- The decision to use original data (primary research) or existing data (secondary research)
- The techniques for choosing participants or the standards for selecting subjects
- The tools and techniques employed to gather data
- The step-by-step process you’ll adhere to when collecting information
- The strategies you’ll use to examine and interpret the collected data
Steps for creating a research design
Creating a research design involves several steps, each of which requires careful consideration and planning.
Step 1: Consider your aims and approach
The first step in creating a research design is to clearly define your research question and objectives. This will help you determine the most appropriate approach for your study.
Research question example
“What factors contribute to employee turnover in the tech industry?”
The approach you choose will depend on the nature of your research question and the type of data you need to collect. There are two main approaches to research: qualitative and quantitative.
| Qualitative Approach | Quantitative Approach |
| Focuses on understanding the meaning and context of social phenomena | Focuses on measuring and quantifying variables and relationships |
| Uses non-numerical data, such as text, images, and observations | Uses numerical data and statistical analysis |
| Aims to explore and generate theories | Aims to test hypotheses and establish causal relationships |
| Emphasizes the researcher’s interpretation and subjectivity | Emphasizes objectivity and reproducibility |
Qualitative research designs
Qualitative research designs are usually more adaptable and employ an inductive approach. These designs allow researchers to modify their methods based on the findings that emerge during the research process.
Qualitative research example
An ethnographic study of the organizational culture in a tech startup, using participant observation and in-depth interviews.
Quantitative research designs
Quantitative research designs are typically more structured and follow a deductive approach. In these designs, the variables and hypotheses are clearly outlined before data collection begins.
Quantitative research example
A survey of employees in the tech industry, measuring job satisfaction, organizational commitment, and turnover intentions.
When designing your research, consider the following practical and ethical factors:
- The available time frame for data collection and completing the research write-up.
- Accessibility to the required data, which may involve traveling to specific locations or reaching out to particular individuals.
- Your proficiency in the necessary research skills, such as conducting statistical analyses or employing interviewing techniques.
- The potential need for obtaining ethical approval before proceeding with the study.
Step 2: Choose a type of research design
Once you have determined your research approach, you can choose a specific type of research design. The choice of design will depend on your research question, the type of data you plan to collect, and the resources available to you.
Types of quantitative research designs
Quantitative research designs aim to quantify and measure variables, test hypotheses, and establish relationships or causality among variables through the use of structured methods such as surveys, experiments, and statistical analysis.
| Type of Design | Purpose and Characteristics |
| Experimental | Tests causal relationships between variables Manipulates the independent variable and measures the effect on the dependent variable Uses random assignment and a control groupQuasi-experimentalSimilar to experimental design, but lacks random assignment Uses pre-existing groups or natural experiments |
| Descriptive | Describes the characteristics of a population or phenomenon Uses surveys, observations, or secondary data Does not aim to establish causal relationships |
| Correlational | Examines the relationship between two or more variables Uses statistical analysis to determine the strength and direction of the relationship Does not establish causality |
Correlational design example
A study examining the relationship between employee engagement and job performance, using surveys and performance evaluations.
Experimental design example
A study testing the effectiveness of a new training program on employee productivity, using random assignment and a control group.
Types of qualitative research designs
Qualitative research designs focus on understanding the subjective experiences, perceptions, and meanings that individuals attach to a particular phenomenon, typically through methods such as interviews, observations, and document analysis.
| Type of Design | Purpose and Characteristics |
| Case study | In-depth investigation of a single case or a small number of casesUses multiple sources of data, such as interviews, observations, and documentsAims to provide a rich, contextualized understanding of a phenomenon |
| Ethnography | Study of a specific cultural group or communityUses participant observation and interviews to understand the group’s beliefs, practices, and social interactionsRequires immersion in the field and a long-term commitment |
| Grounded theory | Develops a theory based on the analysis of qualitative dataUses a systematic, iterative process of data collection and analysisAims to generate a theory that is grounded in the data |
| Phenomenology | Explores the lived experiences of individualsUses in-depth interviews to understand how individuals perceive and make sense of a phenomenonFocuses on the subjective, first-person perspective |
Step 3: Identify your population and sampling method
Your research design should specify the focus of your study and the process for selecting participants or subjects. In research, a population refers to the entire group you aim to draw conclusions about, while a sample is the smaller subset of individuals from whom you’ll actually collect data.
Defining the population
A population can consist of various entities, such as people, organisms, organizations, texts, or countries. In social sciences, it frequently refers to a specific group of people. Clearly defining your population based on demographics, region, background, occupation, medical condition, or product usage will facilitate the process of obtaining a representative sample.
Population example
If you’re researching the impact of social media on body image, studying the entire global population of social media users would be incredibly challenging.
To streamline the research and enable more specific conclusions, you could narrow down your population to a more focused group—for instance, female Instagram users aged 18-25 in the United Kingdom.
Sampling methods
There are two main types of sampling methods: probability sampling and non-probability sampling. In qualitative research, sampling is often purposive, meaning that participants are selected based on their ability to provide rich, in-depth information about the phenomenon being studied.
| Probability Sampling | Non-Probability Sampling |
| Every member of the population has an equal chance of being selected | Some members of the population are more likely to be selected than others |
| Allows for generalization to the larger population | Does not allow for generalization to the larger population |
| More time-consuming and expensive | Less time-consuming and expensive |
| Examples: simple random sampling, stratified sampling, cluster sampling | Examples: convenience sampling, purposive sampling, snowball sampling |
Case selection in qualitative research
In certain qualitative research designs, such as ethnographies or case studies, sampling may not be applicable.
The goal in these designs is to gain a deep understanding of a specific context rather than to generalize findings to a larger population. Instead of sampling, the focus is on collecting as much relevant data as possible about the context under investigation.
When using these designs, it is crucial to provide a clear justification for selecting a particular case or community, demonstrating how it is well-suited to address the research question at hand.
This may involve choosing a case study that sheds light on an overlooked or unique aspect of the research problem or selecting multiple cases that are either highly similar or distinctly different for comparative purposes.
Step 4: Choose your data collection methods
Data collection methods involve directly measuring variables and gathering information, enabling you to acquire first-hand knowledge and original insights into your research problem. You can opt for a single data collection method or employ multiple methods within the same study.
Survey methods
Surveys enable you to collect data on opinions, behaviors, experiences, and characteristics by directly asking participants. There are two primary survey methods: questionnaires and interviews.
| Questionnaires | Interviews |
| Standardized set of questions administered to a large sample | In-depth, open-ended questions administered to a smaller sample |
| Can be self-administered or researcher-administered | Usually conducted face-to-face or over the phone |
| Allows for quantitative analysis | Allows for qualitative analysis |
Observation methods
Observational studies allow for unobtrusive data collection by observing characteristics, behaviors, or social interactions without relying on self-reporting. Observations can be conducted in real-time, with the researcher taking notes, or by making audiovisual recordings for later analysis. These observations can be either qualitative or quantitative in nature.
| Quantitative Observation | Qualitative Observation |
| Structured, systematic observation of behavior | Unstructured, naturalistic observation of behavior |
| Uses predetermined categories and coding schemes | Uses field notes and descriptive accounts |
| Focuses on frequency and duration of behaviors | Focuses on the meaning and context of behaviors |
Other methods of data collection
Other methods of data collection include:
- Experiments: Researchers manipulate one or more variables to measure their effect on a dependent variable, often using control and experimental groups.
- Focus groups: Small, guided discussions with participants to gather qualitative data on opinions, attitudes, and experiences related to a specific topic.
- Document analysis: Systematic review and evaluation of existing documents, such as official records, personal journals, or media articles, to extract relevant data.
- Archival research: Collecting and analyzing historical records, documents, or artifacts to gain insights into past events, trends, or phenomena.
These methods can be used independently or in combination with other data collection techniques, depending on the research question and study design.
Secondary data
If time or resources are limited, you can utilize secondary data previously collected by other researchers, such as datasets from government surveys or prior studies related to your topic. This helps you explore new research questions that were not addressed in the original study.
Using secondary data can broaden the scope of your research by providing access to larger and more diverse samples than you might be able to collect on your own. However, this approach also means that you have no control over the variables measured or the measurement methods used, which may limit the conclusions you can draw from the data.
Step 5: Plan your data collection procedures
Once you have chosen your data collection methods, you need to plan the specific procedures for collecting the data. This involves:
Operationalization
Defining your variables and measures in concrete, observable terms.
If you’re using observations to study employee engagement in a retail store, you could record the number of times employees initiate conversations with customers, offer assistance, or go above and beyond their job duties.
Example: To measure employee engagement, you could create a checklist of specific behaviors, such as greeting customers, offering product recommendations, or handling customer complaints efficiently. Observers would then tally the number of times each behavior occurs during a set period.
If you’re using surveys to evaluate customer satisfaction with a new product, you could design questions that ask about specific features, ease of use, and overall experience.
Example: To measure customer satisfaction, you could create a survey with questions like, “On a scale of 1 to 10, how likely are you to recommend this product to a friend?” or “How well did the product meet your expectations?” with response options ranging from “Did not meet expectations at all” to “Exceeded expectations.”
Reliability and validity
Reliability refers to the consistency of your measurement. If your study is reliable, it means that other researchers could replicate your methods and obtain similar results. In other words, your findings are not just a one-time occurrence but can be consistently reproduced.
Validity, on the other hand, is concerned with the accuracy of your measurement. It refers to the extent to which your research design and data collection methods actually measure the concept or variable you’re interested in studying. A valid study is one that accurately captures the real-world phenomenon it aims to investigate.
| Reliability | Validity |
| Consistency of the measures | Accuracy of the measures |
| Can be assessed through test-retest, inter-rater, or internal consistency | Can be assessed through face validity, content validity, criterion validity, or construct validity |
Sampling procedures
In addition to selecting a suitable sampling method, you must develop a specific plan for contacting and recruiting your chosen sample. This involves making decisions about:
- The number of participants needed for an adequate sample size
- Inclusion and exclusion criteria for identifying eligible participants
- The method of contacting your sample (mail, online, phone, or in person)
When using probability sampling, it’s crucial to ensure that all randomly selected individuals participate in the study. Consider strategies to achieve a high response rate.
If using non-probability sampling, take steps to minimize research bias and ensure a representative sample.
Data management
Create a data management plan for organizing and storing your data. This may involve transcribing interviews or entering observation data. Anonymize and protect sensitive data, and ensure regular backups.
Keeping your data well-organized will save time during analysis and enable other researchers to validate and build upon your findings, enhancing replicability.
Step 6: Decide on your data analysis strategies
The final step in creating a research design is to plan your data analysis strategies. The choice of analysis methods will depend on the type of data you have collected and your research question.
Quantitative data analysis
In quantitative research, statistical analysis is commonly used to summarize sample data, make estimates, and test hypotheses.
Descriptive statistics allow you to summarize your sample data in terms of:
- Distribution: the frequency of each score on a test
- Central tendency: the mean to describe the average score
- Variability: the standard deviation to describe the spread of scores
The level of measurement of your variables determines the specific calculations you can perform.
Inferential statistics enable you to:
- Make estimates about the population based on your sample data
- Test hypotheses about relationships between variables
Regression and correlation tests assess associations between variables, while comparison tests (e.g., t-tests and ANOVAs) examine differences in outcomes between groups.
Your research design, including variable types and data distribution, guides your choice of statistical tests.
Qualitative data analysis
In qualitative research, data is often dense with information and ideas. Rather than summarizing it numerically, you must thoroughly examine the data, interpret its meanings, identify patterns, and extract the most relevant parts to address your research question.
Common approaches include thematic analysis and discourse analysis.
| Thematic Analysis | Discourse Analysis |
| Identifies patterns and themes in the data | Examines how language is used to construct social reality |
| Focuses on the content of the data | Focuses on the form and function of the data |
| Can be inductive (data-driven) or deductive (theory-driven) | Often uses a critical lens to examine power relations and ideologies |