Chapter 9: Conducting Experiments Getting Your Hands Dirty

1
Chapter 9Conducting Experiments(Getting Your
Hands Dirty!)
2
Purpose of Chapter

• Looks at practical issues that need to be
  addressed (apart from the experimental design
  used) when conducting experiments
• Chapter looks at selecting participants,
  manipulating your IV, measuring variables, and
  how to disseminate the results

3
1. Selecting Research Participants

• Goal sample participants from a population of
  interest.
• Use probability sampling when it is important to
  
• generalize results to the population

4
1. Selecting Research Participants

• A great deal of research is interested in testing
  hypotheses about what variables affect behavior
  without worrying about generalizing results.
• Participants may be found in the easiest way
  possible using haphazard or convenience
  sampling methods.
• Sample sizes needed usually large

5
2. Manipulating the IV

• To manipulate an independent variable
• Researcher must first convert a conceptual
  variable into a set of operations (operational
  definition) Examples?
• Researcher must set the stage by
• 1. Provide informed consent information
• 2. Explain why the experiment is being
• conducted (Rationale)

6
2. Manipulating the Independent Variable (cont)

• Types of IV manipulations can be either
• A. Straightforward manipulations
• B. Staged manipulations
• More

7
Manipulating the Independent Variable (cont)

• A. Straightforward manipulations
• Need to have 2 or more levels (values) of
• Written instructions
• Verbal instructions or comments
• Visual stimuli

8
Manipulating the Independent Variable (cont)

• B. Staged manipulations participants are put in
  a situation to see its effect on them
• Often employ confederates
• Events are staged or manipulated to
• 1. Create some psychological state
• 2. Simulate some situation that occurs in
  the
• real world (like peer pressure, for
  example)

9
Manipulating the Independent Variable (cont)

• Researcher needs to consider the strength of the
  manipulation
• Best to maximize differences between levels
  (especially in a new research area!)
• E.g., compare effects of 6 beers vs. no beers on
  math skills
• E.g., compare effects of very hot room
  temperature vs. very cold on reading speed

10
3. NowMeasuring the Dependent Variable

• Types of DV measures often used (each has
  strengths and weakness)
• Self-report measures easy to do BUT very prone
  to have subjective bias errors
• Behavioral measures harder to do but less
  subjective bias (although may be more observer
  bias errors is operational definition is weak)
• Physiological measures may require expensive
  equipment and may NOT be connected to actual
  learning or behavior changes

11
Measuring Dependent Variable (cont)

• Is DV sensitive enough? (Can it detect small
  differences/changes in participant behavior when
  the IV is manipulated?)
• Ceiling effect seen in DV means that task is so
  easy everyone does great regardless of IV (think
  of a VERY easy exam, for example).
• Floor effect in DV means that task is so
  difficult everyone does horribly regardless of IV
  (think of a VERY difficult exam, for example).

12
Measuring Dependent Variable (cont)

• We can also use Multiple DV measures
• Means to measure more than one dependent
  variable
• - A variable can be measured a variety of ways.
• - If the IV has the same effect on several
  measures of the same dependent variable,
  confidence in the results increases

13
NowAdditional Controls for Confounding in
Experiments

• (previously weve discussed some of these)
• Controlling for participant expectations
• Demand characteristics use deception
• Cover story
• Filler items
• - Ask participants their perceptions about the
  purpose of the study

14
Additional Controls (cont)

• Use a Placebo group
• Is a type of Single-blind experiment
  (participants dont know who is getting the
  real treatment so expectations of participants
  in each groups should be same)

15
Additional Controls (cont)

• Controlling for experimenter expectations
• Experimenter bias may be a big source of error!
• Experimenter may look for effects that arent
  there!
• Research on expectancy effects
• Parapsychology studies
• Teacher expectancy (Rosenthal Jacobsen, 1968)

16
Additional Controls (cont)

• Solutions to the experimenter expectancy problem
• Experimenter training about experimenter bias
• Automate the procedures (e.g., computer scoring)
• Use a double-blind study!

17
Additional Considerations to Make Sure Research
Works Well

• Write a research proposal before doing any
  experimentation to thoroughly think out the study
• Pilot the study to work out any kinks
• During pilot, take a quick manipulation check of
  effects of IV on DV
• Pilot Debriefing besides ethical reasons, also
  gives researcher insight into what perception of
  participants is

18
Communicating Research to Others

• After data results are collected, analyzed, and
  interpreted, you should communicate the findings
• Write the report first and then communicate it
  through
• 1. Professional meetings (talk or poster)
• 2. Journal articles
• 3. Popular sources (books, magazines)

19
The End