experimental RESEARCH

1
experimental RESEARCH

• BUSN 364 Week 12
• Özge Can

2

• Stanford Marshmallow Experiment
• Measuring delayed gratification (ability to wait
  in order to obtain something you want)
• http//www.youtube.com/watch?vx3S0xS2hdi4
• A child is given a marshmallow and promised
  another, only if they could wait before eating
  the first one.
• Researchers measure how long the child resists
  the temptation of eating it
• Whether correlated with future success or not

3
Experimental Research

• Builds on the principles of positivist
  perspective and natural sciences
• (1) Starts with a causal hypothesis
• (2) Modify one specific aspect of a situation
  that is closely connected to the cause, and
• (3) Compare the outcome to what existed without
  the modification

4
Example Experiment

• Manipulation of Independent Variable
• Question Do suggestions affect memories?
• Method Randomly assign participants to groups
  and tell some participants that dreams indicate
  prior experience but do not tell that to other
  participants.
• Results Dream interpretations induce changes in
  memories of past events.
• Conclusion We can influence peoples memories
  by giving them erroneous information

5
Attributes of Experiments

• Can powerfully test causal relationships
• An experiment is often artificial gt includes
  independent and dependent variables but exclude
  confounding variables
• Confounding Variables Factors that are not part
  of the intended hypothesis but have effects on
  variables of interest
• Best for issues that have a narrow scope and
  sharpened focus

6
Attributes of Experiments

• Isolates and targets one or a few causal
  variables not effective for considering dozens
  of variables simultaneously
• Best suited for micro-level (individuals,
  small-groups) more than macro-level (entire
  society) theoretical concerns
• Practical and ethical limitations in social
  science experiments gt We cannot manipulate many
  areas/conditions of human life

7
Power of Experimental Design

• Is it better to study these with a survey or
  experiment?
• Playing violent video games increases aggressive
  behavior
• Students who sit at the front of the classroom
  make better grades than those that sit in the
  back

8
Power of Experimental Design

• Consider recent changes that have been made on
  your campus (e.g., changes in graduation
  requirements or student life issues). Did the
  administration use experiments or
  quasi-experiments to determine whether or not to
  make these changes? Can you think of situations
  in which an experimental approach could have been
  used?

9
Parts of the Experiment

• 1. Treatment or independent variable (IV)
• 2. Dependent variable (DV)
• 3. Pretest
• 4. Posttest
• 5. Experimental group
• 6. Control group
• 7. Random assignment

10
Parts of the Experiment

• Treatment gt (stimulus, manipulation) the
  independent variable or a combination of
  independent variables in experimental research
• Dependent variable gt the outcome in experimental
  research. It is measured by paper-and-pencil
  tests, observation, interviews or physiological
  responses (e.g. heart beat, palm sweating)
• Pretest gt the measurement of the independent
  variable prior to the introduction of the
  treatment.

11
Parts of the Experiment

• Posttest gt the measurement of the independent
  variable after the treatment has been introduced
  into the experiment situation
• Experimental group gt the group that receives the
  treatment or in which treatment is present
• Control group gt the group that does not receive
  the treatment

12
Random Assignment

• Participants divided into groups at the beginning
  of experimental research using a random process
  so the experimenter can treat the groups as
  equivalent
• Randomness in statistical or mathematical sense
  All participants have an equal chance of ending
  up in one or the other group
• Increases our confidence that the groups do not
  differ in any systematic way

13
Random Assignment
14
How to Randomly Assign?
15
(No Transcript)
16
Controls in Experiment

• By controlling confounding effects and isolating
  the effects of the treatment, we eliminate
  alternative explanations
• Deception Sometimes we intentionally mislead
  participants to control the experiment setting
• Deception involves using confederates and cover
  stories (as false explanations)
• Using placebo simulated or ineffectual treatment
  intended to deceive the participants
• Blind studies Single blind double-blind

17
(No Transcript)
18
Steps in Experiment (1)

• Hypothesis
• Choose a design
• Design experiment
• How to introduce IV
• How to measure DV
• Locate subjects
• Randomly assign subjects

19
Steps in Experiment (2)

• Gather pretest data
• Run experiment
• Introduce treatment
• Measure DV
• Gather posttest data
• Debrief
• Analyze data

20
Types of Experimental Design

• Classical Experimental Design
• Includes random assignment, a control group and
  experimental group, and a pretest and posttest
  for each group
• Pre-Experimental Design
• Lack control group and random assignment
• Weaker than classical experiment
• Substitude classical experiments when they are
  not possible

21
Types of Experimental Design

• Quasi-Experimental Design
• We call them quasi (apparently, as if) becuse
  they are variations of the classical design
• Some have control group and randomization but
  lack a pretest

22
Types of Experimental Design
23
Classical Experimental Design

• Figure 8.6 Pretest-posttest control-group
  design.

24
Classical Experiment Example

• You have 40 newly hired waiters you instruct
  them not to introduce themselves with first name
  and not to return during the meal to check on the
  customers
• You randomly divide the waiters into two groups
  of 20 persons (random assignment) and send each
  group to one of two restaurants to begin working.
  
• You record the amount of tips for all
  participants for one month (pretest score)

25
Classical Experiment Example

• Next, you retrain the group at restaurant A
  (experiment group) and instruct them to introduce
  themselves with first name and ask Is everything
  fine? 10 minutes after delivering the food
  (treatment).
• You remind the group at restaurant B (control
  group) to continue as before.
• Over the second month, you record the amount of
  tips for both groups (posttest score)

26
Internal Validity

• Occurs when the independent variable, and nothing
  else, influences the dependent variable.
• Anything other than independent variable
  threatens internal validity
• Artifacts gt unwanted or confounding variables
  that are due to the particular experimental
  arrangement

27
Threats to Internal Validity

• Selection bias
• A bias that arise when groups in an experiment
  are not equivalent with regard to the DV
• History
• Something occurs and affects the DV during an
  experiment is unplanned and outside the
  researchers control
• Maturation
• Natural processes of growth, boredom, fatigue
  that occur during the experiment and affect DV

28
Threats to Internal Validity

• Testing The very process of measuring, the
  pretest measure itself has an effect on DV
• Instrumentation Occurs when the instrument or
  measure of the DV changes during the experiment
• Experimental mortality
• Participant fail to participate throughout the
  entire experiment they leave the experiment
• Demand characteristics
• Participants guess the study hypothesis and
  respond to what they think the experimenter
  demands from them

29
Threats to Internal Validity

• Diffusion of treatment
• The treatment spills over from the experimental
  group and control group participants modify their
  behavior because they learn of the treatment
• Compensatory behavior
• When participants in the control group modify
  their behavior to make up for not getting the
  treatment
• Experimenter expectancy
• Experimenter indirectly makes participants aware
  of the hypothesis or desired results

30
External Validity

• The ability to generalize experiment findings
• It addresses two major questions
• 1. Can we generalize from the specific collection
  of participants in one experiment to an entire
  population gt Population generalization
• 2. Can we generalize from what occurs in a highly
  controlled and artificial setting to most
  natural, real world situations gt Naturalistic
  generalization

31
External Validity
32
Field Experiments

• An experiment that takes place in the field a
  natural setting instead of an artifical one
  (laboratory)
• They have lower internal validity but higher
  external validity than laboratory experiments
• Participants are usually unaware that they are
  involved in an experiment and react in a natural
  way
• Example A confederate fakes a heart attack on
  the subway to see how the bystanders react

33
Field Experiments

• Helping behavior and gender
• http//www.youtube.com/watch?vBuyIOM4d-OAfeature
  related

34
Practical Considerations

• Planning and pilot testing
• We anticipate alternative explanations or threats
  to internal validity during a good planning phase
• Instructions to subjects
• Preparing instructions carefully so that all
  participants understand the exact same thing
• Post-experiment interview
• Ethically debriefing research participants the
  deception understanding what participants
  thought and felt during the experiment

35
Online Resources

• This University of Denver site provides links to
  various online experiments in which students can
  participate.
• http//www.du.edu/psychology/methods/

36
Watch it ?

• Super Size Me
• Field experiment by Morgan Spurlock (2004)
• http//www.youtube.com/watch?v4Q-ezCOCVmgfeature
  relmfu