RESEARCH DESIGNS/STUDY DESIGN
Study or Research design defines the approaches, methods and the rationale of picking that appropriate research design
- Eg descriptive cross sectional design
- Approaches can be Quantitative/qualitative or both
- Note that it is advisable to use one of these at our level.
Research design acts as a blueprint for conducting a research study, outlining how variables will be operationalized for measurement, the selection of the sample of interest, data collection methods, and the intended means of data analysis.
Zikmund (1988) defines research design as a master plan that specifies the methods and procedures for measuring, collecting, and analyzing data.
At the core of a research design are answers to crucial questions:
- How will the study be conducted?
- What procedures will be adopted to obtain answers to research questions?
- What kind of data needs to be collected?
- How will the tasks required to complete the various research components be carried out?
🏗 Importance of Research Design 📊
- Foundation for Research: 🏠 Acts as the solid base upon which the entire
research stands. Acts as a firm foundation for the research.
- Smooth Research Operations: ⚙️ Ensures all research activities run smoothly and efficiently. Because you know what next!
- Efficiency Maximization: 💪Provides maximum information with minimal effort, time, and cost. It makes research as efficient as possible by giving maximum information with minimum expenditure of effort, time and energy.
- Blueprint for Research: 📐 Just as an architect needs a blueprint for building a house, research needs a proper design for conducting a study.
- Simplifies work: 😍This is by ensuring that limitations are predetermined and solutions are already at hand, so you overcome them.
A well-planned research design is like a strong foundation for your study,
making the research process efficient and effective. 🌟🧱📈
🧐 FACTORS THAT INFLUENCE CHOOSING A RESEARCH DESIGN 📝
Researcher’s Knowledge 🧠: The researcher’s familiarity with a particular design.
- Example: If a researcher is well-versed in qualitative research methods, they may choose to conduct an ethnographic study to gain an in-depth understanding of a specific community.
Resource Availability ⏳💼: Availability of time, human resources, and willing respondents.
- Example: In a time-sensitive study, a researcher might opt for a cross sectional design due to its efficiency in data collection and analysis.
Ethical Considerations 🤝📜: Ethical aspects, including the ethical treatment of respondents.
- Example: In a study involving vulnerable populations, such as children, ethical considerations may lead the researcher to choose a design that prioritizes the protection of participants, like an experimental design.
Feasibility and Relevance 🎯📊: The practicality and relevance of the design to the study.
- Example: A large-scale public health survey may require a design that is both feasible and relevant, such as a cross-sectional study that provides a snapshot of health trends in a population.
Geographical Scope 🌍🗺: The extent of the geographical area to be covered.
- Example: A study investigating regional variations in climate change impacts might choose a design that covers multiple countries and regions to capture a broad geographical scope, such as a comparative case study.
Equipment Availability 🛠💻: Access to necessary research equipment and tools.
- Example: Research requiring advanced scientific equipment, like electron microscopes, would naturally be influenced to adopt experimental research designs.
Research Type 🔍📈: The specific type of research, e.g., cross-sectional or longitudinal.
- Example: If a researcher aims to understand changes over time, they may select a longitudinal design to follow the same group of participants at multiple points in time.
Control 🎮🔒: The level of control the researcher can maintain over the study.
- Example: In a medical study testing a new drug’s effectiveness, a randomized controlled trial (RCT) design would provide a high level of control over the research conditions and variables.
Population Type 👥🌐: The characteristics of the population under study.
- Example: Research on consumer preferences might employ a certain design to ensure the population’s characteristics are accurately represented in the study.
Types of research designs
There are three main overall methodologies (or designs): Qualitative , quantitative and Mixed.
Qualitative Research Designs:
- Qualitative research designs are primarily exploratory and descriptive in nature.
- They aim to understand and interpret the underlying meaning of phenomena.
- Qualitative studies often involve collecting non-numerical data such as text, images, or observations.
- Common qualitative research designs include phenomenology, ethnography, grounded theory, and case studies.
Quantitative Research Designs:
- Quantitative research designs focus on collecting and analyzing numerical data.
- They are structured, objective, and use statistics to establish relationships between variables.
- The primary goal of quantitative research is to measure, describe, and explain phenomena through numerical data.
- Common quantitative research designs include descriptive, analytical, experimental, and correlational studies.
Mixed Research Designs:
- Mixed research designs combine elements of both qualitative and quantitative research within a single study.
- These designs aim to provide a comprehensive understanding of complex research questions.
- Researchers typically collect and analyze both numerical and non-numerical data, often in two distinct phases, with one method informing the other.
- The choice of mixed design depends on the research objectives and the need for a holistic approach to the research question.
|Quasi & Comparative
|Randomized Controlled Trial
Qualitative vs. Quantitative Research Designs
Qualitative research deals with phenomena that are difficult or impossible to quantify mathematically, such as beliefs, meanings, attributes, and symbols.
Quantitative research deals with phenomena that are possible to be quantified mathematically. For example, experimental studies are quantitative in nature.
|Nature of Data
|Non-numerical (text, words)
|Numerical (numbers, measurements)
|Number of Respondents
|Larger number of respondents
|Formulation of hypotheses
|Testing formulated hypotheses
|Limited knowledge on the topic
|More knowledgeable about the topic
|Nature of Data Collection
|Subjective (relies on respondents)
|Objective (researcher gathers data)
|Conclusiveness of Results
|Inductive (results inconclusive)
|Deductive (results are conclusive)
|General and broad questions
|Specific and narrow questions
|Bias in Research
|Less biased (researcher in control)
Experimental Study Design:
Experimental study design involves introducing an intervention assumed to be the ’cause’ of change and waiting until it has produced or has been given sufficient time to create that change. Some essential experimental study designs include the following:
i) After-only design: In this approach, the researcher knows that the population has been exposed to an intervention. They aim to study its impact on the population. Information about the baseline is usually gathered from respondents’ recall of the situation before the intervention or from existing records. This design is commonly used in impact assessment studies.
ii) Before-and-after design: This design addresses the issue of retrospectively constructing the ‘before’ observation by establishing it before introducing the intervention to the study population. Subsequently, when the program is fully implemented or assumed to have its effect on the population, an ‘after’ observation is conducted to determine the intervention’s impact.
iii) Control group design: In the control group design, the researcher selects two population groups – a control group and an experimental group. These groups are expected to be as comparable as possible in every respect except for the intervention. The experimental group receives or is exposed to the intervention, while the control group is not. Initially, the ‘before’ observation is made on both groups simultaneously. Subsequently, the experimental group is exposed to the intervention. When it is assumed that the intervention has had an impact, an ‘after’ observation is conducted on both groups. Any difference in the ‘before’ observation between the groups concerning the dependent variable(s) is attributed to the intervention.
General Characteristics of Experimental Designs:
- Direct manipulation of independent variables in a controlled environment.
- Examination of the effect of independent variable(s) on one or more dependent variables.
- Focus on testing hypotheses.
- Total control of extraneous variables.
- Generation of empirical findings with internal and external validity.
- Some experimental designs use control and experimental groups (pure experimental design).
Advantages of Experimental Design:
- i) It enables the researcher to control the situation, helping answer the question, “What causes something to occur?”
- ii) It allows identification of cause-and-effect relationships between variables and distinguishes placebo effects from treatment effects.
- iii) Experimental research designs support the ability to limit alternative explanations and infer direct causal relationships in the study.
- iv) This approach provides the highest level of evidence for single studies.
Disadvantages of Experimental Design:
- i) Experimental design can be artificial, and results may not generalize well to the real world.
- ii) The artificial settings of experiments may alter participants’ behaviors or responses.
- iii) Experimental designs can be costly if they require special equipment or facilities.
- iv) Some research problems cannot be studied using experiments due to ethical or technical reasons.
- v) It can be challenging to apply ethnographic and other qualitative methods to experimentally designed studies.
Cross-Sectional Study Design
Cross-sectional study design is a research approach that involves different groups of people who vary in the variable of interest but share other characteristics, such as socioeconomic status, educational background, and ethnicity. For instance, researchers may select groups of people who are quite similar in most aspects but differ only in age. This allows any observed differences between groups to be attributed to age rather than other variables.
Cross-sectional studies are typically observational and are commonly used in descriptive research.
Characteristics of Cross-Sectional Studies:
- Researchers record existing information within a population without manipulating variables.
- Data collection occurs at a single point in time.
- This research design describes characteristics within a population but does not establish cause-and-effect relationships between different variables.
- It enables researchers to investigate multiple aspects simultaneously, such as age, income, and gender.
What Cross-Sectional Studies Can Reveal:
- Cross-sectional studies provide a snapshot of outcomes and associated characteristics at a specific point in time.
- Unlike experimental designs that involve active intervention to induce and measure change or create differences, cross-sectional studies focus on studying and drawing inferences from existing differences among people, subjects, or phenomena.
- Data is collected at a particular moment in time.
- While longitudinal studies require multiple measurements over an extended period, cross-sectional research aims to identify relationships between variables at a single moment.
- Groups chosen for study are deliberately selected based on existing differences in the sample rather than random sampling.
- This method can estimate the prevalence of an outcome of interest because the sample typically represents the entire population.
- Cross-sectional designs often use survey techniques for data collection, making them relatively cost-effective and efficient.
Disadvantages and Challenges of Cross-Sectional Design:
- Finding individuals, subjects, or phenomena that are very similar except for a specific variable can be challenging.
- Results are static and time-bound, offering no insights into the sequence of events or historical contexts.
- Cross-sectional studies cannot establish cause-and-effect relationships.
- This design provides only a single snapshot of analysis, which may yield different results if conducted at another time.
- There is no follow-up to the findings.
Characteristics of Cross-Sectional Research Design:
- No time dimension.
- Focuses on existing differences rather than changes following intervention.
- Groups are selected based on existing differences rather than random allocation.
Note: Cross-sectional research design differs from longitudinal research, which involves taking multiple measurements over an extended period of time.
Longitudinal Research Design:
This is a study design that involves a series of sample measurements taken over a period of time. The study focuses on specific respondents or a sample over a specified period, examining changes in behavior.
- A study of a specific group of children’s academic performance over time.
- A study of the stages of human development by following a child from birth to adulthood.
Longitudinal studies provide deeper insights into issues. However, they can be expensive in terms of both time and money. Such studies may face setbacks if a respondent dies or transfers.
Longitudinal studies are further subdivided into Panel and Trend studies.
Panel Study: Involves using the same group or individual throughout the study period. One sample is studied continuously.
Trend Study: Involves different homogenous groups or samples in phases over the study. The results obtained from the study are then analyzed and compared to investigate trends of change. For example, one can observe the trend of change in the behavior of 1st-year students admitted in 2001 and 1st-year students admitted in 2002 using a trend longitudinal study.
What Longitudinal Studies Tell You:
- Longitudinal data help analyze the duration of a specific phenomenon.
- Enables researchers to approach causal explanations usually achievable only through experiments.
- Permits the measurement of differences or changes in a variable from one period to another, describing patterns of change over time.
- Facilitates predicting future outcomes based on earlier factors.
What Longitudinal Studies Don’t Tell You:
- The data collection method may change over time.
- Maintaining the integrity of the original sample can be difficult over an extended period.
- It can be challenging to show more than one variable at a time.
- This design often requires qualitative research data to explain fluctuations in the results.
- A longitudinal research design assumes that present trends will continue unchanged.
- It can take a long time to gather results.
- A large sample size and accurate sampling are needed to reach representativeness.
Case Study/Case Report Designs:
A case study is an in-depth examination of the behavior of a specific phenomenon, such as a person, a group of people, an object, or a situation within a limited environment. The findings from a case study can be generalized to represent other cases within a population of interest. For instance, one can study the behavior of a child or a group of children from birth to adulthood and then generalize the results to other children.
What Case Studies Don’t Tell You (Disadvantages):
- A single or a small number of cases offer limited grounds for establishing reliability or for generalizing the findings to a broader population of people, places, or things.
- Intensive exposure to the study of a case can bias a researcher’s interpretation of the findings.
- This design does not facilitate the assessment of cause-and-effect relationships.
- Vital information may be missing, making the case challenging to interpret.
- The case may not be representative or typical of the larger problem under investigation.
- If a case is selected because it represents a very unusual or unique phenomenon or problem for study, then the interpretation of the findings can only apply to that specific case.
Retrospective and Prospective Study:
A retrospective study is a longitudinal study that examines data from the past. For example, a researcher may review medical records from previous years to identify trends. In essence, retrospective studies “look back” in time. On the other hand, prospective studies “look forward” by collecting data as events unfold.
Retrospective Study Example in Health: A retrospective study in health might involve examining historical medical records of cancer patients to determine the effectiveness of a particular treatment protocol used in the past. Researchers can analyze the outcomes of these patients based on the treatment they received and other variables to gain insights into the treatment’s success rates.
Prospective Study Example in Health: A prospective study in health could involve tracking a group of pregnant women from early pregnancy through childbirth and beyond. Researchers would collect data on factors like prenatal care, diet, and lifestyle, and follow the women to monitor pregnancy outcomes, birth complications, and the health of their babies after birth. This type of study helps in understanding the factors that influence maternal and child health during and after pregnancy.
Cohort Study Design:
A cohort is a group of individuals who share a common characteristic or experience. For instance, a group of people born on the same day or during a specific period, like 1981, can form a birth cohort. The cohort design is a type of observational study in which one or more groups (cohorts) are followed over time, and subsequent assessments are made to determine the association between the initial characteristics or risk factors of the participants and specific outcomes or diseases. As the study progresses, the outcomes of participants in each cohort are measured, and researchers analyze the relationships with particular characteristics.
Example of a Cohort Study:
To investigate whether tobacco exposure is associated with the development of lung cancer, an investigator forms two groups (cohorts) of adolescents. One group consists of individuals who have never smoked tobacco and continue to refrain from smoking (unexposed), while the other group comprises tobacco smokers (exposed). The investigator then follows both groups for a specified period and observes how many individuals in each cohort develop the disease and how many do not. Below is a table illustrating the outcomes:
|Do Not Develop Disease
|Incidence of Disease
|Do not smoke tobacco
Advantages of Cohort Studies:
- Subjects in cohorts can be matched to limit the influence of confounding variables.
- Cohort studies can demonstrate that potential causes precede the outcomes.
- Original or secondary data can be used in this design.
- Cohort studies are often less expensive and easier to conduct than randomized controlled trials.
Disadvantages of Cohort Studies:
- Identifying cohorts can be challenging due to confounding variables.
- Lack of randomization may lead to imbalances in patient characteristics.
- Blinding or masking is difficult in cohort studies.
- Outcomes of interest may take time to occur.
Additionally, they might take a long time to complete, potentially impacting the validity of findings. The lack of randomization in cohort studies also reduces their external validity compared to randomized studies.
Randomized Controlled Trial (RCT):
The primary difference between a Randomized Controlled Trial (RCT) and a cohort study is the random selection of samples. RCTs involve the random allocation of participants to different clinical interventions, including a control group. These trials seek to measure and compare the outcomes following the interventions. RCTs are purely experimental and quantitative in nature.
Example: In a study to determine whether painkillers are necessary for males who have just undergone circumcision, 200 eligible men were randomly assigned to two groups. One group received Panadol tablets as pain relief immediately after surgery, while the other group received a placebo. The results showed that a majority (90%) of patients given Panadol reported no pain, whereas over 90% of patients on the placebo reported significant pain.
Advantages of RCT:
- RCTs are considered the most reliable form of scientific evidence.
- They reduce spurious causality.
- RCTs influence healthcare policy and practice.
Limitations of RCT:
- The external validity of RCT results may be limited.
- Ethical concerns may arise in some situations.
- The time required to observe outcomes can be extensive.
Case Series/Clinical Series:
A case series is a descriptive study that follows a group of patients or subjects with known exposure, such as individuals who have received similar treatments or whose medical records are examined for exposure and outcomes. Case series can generate hypotheses for further studies but cannot establish causal relationships.
Internal validity in case series studies is generally low because they lack a comparator group exposed to the same intervening variables.
Correlation Study Design:
A correlation study investigates whether two variables are correlated, meaning an increase or decrease in one variable corresponds to a change in the other. However, correlation does not imply causation. Three types of correlation exist: positive, negative, and no correlation.
A correlation coefficient, which ranges from +1 to -1, is typically used to quantify the strength and direction of the correlation between variables.
There are three primary types of correlation:
Positive Correlation: This type of correlation occurs when an increase in one variable is associated with an increase in another, and a decrease in one variable is linked to a decrease in the other. For example, an individual’s wealth may exhibit a positive correlation with the number of rental houses they own. This suggests that as wealth increases, so does property ownership.
Negative Correlation: Negative correlation is observed when an increase in one variable is connected to a decrease in another, and vice versa. For instance, there might be a negative correlation between the level of education in a country and its crime rate. If education levels improve in a country, the crime rate tends to decrease. However, it’s essential to note that this does not mean that a lack of education directly causes crime. Both factors might share a common underlying cause, such as poverty.
No Correlation (Uncorrelated): In cases of no correlation, changes in one variable do not correspond to changes in the other, and vice versa. For example, among millionaires, there may be no correlation between their level of wealth and their level of happiness. This indicates that an increase in wealth does not necessarily lead to an increase in happiness.
Comparative Study Design:
Comparative studies involve examining two or more cases, specimens, or events that share similarities in some aspects but differ in others. The goal is to determine the reasons for these differences and to generalize findings to larger groups from which the cases originated. The plausibility of generalization increases when several cases from the same group are considered.
Ethnographical Research Design:
Ethnographical research, often referred to as ethnography, involves the in-depth study of naturally occurring behavior within a particular culture or social group. Its primary aim is to understand the relationship between culture and behavior. Culture in this context refers to the beliefs, values, and attitudes of a specific group of people. Ethnographic research methods were developed by anthropologists to study and describe human cultures.
Characteristics of good ethnography, as described by Spindler & Hammond (2000), include:
Extended Participant Observation: Researchers immerse themselves in the culture or social group they are studying and observe behavior over an extended period.
Long Time at Site: Researchers spend a significant amount of time within the community or culture being studied to gain a deep understanding of their way of life.
Collection of Large Volume of Materials: This includes detailed notes, audio recordings, video tapes, and other data, often without having specific hypotheses or predefined categories at the outset of the study.
Phenomenological Research Design:
Phenomenological research is the study of phenomena, which can be events, situations, experiences, or concepts as they are perceived by individuals. This approach aims to understand and describe the essence of lived experiences from the perspective of the participants. Phenomenology begins with the acknowledgment that there is a gap in our understanding and that clarification will be beneficial.
In phenomenological research, researchers seek to explore the essence and meaning of experiences as they are lived. It often involves in-depth interviews and analysis to uncover the common themes and structures that underlie these experiences. This research design is valuable in understanding how individuals make sense of the world and their experiences.
Grounded Theory Research Design:
Grounded theory is a systematic research methodology that aims to develop theories or concepts that are grounded in data. It was developed by sociologists Barney Glaser and Anselm Strauss in the 1960s. The primary goal of grounded theory is to generate new, abstract theories based on empirical observations, rather than testing existing theories.
Key features of grounded theory research include:
Data-Driven Approach: Grounded theory starts with data collection and analysis. Researchers gather and analyze data without preconceived notions or hypotheses.
Constant Comparison: Data is constantly compared to previously collected data, allowing researchers to identify patterns and categories.
Theory Development: Through iterative data analysis, researchers develop theories or concepts that explain the phenomenon under investigation.
Sampling: Researchers use purposeful sampling to select participants and gather data that is relevant to the research question.
Grounded theory is widely used in the social sciences, particularly in fields like sociology and psychology, to develop new theories and understand complex social phenomena.