Youve Survived 48 Hours

1
Youve Survived 48 Hours

• Congratulations!

2
VI. Process of Research

• A.      Statement of a problem to be
  investigated.
• B.       Delineation of a method for
  investigation of the problem.
• C.       Presentation of the results of this
  investigation.
• D.      Drawing conclusions from the results
  about the problem.

3

• VII. Variables
• A. Dependent Variables
• B. Independent Variables
• C. Active and Assigned (Attribute) Variables
• D. Continuous Versus Categorical Variables

4
VIII.Types of Research

• A. Experimental Research Defining
  Characteristics
• 1. Experimenters begin with a purpose, question,
  or hypothesis that allows them to know when to
  observe certain aspects of behavior.
• 2.       Experimenters can control the occurrence
  of events, thus observe changes in behavior when
  they are best prepared to make the observations.
• 3.       The results of the experiment can be
  replicated.
• 4.       Because the experimenters can control
  the conditions of observation, they can
  systematically manipulate certain conditions to
  measure the effects of these manipulations on
  behavior.
• B. The Importance of Experimental-Clinical
  Research
• C. Parametric Research
• D. Descriptive Research
• E Developmental Research

5

• F. Correlational Research

6
(No Transcript)
7

• G. Survey Research
• H. Comparative Research
• I.   Combined Experimental-Descriptive Research

8

• IX. The Relation Between Research Types and
  Questions
• X. Encountering Problems and Their Causal
  Relations
• A. Populations
• B.       Samples
• XI. Research Design in Communicative Disorders
• A.                  Group Designs
• 1.        Between-Subjects Designs
• a. randomization.
• b. matching
• 2. Within Subject Studies
• B. Single-Subject Research Designs

9

• XII. External Validity
• A. Inferential Validity
• B. Across Settings
• C. Across Experimenters
• D. Across Response Classes
• XIII. Factors That Affect External Validity
• A. Subject Selection
• B. Reactive or Interactive Effects of
  Pretesting
• XIV. Controlling Variability

10

• XV. Measurement Issues in Communicative Disorders
  Research
• A. Traditional Levels of Measurement
• 1. Nominal
• 2. Ordinal
• 3. Interval
• 4. Ratio

11

• B. Measuring Communicative Behaviors
• 1. Frequency
• 2. Durational Measures
• 3. Interresponse Time Measures
• 4. Latency measures
• 5. Time Sampling Measures
• 6. Momentary Time Sampling
• 7. Verbal Interaction Sampling
• C. Client Assisted Measurement
• D. Indirect Measures Self-Reports
• E. Covert Measurement

12

• XVI. The Observer
• A. Observer Bias
• B. Observer Training
• C. Mechanical Aids
• XVII. Reliability of Measurements
• A. Stability Measures
• B. Equivalence Measures
• C. Internal Consistency Measures

13

• XVIII. Validity of Measurements
• A. Content validity
• B. Criterion validity
• 1. concurrent validity
• 2. Predictive validity
• C. Construct validity

14

• XIX. Evaluating Treatment Efficacy Research
• XX. Threats to Internal Validity
• A.                  History
• B.                  Maturation
• C.                  Testing
• D.                  Instrumentation
• E.                  Statistical Regression
• F.                  Subject Selection Biases
• G.                  Attrition (or Mortality)
• H.                 Hawthorne Effect
• I.                   Diffusion of Treatment

15

• XXI. Other Important Issues in Research Design
• A. Systematic Variance
• B.       Error Variance
• XXII. Some Research Designs
• A. Nonexperimental Designs.
• 1.       Descriptive Studies
• 2.       Predictive Studies

16

• B. Experimental Designs
• 1. Group Designs
• a. Pre-Experimental Designs
• One-Shot Case Study
• One Group Pretest-Posttest Design
• Static Group Comparison

17

• b. True Experimental Designs
• Control Groups--
• Randomization--
• Matching --
• Pretest-Posttest Control Group Design
• Posttest Only
• Solomon Four-Group Design

18

• Designs to Evaluate Multiple Treatments
• Multigroup Pretest-Posttest Design
• Multigroup Posttest-Only Design
• d. Factorial Designs
• e. Randomized Blocks Design
• f. Completely Randomized Factorial Design
• Quasi-Experimental Research
• g. Nonequivalent Control Group Designs
• h. Separate Sample Pretest-Posttest Design

19

• 2. Single Subject (Time Series) Designs
• a.  Single-Group Time Series Design With
  Temporary Single Treatment
• b.  Single-Group Time Series Design With
  Continuous Treatment and Withdrawal
• c.  Single Group Time Series Design With Single
  Continuous Treatment
• d.  Single Group Time Series Design with Multiple
  Temporary Treatments

20

• 3. Counterbalanced Within-Subjects Designs
• a.       One Group Single-Treatment
  Counterbalanced Design
• b.       Crossover Design
• 4. Limitations of Counterbalanced Within-Subjects
  Designs
• a. Order Effects
• b. Carryover Effects
• c. Ceiling and Floor Effects
• 5. Correlational Analysis

21

• XXIII. Single Subject Designs
• A. Control Mechanisms in Single Subject
  Designs
• 1. Replication
• 2. Withdrawal
• 3. Reversal
• 4. Reinstatement of Treatment
• 5. Baselines
• 6. Multiple Baselines
• B. ABA Single Subject Design
• C. ABAB

22

• D. Multiple Baseline Designs
• 1. Multiple Baselines Across Behaviors
• 2. Multiple Baselines Across Settings
• 3. The Problem of Repeated Baselines

23
(No Transcript)
24
(No Transcript)
25

• E. Designs for Multiple Treatment Comparisons
• 1. ABACA/ACABA Design
• 2. Alternating Treatments Design
• F. Ineffective Treatments in Multiple Treatment
  Evaluations
• G. The Interactive Design
• H. Changing Criterion Design
• I. Designs to Assess Response Maintenance
• XXIV. Research Question

26

• XXV. Organization and Analysis Data
• A. Graphic Displays of Frequency Distributions
  (parts 3-4)
• B. Types of Data Distributions
• 1. Measures of Central Tendency
• 2. Measures of Variability Range and
  Standard Deviation
• 3. Standard scores (z-scores) and percentile
  ranks

27

• XXVI. Basic Concepts of Data Analysis
• A. The Normal Curve Model
• B. Parametric Nonparametric Stats
• 1. Parametric
•          The population parameter is normally
  distributed
•          The parametric measure is an interval
  or ratio measurement
•          If we have two or more data
  distributions to be analyzed (e.g., experimental
  and control groups) the variances should be
  similar
•          The sample should be large (assume 30
  or more)
• 2. Testing a Null Hypothesis
• 3. Type I and Type II Errors
• 4. Degrees of Freedom

28

• C. Measures of Relationships Correlation and
  Regression
• D. One-Sample Case
• E. The t-test
• F. Three or More Groups (ANOVA)
• G.  Post-Hoc Tests
• H. Non-Parametric Alternatives (p. 185, Table
  6.18)
• XXVII. How to Formulate Research Questions

29

• XXVIII. How to Write Research Reports
• a.       Introduction
• b.       Methods
• c.       Results
• d.       Discussion
• e.       References