HSBH 1007: Health Science and Research Lecture Week 13

1
HSBH 1007Health Science and ResearchLecture
Week 13

• Longitudinal Research and Typical Data Analyses
• Dr Andrew Campbell
• A.Campbell_at_usyd.edu.au
• www.psyberspace.info

2
Study Reference

• Textbook Research Methods for Health Sciences
• Topic 7 Pages 203 - 224

3
Overview of todays lecture

• The purpose of unique methods and research
  designs in health
• Single Case Experimental Designs- Reversal
  Designs
• Multiple Baseline Designs
• Replications in Single Case Designs
• Quasi-Experimental Designs- Program Evaluation-
  One-Group Posttest-Only Design- One-Group
  Pretest-Posttest Design

4
Continued

• Non-Equivalent Control Group Design
• Interrupted Time-Series Design
• Control Series Design
• Developmental Research Designs- Longitudinal
  Design- Cross-Sectional Design-
  Cohort-Sequential Design

5
A Recap from Lecture 1 Research is important
because

• All practice comes from research
• All teaching is research led
• Research becomes habit in Allied Health

6
The purpose of unique methods and research
designs in health and science

• Quantitative and Qualitative Research has many
  different purposes and designs.1. Drug
  research2. Surgical Procedures3. Mental
  Health4. Management5. Education

7
Single Case Experimental Designs

• Meaning A single-subject design.
• Skinner (1953) used this for reinforcement of
  new behaviour.
• Application example in my researchPilot study
  for ADHD video game treatment

8
Single Case Experimental Designs- Reversal
Designs

• How do you know if the treatment really
  worked??
• Do the experiment backwards!1. Measure
  Baseline2. Introduce treatment or
  manipulation3. Stop treatment and measure
  baseline.
• Example Drug treatment for Cancer.
• Problem Ethical consideration?

9
Multiple Baseline Designs

• Multiple Baseline Designs are perhaps more
  ethical.
• They dont generalise.
• Different conditions can be introduced to improve
  behaviour.
• Different settings also can change behaviour!

10
Replications in Single Case Designs

• This design is still usefulBUT1. Is normally
  published in relation to many other single case
  designs.2. Produces the data for each
  individual rather than as a group.
• By producing individual results we specify
  individual effects. Sometimes this is more
  accurate than group means.
• Example ADHD Biofeedback works better for 1
  week intervention per child, than 3 times a week
  per child.

11
Quasi-Experimental Designs- Program Evaluation

• Often employed to examine hard-to-control
  conditions.
• Question asked Are we going to achieve what
  weve planned and implemented?
• Methods often pulled from similar, but not exact,
  studies.
• Example Health in Space cant be replicated on
  Earth!

12
Quasi-Experimental Designs- Program Evaluation
Stages

• Program Evaluation Problem Solving
• What is needed to keep Astronauts healthy on a
  trip to Mars?
• Needs Assessment (The Beginning Thoughts)
• Program Theory Assessment (Thoughts and arguments
  based on prior Research)
• Process Evaluation (Simulation)
• Outcome Evaluation (Did we succeed?)
• Efficiency Assessment (Can we replicate and
  improve is it worth it?)

13
One-Group Posttest-Only Design

• Theory Babies enjoy peeing on President G.W.
  Bush.
• Each baby he hold pees on him.
• Average peeing speed is 4.6 seconds from
  contact.
• Does this design validate that all babies will
  pee on him?
• Design is flawed no environmentally
  controlled condition baby may have peed on
  anyone!

14
One-Group Pretest-Posttest Design

• What if their was a controlled condition pick up
  after baby peed and pick up after baby drank a
  lot?
• Problem - Still does not account for the
  following
• History (Bladder Problem)
• Maturation (Age of Baby)
• Testing (Accuracy of the measure when did baby
  last drink?)
• Instrument Decay (Repeating test can cause
  inaccuracy due to external factors e.g. parent
  disinterested).

15
One other factor

• Regression toward the Mean (we all regress toward
  the mean in repeat conditions).
• Sometimes bad, because we want a true score.
• To over come this, we need a control condition
  that does not receive treatment or
  manipulation!
• Example Baby is picked up by parent in condition
  1 (control). Baby is picked up by President G.W.
  Bush.

16
Non-Equivalent Control Group Design

• Example Quit Smoking Program in industry.
• Smokers in each condition control and
  treatment.
• Is this a good design?
• No Other factors to consider like smoking
  history and job conditions.
• Would a controlled Post-Test Pre-test design
  help?
• Yes Looking at pre-test scores can determine if
  difference are present before test condition.

17
Interrupted Time-Series Design

• 1955 car crash fatalities in Connecticut, USA.
• Seatbelts came in.
• Compared a year later in 1956 reduced
  fatalities.
• Simple way to improve that validity extend the
  time-series design!
• Even better improvement find a control group!
  Compare with a non-seatbelt state in the US.

18
Longitudinal Research

• This type of research considers groups of
  subjects over a period of time.
• To study change in language
• The development of Personality and Aggression
• How intelligence develops
• Perceptual abilities across the lifespan
• Impact of Life choices! 7 Up Series.

19
Weaknesses of Longitudinal Research

• The time it takes to complete
• The cost and commitment of the researchers
• Drop out rates of subjects
• Death

20
Cross-Sectional Research

• This type of research considers groups of
  subjects of different ages at the same point in
  time.
• Each of the age groups is called a Cohort
• A cohort is roughly a group of people of the
  same age or developmental level.
• Example Online Gamers and Identity Development.

21
Cohort-Sequential Research

• This type of research begins as a cross-sectional
  study and then follows the cohorts longitudinally
  for a small number of years.
• Example 1st year BHS to Graduation!

22
Last Chance to see me about Electives and Majors!

• Come see me now about any questions you have
  about BHS electives, majors or to discuss the
  meaning of life!

But First A message from Karen Pepper about your
assignments and Exams