Research Design Flashcards
Overview of Research Design Types
HDFS 361 | Applied Research Methods focus on four primary research designs: Descriptive, Correlational, Experimental, and Quasi-Experimental.
Core Objective: These notes explore how research design shapes the conclusions a researcher can and cannot draw.
Career Connection: Professionals in social services, policy, public health, and education frequently encounter research claims and must develop the skills to evaluate them critically.
Descriptive Research
Primary Goal: To describe existing phenomena and generalize those descriptions to a larger population.
Key Questions Answered by Descriptive Research: * Central Tendency: What is the average? * Variability: How spread out are the values? * Frequencies / Percentages: How common is a specific occurrence?
Critical Constraint: Descriptive research does not establish connections between variables.
Methodological Visualization Tools: * Pie Charts * Histograms * Bar Charts
Detailed Research Examples: * Academic Performance: Determining the average GPA of high-schoolers from low-income families and measuring the observed variability. * Healthcare Access (Frequency/Needs Assessment): Determining the percentage of people with limited English proficiency who receive interpreter support during medical visits. * Program Planning Baseline: Determining the proportion of youth in a specific county meeting physical activity guidelines.
Correlational Research
Primary Goal: To describe how multiple variables naturally co-exist without any external manipulation by the researcher.
Nature of Study: Usually cross-sectional; occasionally longitudinal. It is noted as the most common and least invasive research design.
Function: Examines the association between at least two variables.
Types of Analysis in Correlational Research: * Between-group differences: For example, comparing if students from public versus private high schools are more likely to attend college. * Correlations: For example, determining the association between math and reading test scores. * Regression: For example, determining how factors like race, gender, and perceived social support predict academic achievement.
Terminology: * Predictor: Equivalent to the independent variable (IV). * Outcome: Equivalent to the dependent variable (DV).
Understanding the Correlation Coefficient
Mathematical Range: The correlation coefficient ranges from to .
Strength Levels: * Closer to indicates a weaker association. * Closer to indicates a stronger association. * is considered Weak. * is considered Medium. * is considered Large.
Direction and Meaning: * : No relationship. * : Perfect positive relationship. * : Perfect negative relationship.
Limitations of Correlational Research
The Third Variable Problem: Researchers must always consider alternative explanations. For example, ice cream sales and drownings show a positive correlation (), but the actual driving force is the "Third Variable" of Summer Heat, which causes both to increase.
The Directionality Problem (Chicken and Egg): Even if two variables are related, it is often unclear which causes which. * Example: In the relationship between after-school tutoring and academic motivation, it is unclear if tutoring raises motivation or if highly motivated students seek out tutoring.
Advanced Structural Relationships: Moderation and Mediation
Moderation: The association between variable and variable depends on the level of a third variable, the Moderator (M). * Example: High stress () predicts depressive symptoms () only at low levels of perceived social support (). At high levels of social support, the link between high stress and depressive symptoms disappears.
Mediation: A Mediator (M) explains "how" or "why" affects . It is the mechanism in the middle (). * Example: Increased physical activity () leads to higher endorphins (), which then leads to a better mood ().
Experimental Research
Primary Goal: To establish causal associations by testing conditions.
Defining Characteristics: * The researcher actively manipulates the independent variable (IV). * Participants are randomly assigned to specific conditions.
Logic of Causality: * Control: Random assignment controls for many third variables simultaneously. * Temporal Precedence: The researcher decides what happens first (the treatment), making directionality clear.
Validity Trade-offs: * Internal Validity: The degree to which changes in the outcome can be confidently attributed to the treatment rather than alternative explanations. * External (Ecological) Validity: The degree to which findings from a controlled study generalize to real-world settings, populations, and conditions.
Mechanics of the Experimental Design
The Power of Random Assignment: This process distributes all possible third variables (both known and unknown) equally across groups. Differences observed after treatment cannot be attributed to pre-existing third variables.
Experimental Study Flow Chart: 1. Recruit Sample. 2. Collect Pretest Data. 3. Randomize Into Groups. 4. Groups diverge (Control Group does its usual thing while the Treatment Group receives the program). 5. Collect Posttest Data.
Group Classifications: * Treatment Group: Receives the actual program or intervention being tested (e.g., an 8-week mentoring program). * Control Group: Does not receive the treatment; provides a comparison baseline (e.g., a waitlist group continuing life as usual). * Placebo Group: Receives something inert or an attention-equivalent condition to control for the expectation of receiving something (e.g., attending unrelated workshops for the same amount of time).
Ethical Standard: After the experiment concludes, the control group should be offered the opportunity to receive the treatment.
Quasi-Experimental Research
Primary Goal: To get as close as possible to establishing causality when true random assignment is not feasible.
Key Distinguishing Factor: Absence of random assignment. Participants enter groups via natural circumstances, self-selection, or eligibility criteria rather than a flip of a coin.
Common Designs: * One-Group Pretest-Posttest: Measuring subjects before and after an intervention without a control group. * Natural Experiments: Exploiting naturally occurring variation, such as comparing a school implementing a new intervention to one that is not.
Why Quasi-Experimental is Used: * Ethical Constraints: It is unethical to withhold necessary services from people for the sake of an experiment. * Practical Constraints: Some variables cannot be assigned (e.g., a researcher cannot force people to smoke to study its effects) or are rolled out district-wide.
The Real-World Trade-off: While random assignment is the only way to confidently establish causality, quasi-experimental designs are stronger than purely correlational studies, even though alternative explanations may remain.
Research Activity and Application
Activity Prompt: Create a mini research study by choosing a topic (e.g., personal phone usage per hour, a pet's favorite toy) and selecting a design (Descriptive, Correlational, Experimental, or Quasi-Experimental).
Operationalization: Define and measure variables and predict results based on the chosen design structure.
Comprehension Check: Questions & Discussion
Question 1: Which strategy is best for establishing cause-and-effect? * Answer: (c) Experimental.
Question 2: Study on older adults, social support, and well-being relationship? * Answer: (a) Correlational.
Question 3: Parents reporting closer relationships after a parenting course? * Answer: (d) Quasi-Experimental.
Question 4: What is necessary after finding parents benefit from a course? * Answer: (b) Offer the control group the opportunity to complete the course as well.
Question 5: Example of a confounding third variable? * Answer: (d) SES influencing both the number of books at home and the quality of teachers (Master's degrees), creating a false correlation between teacher degree level and reading ability.
Question 6: Examining kindergarten readiness without random assignment? * Answer: (d) Quasi-Experimental.
Question 7: Percentage of students from single-parent households graduating high school? * Answer: (b) Descriptive.
Question 8: Description of a negative correlation? * Answer: (b) Lower levels of household chaos are associated with higher levels of family cohesion.
Question 9: Tutoring associated with motivation only when teacher connection is strong? * Answer: (b) Teacher connection moderates the relationship.
Question 10: Limitation of a highly controlled lab experiment with significant stress reduction? * Answer: (c) The findings may not generalize to real-world settings where conditions cannot be controlled.