CDS 501 Week 4

  1. Research Design
    • plan that describes all protocols for the research project 
    • types of subjects
    • subject selection 
    • obtaining informal consent 
    • list and explain variables 
    • how to ID survey population
  2. Goal of Research Design
    • answer the hypothesis and other questions asked 
    • control extraneous variables 
    • make sure the relationship between the IV and DV is sound and logical and can be measured and interpreted
  3. Poor Research Design
    • leads to faulty conclusions 
    • has low internal validity and replicability
  4. Good Research Design
    • leads to meaningful conclusions 
    • high internal validity and replicability 
    • answer posed questions 
    • control for alternative explanations and factors
  5. Experimental vs Quasi-Experimental
    • Experimental: has random assignment to groups and manipulated IV 
    • Quasi-experimental: no random assignments to groups 

    • *most research for SLPs is quasi-experimental
    • we work with subjects we have and align as many critical groups as possible
  6. Steps for the Research Plan
    • 1. ID target population - make the sample group representative of the population as a whole 
    • 2. Develop sampling protocol - decide how you'll assign subjects to control/experimental group 
    • 3. Select design that best answers research questions 
    • 4. Choose appropriate statistical test
  7. Non-experimental Research Designs
    • investigate existing conditions without manipulating them 
    • observe/describe behaviors 
    • determine relationship of measures with different skills 
    • compare people with different characteristics
  8. Survey Research
    • subject of survey is the most important factor when determining subjects 
    • has to answer the questions brought up by research design
  9. Survey Research Questions
    • yes/no 
    • categorical response 
    • rating scale
    • open-ended
    • Like-hart scale: strongly agree (5) to strongly disagree (1)
  10. Advantages of Written Surveys
    • include graphics
    • greater privacy for respondents 
    • low cost 
    • recruit more participants
  11. Advantages of Interviews
    • explain the survey 
    • answer participant questions 
    • modify questions
    • ask follow-up questions 
    • include people who can't read and/or write
  12. Case Studies
    • descriptive studies of on individual 
    • descriptive studies of units such as specific classroom, acute care facility, industrial setting etc. 
    • use multiple sources of evidence (combo of speech samples, work samples, results of speech, language, and hearing tests)
  13. Longitudinal Studies
    individual or group followed over long period of time
  14. Longitudinal Case Studies
    correlation study that tries to predict future abilities
  15. Cohort Study
    • allows study of emergence of disorders over time 
    • view long term effects of particular disorders 
    • view long term effects of treatment
  16. Regression Studies
    determine the strength of relationship between two variables
  17. Correlation Studies
    • determines how two variables are related in some way
    • cannot establish causality
  18. Most Common Experimental Research Designs
    • Single Subject
    • Group 
    • Mixed-methods
  19. Group Designs
    one or more groups of subjects exposed to one or more levels of IV
  20. Single Subject Designs
    data on subjects is considered individually, not as a group average
  21. Posttest Only Designs
    • assigned randomly 
    • experiment ensues
    • subjects tested only after treatment is administered
  22. Test-teach-retest Designs
    • test subjects at the beginning of the study; form a baseline 
    • administer treatment 
    • retest subjects 
    • most frequently used 
    • can measure subject growth
  23. Between Subjects Design
    • different groups get different levels of the IV
    • measure behavior between two groups to determine the effect of the IV
    • may be bivalent or multivalent 
    • try to get the groups to be as much alike in number and characteristics as possible (group equivalence)
  24. Within Subjects Design
    • performance of same subjects compared across different situations/conditions 
    • everything but the IV should stay the same 
    • make sure changes seen are from the IV and not nuisance variables 
    • subjects might change based on previous treatment given, not because of particular change in the IV (sequencing effect)
  25. Mixed Designs
    combo of between and within designs
  26. Nonequivalent Control Group Designs
    • two preexisting groups 
    • one group gets the experiment and the other is control/alternative treatment
  27. Repeated Measures Research Design
    • 2+ measurements from the same group 
    • participants are their own controls 
    • participants experience all levels of the IV 
    • can be attached to any research design
    • allow time between measures to limit sequencing effect
Card Set
CDS 501 Week 4
CDS 501