Research approaches and design Flashcards
What is research and what does it do?
Creative and systematic work aimed at increasing knowledge and using this knowledge to develop new applications
Research is used to:
1. Establish or confirm facts
2. Reaffirm results of previous work
3. Solve new or existing problems
4. Support theorems
5. Develop new theories
What are two types of research?
- Fundamental/pure research: Advance knowledge without/with little concern for immediate, practical benefits
- Applied research: Research designed with a practical outcome in mind, and with the assumption that some group/society will gain specific benefit from it
10 general steps in applied research
- Identify research problem
- Literature review
- Specify research purpose
- Determine research questions
- Specify conceptual framework - sometimes including hypothesis
- Data collection
- Data verification
- Data analysis and interpretation
- Reporting and evaluation of research
- Communicate findings and possible recommendations - in a report
Definition research approaches vs research design
Approaches: Plans and procedures for how to go from general assumptions to detailed methods for data collection, analysis and interpretation
Design: The structure and framework within which data is collected, link between the research question and the evidence –> plan for answering research questions using empirical data
Three types of research approaches and what do you select approach based on
- Quantitative
- Qualitative
- Mixed-method
Select based on
1. Worldview
2. Nature of research problem
3. Researchers experience
4. Study audience
Definition and characteristics of quantitative approach
Test objective theories by examining the relationship among variables
- Data nature: Numbers and quantified observations
- Sampling: Representative samples for generalizability
- Data collection: Predetermined instruments and variables, closed-ended questions
- Analysis: Statistical procedure
- Focus on replicability
- Set structure on the written report
Definition and characteristics of qualitative approach
Exploring and understanding the meaning individuals/groups assign to a social or human problem
- Data nature: Word, text, recordings
- Sampling: Selective sampling for depth and detail
- Data collection: Open-ended questions, interviews, case studies, natural observations
Analysis: inductive, building general themes for specific details - Emphasizes detail, complexity and context-specific over breadth
- Flexible structure on report
Definition mixed approach
Combination of quantitative and qualitative approaches to generate deeper insights and bridge the gaps between broad generalization and nuanced interpretations
Three types of quantitative research design
- Non-experimental design: Observe and measure variables that have already occurred without manipulation or influence, often used to describe phenomena and explore correlations. Data collection methods include surveys, observations and secondary data.
Two types: Descriptive design and correlational design
- Experimental design: Determines if a specific treatment or intervention influences an outcome, suitable for studying cause-and-effect relationships and involves control groups or random assignment to eliminate bias. Data collection: Multiple iterations of an experiment.
Pro: Clear identification of causal relationship, con: findings might not always translate to real-world scenarios/general populations
- Quasi-experimental: Investigates causal relationship when random assignment is not possible, sues redefined groups based on existing IV (gender, geographic location) -> therefore useful for studying naturally occurring group differences but lacks full control of true experimental design. Data collection: Survey, observations
Pro: Applicable to real-world contexts where randomization isn’t possible, con: increased risk for bias due to lack of random assignment
Experimental design
It tests the impact of an intervention on an outcome, controlling for all other factors that might influence the outcome
- A way to plan experiments in advance to make results objective and valid.
Tries to prove/disprove a hypothesis with statistical analysis
Useful for establishing causality between IV and DV
What eight step does an experimental design plan cover to systematically test a hypothesis?
- Hypothesis
- What are your IV and DV?
- How will you balance internal and external validity?
- How will you manipulate your IV? (random assignment ensures consistent manipulation without bias)
- How will you measure your DV? (data collection)
- How are participants selected? (sampling)
- How are participants allocated to different conditions in an experiment?
- How will you evaluate your experiment?
Definition hypothesis
A testable prediction about the relationship between two variables
Internal validity and external validity
Internal validity: Ensuring observed effects are due to IV and not confounding variables (variables that correlate with both the IV and DV which may distort results)
External validity: Ensuring results generalize to the real world
What are five main concerns in experimental design?
- Ethics, anonymity, GDPR
- Validity: Internal and external
- Reliability: Consistency of measurements and results when repeating the experiment
- Replicability: The ability to repeat the experiment under the same conditions and obtain similar results
- Sampling: How are participants selected
How do you create a representative sample?
Know your population, identify groups or clusters that might make some groups distinguishable from other groups, ensure proportional representation within your sample to the population
What experimental design concerns can lack of proportional representation lead to?
This is an acknowledged problem that can cause bias and disrupt external validity.
What are four good practices for sampling/ensuring a sample representative of the population?
- Understand your population
- Proportional sampling
- Survey design: Ask participants background questions to avoid bias and overrepresented groups
- Combine sampling methods
Four sampling methods
- True random sampling: Every individual has equal change of being selected
- Systematic ordered sampling: Select participants at regular intervals from an ordered list, ex every 10th person
- Stratified sampling: Divide popluation inte distinct strata and randomly sample from each subgroup in proportion to its size
- Cluster sampling: Divide population inte clusters and randomly select entire clusters
Three data collection methods
- Surveys (stated preference - asking ppl directly about behavior and preference): Include questioners (quantitative) and interviews (qualitative)
- Survey (observations ): Systematically measuring or counting events (quantitative), detailed notes/recordings/participant observ. (qualitative)
- Secondary data: Government surveys, census data, socio-economic data, geodata
How to write a good survey
- Clear, short and straightforward questions
- Ensure consistent interpretation of questions by all people
- Not too long
- Careful writing, editing, reviewing and rewriting
- Pilot-test it on a representative sample of target audience
