Types of Research Design PDF

Types of Research Design Two overarching categories Qualitative Quantitative Qualitative Research Specific procedures determined during the research and before. Design emerges step by step. Types of Qualitative Research Ethnographic Designs Observation Interviews (Narrative re...

Types of Research Design Two overarching categories Qualitative Quantitative Qualitative Research Specific procedures determined during the research and before. Design emerges step by step. Types of Qualitative Research Ethnographic Designs Observation Interviews (Narrative research) Document Analysis Analytical Designs – Studying the past – Historic Document analysis Legal Historical – Contexts and trends Conceptual Ethnographic Design Example Example #2 Quantitative Research Experimental Research Cause-and-effect investigation One or more independent variables are manipulated Measure the dependent variable Intervention is pre-determined and set-up by the investigator Typically (ideally) two or more groups Design for maximum control Threats to Internal Validity What is internal validity? Threats Selection of subjects History Maturation Testing Attrition Instrumentation Threats to External Validity What is External Validity? Threats Population validity Personologicial Variables Ecological Validity Essentials for research design Reliable Measurement Repeated measurement Description of the conditions Baseline and treatment condition Control condition? Single-variable rule Types of Quantitative research Non-experimental Experimental Case-study Non-experimental Quantitative Research Descriptive Correlational Ex-Post Facto Survey Descriptive Research Tells us what something is Correlational Research Non-experimental research method that examines the relationship between two or more variables by measuring them with any manipulation. Allows researchers to identify associations/patterns between variables without establishing a cause-and-effect relationship. “How variables change together” Non-kinesiology examples Study habits and exam results Ice cream sales and crime rates (both increase during warmer weather) Limitations of Correlations Research Can’t identify cause-and-effect Less rigorous than experimental research Prone to identify spurious relational patterns Relational patterns are often arbitrary and ambiguous Encourages a “shot-gun” approach to research Types of Correlational Research Simple Correlational Studies Prediction Studies Shotgun Approach Simple correlation Predictive Correlational design Uses statistical analysis to predict the value of a variable based on its relationship with other variables Trying to predict an outcome Shotgun correlational studies Cautions of correlational research Spurious correlations Restriction of range “Significant” correlations can be small  Ex-post Facto study Translates to “after the fact”. Study something that has already occurred Means that the independent variable can’t be manipulated Looks for potential effects of the independent variable on the dependent variable. Survey Research Design Data collected by asking questions in a structured questionnaire or interview, allowing researchers to gain insights into people’s thoughts, behaviors, and experiences on a specific topic, typically used to gather information about a larger population by surveying a representative individuals within that group. Experimental Research True Experimental Design Randomized controlled trial Independent Measures design (between groups) Repeated Measures design (within groups) Counter-balancing Matched pairs design Quasi-experimental Design Regression discontinuity design Randomized Control Trial Best type of study if possible Different participants used in each group Each group has a different condition Assigned to a group randomly Pro: Con -Avoids order effects -More people needed than Repeated measures design -Differences between participants may affect results (extraneous variables)  Control Test DV Test DV Population Randomization Test DV Treatment Test DV RCT – Independent groups example Randomized Control Trial – Repeated Measures Same participant participates in each condition Need to counterbalance Pro Con -Same participants in each condition, reduces -Order effects may exist individual differences (extraneous variables) -Fewer people needed (1/2 as many!) No counterbalancing RCT RM – No counterbalancing Control Control Treatment Population Test DV Test DV Test DV RCT RM – with counterbalancing Control Treatment Test DV Test DV Test DV Population Randomization Test DV Treatment Test DV Control Test DV RCT RM Crossover Design RCT RM example Quasi-experimental Looks just like an experimental study, just without the randomization. Other types of research Case report Detailed account of a patient’s diagnosis, treatment, and follow-up, often written in the form of a story. Key part of medical progress Can provide new ideas, especially in unusual or novel cases Not technically research