1.3 Research Decisions
Type of Research Strategy
Research paradigms guide how a study is conducted; research design involves planning key ideas before carrying out an investigation. Major decisions include:
- The type of research strategy to use
- The type of scientific reasoning (deductive or inductive)
- The type of data to collect and analyze
- These decisions shape how research is conducted and what conclusions can be drawn.
Quantitative research (numbers and measurements):
- Focuses on results that can be measured and expressed as numbers (e.g., sales figures, average hospital stay lengths).
- Produces mathematically precise results that, if done correctly, are hard to dispute.
- Representativeness is important in quantitative work. Results from a sample should reflect the whole population.
- Large samples are typical to test hypotheses and summarize data.
- Example: election polls survey people to predict overall voting patterns.
Qualitative research (text-based or non-numerical data):
- Interprets individual cases; results come from interpretation and are subjective and open to debate.
- Usually does not aim for representativeness; selects participants who can provide deep insights.
- Sample sizes are usually small due to complexity of collecting and analyzing data.
- Examples:
- Interviewing a few people to understand why they don’t vote, exploring motives in detail.
- Exploring why employee turnover is higher in some industries, capturing reasons that standardized questionnaires might miss.
- Goal: to obtain rich, meaningful information rather than broad, generalizable data.
- Note: Quantitative results are not guaranteed to answer the research question even if they are numerically accurate. For example, correlation does not imply causation.
- In text: “correlation does not equal causation” (Reed, 2005 as cited in Saunders et al., 2019).
- Critics argue qualitative research can be seen as storytelling or anecdotal evidence; researchers can find what they expect.
- Example: asking women in middle management about the “glass ceiling” often confirms its existence—because the question assumes it.
Reflecting on methods in qualitative research is important to avoid biased results and over-interpretation.
There isn’t a single best approach to research design; each approach works best for different goals.
- Quantitative is ideal for large samples, measurable data, and generalizability (e.g., measuring employee satisfaction across a company).
- Qualitative is better for exploring complex, less-studied topics with text-based data (e.g., conflicts or trauma in teams).
Triangulation: combining quantitative and qualitative methods to strengthen research by using multiple methods.
Types of Scientific Reasoning
The second key decision is the type of scientific reasoning: deductive or inductive.
- Deduction: Start with a theory or hypothesis and test it with data.
- Induction: Start with observations and develop a theory from the data.
Induction details:
- Looks at individual cases and draws general conclusions from them.
- Example: if you see a Ferrari on the road and it is very fast, you might conclude all Ferraris are fast. While this could be misleading, induction is useful for developing theories from a few observations to be tested later.
- References: (Saunders et al., 2019; Veal, 2018; Sheppard, 2004).
Deduction details:
- Starts with a general theory and tests it with specific cases.
- Example: if the theory says “all Ferraris are fast,” then any Ferrari you see should be fast. Test by measuring speed; if it isn’t fast, the theory is challenged.
- References: (Saunders et al., 2019; Veal, 2018; Sheppard, 2004).
Relationship to data types:
- Quantitative research usually uses deduction (theory-driven testing with data).
- Qualitative research often uses induction (theory emerges from observations), though deduction can also occur in qualitative work.
Type of Data
Primary data:
- Collected specifically for the current study.
- Examples: interviews or answers from questionnaires gathered for the project.
- References: (Veal, 2018; Rea & Parker, 2014).
Secondary data:
- Originally collected for another purpose but reused for a new study.
- Examples: economic data collected by a market research institute used to study links between religion and economic trends.
- References: (Saunders et al., 2019; Tantawi, 2021).
Type of Research to Be Carried Out
Two main types: experimental and non-experimental investigations.
- Non-experimental research often happens in real-life settings (field research).
- Example: interviewing managers at their workplace or observing behavior in its natural context.
- Experimental research takes place in a controlled environment where outside influences are minimized.
- Example: marketing researchers using laboratory supermarkets to test product placement effects (e.g., sweets near the checkout to encourage impulse buys).
- References for concepts: (Creswell & Creswell, 2018; Arrington, 2021; Stoica, 2021).
These decisions—experimental vs. non-experimental—shape how data is collected and analyzed.
Experimental vs Non-Experimental: Key Distinctions and Examples
- Non-experimental / Field research: Real-life settings, natural observation.
- Experimental research: Controlled environment, minimization of outside influences.
Key Concepts and Practical Implications
- Representativeness: The extent to which sample results reflect the population.
- Generalizability: The ability to extend findings beyond the sample.
- Interpretative vs numerical data: Qualitative vs quantitative data emphasis.
- Correlation vs causation: Not all observed associations imply a causal relationship.
- Notation:
- Sample size: Often larger in quantitative studies, smaller in qualitative studies.
- In-depth insights: Often the aim of qualitative work.
- Reflective practice: Qualitative researchers should reflect on their methods to avoid biasing results.
Keywords (as highlighted in the transcript)
- Research paradigms: Guide study design; assumptions behind research.
- Research design: Planning key ideas; includes strategy, reasoning, and type of data.
- Type of research strategy:
- Quantitative: Numerical data, measurable, large samples, generalizable, experiments, correlation vs causation.
- Qualitative: Text-based data, small samples, subjective, in-depth insights, anecdotal evidence, reflective.
- Triangulation: Combining quantitative and qualitative methods.
- Scientific reasoning:
- Deduction: Theory → test cases; common in quantitative research.
- Induction: Observations → general theory; common in qualitative research.
- Type of data:
- Primary data: Collected for the current study.
- Secondary data: Previously collected for other purposes.
- Type of research execution:
- Experimental: Controlled environment, minimize outside influence.
- Non-experimental / Field research: Real-life settings, natural observation.
- Additional key concepts:
- Representativeness
- Generalizability
- Interpretative vs numerical data
- Correlation vs causation
- Sample size
- In-depth insights
Real-World and Foundational Connections
- The choice between quantitative and qualitative methods depends on the research goal (generalizability vs deep, contextual understanding).
- Triangulation can help address weaknesses inherent in each approach by leveraging multiple methods.
- Reflective practice and critical analysis of methods are essential to mitigate bias and ensure the validity of findings.
Formulas and Notation Used in Examples
- Correlation does not imply causation:
- Inductive example (Ferrari speeds):
- If the theory says orall x ig(Ferrari(x)
ightarrow Fast(x)ig) then observing a Ferrari that is not fast would challenge the theory.
- If the theory says orall x ig(Ferrari(x)
Quick Recap / Takeaways
- Research design involves choosing a strategy, reasoning approach, and data type that align with the research goals.
- Quantitative vs Qualitative: trade-offs between generalizability and depth of understanding.
- Scientific reasoning shapes how theory and data interact (deduction vs induction).
- Data types (primary vs secondary) determine data collection methods and potential biases.
- Experimental vs Non-experimental approaches define the control over variables and settings.
- Triangulation and reflexivity enhance credibility and robustness of findings.
- Key concepts to monitor: representativeness, generalizability, interpretative vs numerical data, and the pitfall that correlation does not equal causation.