Thinking Scientifically in Everyday Life - PowerPoint PPT Presentation

Loading...

PPT – Thinking Scientifically in Everyday Life PowerPoint presentation | free to download - id: 81c8aa-YWYyO



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Thinking Scientifically in Everyday Life

Description:

Title: Science and Everyday Life Subject: Lecture 2a Author: Janina Jolley & Mark Mitchell Keywords: Validity Last modified by: Mark Mitchell Created Date – PowerPoint PPT presentation

Number of Views:48
Avg rating:3.0/5.0
Slides: 71
Provided by: Jani243
Learn more at: http://www.jolley-mitchell.com
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Thinking Scientifically in Everyday Life


1
Thinking Scientifically in Everyday Life
Quick links
2
Quote
  • The whole of science is nothing more than the
    refinement of everyday thinking.
  • --Albert Einstein

3
Refining Everyday Logic by Questioning 3
Validities
  1. External Validity
  2. Internal Validity
  3. Construct Validity

4
What is external validity?
  • The ability to generalize to a broad range of
    people and situations.

Larger world
Results from the study (a mini-world)
Generalize To
5
External Validity
  • External means beyond, outside, okay to use
    on the outside
  • External validity means that you can generalize
    the results to other
  • people,
  • places, and
  • (time) periods.

6
Why do peoples conclusions often lack external
validity?
  1. Using the false consensus effect to generalize
    to most people based on a sample of 1
    (themselves)

7
1. The False Consensus Effect (Sometimes, Its
Just You)
  • Definition Almost everyone does what I do and
    feels the way I do
  • Example 1 Perot voters thought Perot was going
    to win the 1992 Presidential election (Clinton
    won, Perot got only 19 of the vote and got no
    electoral votes).
  • Example 2 the extinct McDonalds Hula burger
    (only 1 sold--but Ray Kroc liked it).
  • Example 3 If people help, they think most
    people would help if they do not help, they
    think most people would not help.

8
Why do peoples conclusions often lack external
validity?
  1. Using the false consensus effect to generalize
    to most people based on a sample of 1
    (themselves)
  2. Basing conclusions on samples that are not
    representative because samples are small or
    biased.

9
2. Problems With Nonrepresentative Samples
  • Shere Hites study (nonrepresentative sample) 7
    of women happily married.
  • Most national surveys (representative sample)
    90 of women happily married.

10
Dangers of generalizing from small samples
  • Depression Overgeneralizing from a few failures.
  • Stereotyping Overgeneralizing from a few people
    Theyre all like that.
  • Foolishness Being convinced by single examples
    believing in stories over statistics. I know a
    guy who (smoked and lived to be 105, was cured
    by being thrown into the river) Remember, almost
    anything happens sometimes (there are people who
    won millions playing the lottery)but we want to
    know is what will probably happen.

11
You might question a studys external validity if
  • Only rats were studied
  • Only men were studied
  • Most of the older adults dropped out of the study
  • The study was a 1950s survey of peoples
    attitudes toward women in the work place
  • The study involved only 6 participants

12
Achieving External Validity
  • External validity is an issue because
    individuals, settings, and times differ.
  • Solution Study a variety of individuals in a
    variety of settings at different times.
  • Hard to do in one study Easier to do with
    several replications of a study.

13
What is internal validity?
  • Having a study that allows us to determine
    whether, within that study, a factor causes a
    certain effect.
  • causes makes, leads to, determines, produces,
    influences, affects, changes, (increases or
    decreases).

14
Why do we care about causes?
  • Knowing about causes leads to understanding
  • All explanations are causal explanations.
  • Knowing causes allow us to answer why questions
    by saying because...
  • 2.

15
Why do we care about causes?
  • Knowing about causes leads to understanding
  • Knowing causes can lead to cures (control)

16
Why do we care about causes?
  • Knowing about causes leads to understanding
  • Knowing causes can lead to cures (control)
  • Until you know what is causing the problem, it is
    hard to solve the problem.

17
Why do we care about causes?
  • Knowing about causes leads to understanding
  • Knowing causes can lead to cures (control)
  • Until you know what is causing the problem, it is
    hard to solve the problem.
  • Psychologists want to change things for the
    better, and that requires knowing the causes.

18
Why do we care about causes?
  • Knowing about causes leads to understanding
  • Knowing causes can lead to cures (control)
  • Until you know what is causing the problem, it is
    hard to solve the problem.
  • Psychologists want to change things for the
    better, and that requires knowing the causes.
  • If we misdiagnose the cause (we think reading
    problems are due to sporadic eye movements), we
    will not cure the problem False causes lead to
    false cures.
  • Thus,

19
Why do we care about causes?
  • Knowing about causes leads to understanding
  • Knowing causes can lead to cures (control)
  • Until you know what is causing the problem, it is
    hard to solve the problem.
  • Psychologists want to change things for the
    better, and that requires knowing the causes.
  • If we misdiagnose the cause (we think reading
    problems are due to sporadic eye movements), we
    will not cure the problem False causes lead to
    false cures.
  • Thus, causal relationships are not casual
    relationships. Instead, causal relationships tell
    us that if we do this one thing, this second
    thing will probably result.

20
Why is it hard to identify causes?
  • Illusory correlations We may see a change that
    is a mirage made up by our mind or by random
    events seeing a relationship that does not
    really exist.
  • 3 Examples

21
Why is it hard to identify the cause?
  • Illusory correlations We may see a change that
    is a mirage made up by our mind or by random
    events.
  • 3 examples
  • 1. The Sports Illustrated Jinx
  • 2. The hot hand phenomena in
    basketball
  • 3. Belief that, on multiple-choice tests,
    students usually change right answers to wrong
    answers.

22
  • Teams and individuals who are on the cover of SI
    actually do better after being on the cover.
  • Players are actually less likely to make the next
    shot after making one.
  • More answers are changed from wrong to right than
    from right to wrong.

23
Why is it hard to find the cause?
  • We may mistakenly see relationships that dont
    exist Illusory correlations.
  • Finding a real connection between two variables
    is not conclusive evidence that a particular
    variable had an effect because
  • We may not have established which variable came
    first. So, we dont know which is the cause and
    which is the effect (the chicken-egg problem).
  • We dont know whether both variables are side
    effects (symptoms) of some other (third)
    variable We may not have ruled out other
    potential causes.

24
Review Why doesnt showing a relationship mean
there is a causal relationship?
1.
2.
Third variable problem
25
Example of Point 1 Effects may be causes
  • Listening to country music and depression?
  • Authoritative parents and well-behaved children?
  • Self-esteem and success?

26
Example of Point 2Both factors may be side
effects or symptoms of another variable that is
the cause
  • Baldness and length of marriage
  • Elementary school childrens weight and number of
    questions correctly answered on a vocabulary test

27
Problems with 3 types of studies that are
foolishly trusted to identify causes
  • 1. Problem with before-after reasoning Person
    may change even without treatment.

I always get better, even without the
leeches. --Woody Allen If
you have a cold and go to the doctor, you will
get better in a week. If you dont, it will
take you seven days.
Plants and people grow.
People vary over time.
28
Problems with 3 types of studies that are
foolishly trusted to identify causes
  • 2. Problem with comparing natural groups
    Groups may differ on a variety of variables.
  • Smokers eat differently than non-smokers.
  • Heavy drug users differ from non-users even
    before they start using drugs.

Dont compare apples with oranges.
In this class, can you pick out two identical
groups?
29
Problems with 3 types of studies that are
foolishly trusted to identify causes
  • 3. Problem with self-report People dont know
    why they do things Why ask why?
  • Young children often cant provide reasons for
    their behavior I dont know
  • Nisbett Wilsons (1975) research shows adults
    can provide reasons, but the reasons may be
    wrong.

30
How can we design a study that will help us
isolate a cause?
  • Experiments Special types of studies that
  • Solve the chicken-egg problem by introducing the
    treatment first to see if an effect follows.
  • Solve the other variable problem by
  • Isolating participants from other variables
    (difficult to do completely, people usually cant
    live in bubbles)

31
How can we design a study that will help us
isolate a cause?
  • Experiments Special types of studies that
  • Solve the chicken-egg problem by introducing the
    treatment first to see if an effect follows.
  • Solve the other variable problem by
  • Isolating participants from other variables
    (difficult to do completely)
  • Randomizing other variables to prevent those
    variables from having a systematic effect and
    then using statistics to estimate (subtract out)
    their unsystematic effects.

Like a blender, randomization helps spread
things around.
32
Review External Validity vs. Internal Validity
  • Suppose you do a study. In that study, you
    observe that people who smile get helped more
    than people who dont.

33
Review External Validity vs. Internal Validity
  • In your study, suppose you observe that people
    who smile get helped more than people who dont.
  • If you have internal validity, you can conclude
    that, in your study, smiling causes helping.

34
Review External Validity vs. Internal Validity
  • In your study, suppose you observe that people
    who smile get helped more than people who dont.
  • If you have internal validity, you can conclude
    that, in your study, smiling causes helping.
  • If you have external validity, can conclude that
    people who smile get helped more generalizes
    outside of your study to other samples and
    situations.

35
Review External Validity vs. Internal Validity
The Case of Random Sampling (RS)
Vs. Random Assignment
(RA)
  • To get internal validity, you could use random
    assignment by randomly dividing your participants
    into two groups and then smiling at participants
    in one group, but not smiling at participants in
    the other group.

36
Review External Validity vs. Internal Validity
The Case of Random Sampling Vs. Random Assignment
  • To get internal validity, you could use random
    assignment by randomly dividing your participants
    into two groups and then smiling at participants
    in one group, but not at participants in the
    other group.
  • Because the groups were probably similar to each
    other before you approached them, if the group
    that was smiled at ends up being more helpful,
    you could conclude that, in your study, smiling
    causes helping.

37
Population (larger group)
  • RA


RA goal Two slices arebefore treatmentalike
want to see whether treatment makes them differ.
Dont care about larger pie. Just want two chips
off the same block.
38
Review External Validity vs. Internal Validity
The Case of Random Sampling vs. Random Assignment
  • To get external validity, you could use random
    sampling by randomly selecting your participants
    from a larger group (e.g., everyone at your
    school).

39
Review External Validity vs. Internal Validity
The Case of Random Sampling Vs. Random Assignment
  • To get external validity, you could use random
    sampling by randomly selecting your participants
    from a larger group (e.g., everyone at your
    school or everyone in your town).
  • Because the group you studied was probably
    similar to the larger group you sampled from, you
    could probably conclude that your finding that
    people who smile get helped more generalizes to
    the larger group you sampled from.

40
Review External Validity and Random Sampling (RS)
RS goal Slice is like whole pie want to know
about larger whole.
41
Review External Validity vs. Internal Validity
The Case of Random Sampling Vs. Random Assignment
  • You could get both external validity and internal
    validity by using both random sampling and random
    assignment.

42
Review External Validity vs. Internal Validity
The Case of Random Sampling Vs. Random Assignment
  • You could get both external validity and internal
    validity by using both random sampling and random
    assignment.
  • You could use random sampling to select your
    participants
  • (by making it likely that your participants
  • represent the group you sampled from,
    you
  • could generalize your findings to that
    larger
  • group).

43
Review External Validity vs. Internal Validity
The Case of Random Sampling Vs. Random Assignment
  • You could get both external validity and internal
    validity by using both random sampling and random
    assignment.
  • You could use random sampling to select your
    participants
  • (by making it likely that your participants
    represent the group you sampled from, you could
    generalize your findings to that larger group).
  • Then, you could use random assignment to assign
    those participants to either receive or not
    receive smiles
  • (by making it likely that your smile and
    no-smile groups
  • were similar to each other before the
    smiling
  • manipulation, you could conclude
    that differences
  • between them in helpfulness after the
    smiling
  • manipulation was caused by the smiling
    manipulation).

44
Review External Validity Vs. Internal Validity
Random Sampling (RS) Vs. Random Assignment (RA)
RS goal Slice is like whole pie want to know
about whole.
RA goal Two slices arebefore treatmentalike
want to see whether treatment makes them differ.
45
Review External Validity Vs. Internal Validity
Random Sampling (RS) Vs. Random Assignment (RA)
RS goal Slice is like whole pie want to know
about whole.
RA goal Two slices arebefore treatmentalike
want to see whether treatment makes them differ.
46
What is Construct Validity?
  • Definition the degree to
  • which a study actually
  • measures and
  • manipulates the variables
  • that the researcher
  • claims to be measuring.
  • (laughing, smiling may
  • not equal happiness)

Happiness
smiling
47
  • If you have construct validity, you really are
    measuring and manipulating the psychological
    constructs that you claim you are. In other
    words, we can believe the labels you put on your
    variables.

48
  • Labels dont always deserve your unquestioning
    trust .

Green beans
100 pure spring water
49
3 examples of not questioning labels
  • People are quick to give labels to others (e.g.,
    stupid, paranoid)but slow to question those
    labels.
  • Is the person avoiding eye contact
  • Shy?
  • Stuck up?
  • Lying?

50
Examples of not questioning labels
  • 1. People are quick to give labels to others
    (e.g., stupid, paranoid)but slow to question
    those labels
  • 2. People are quick to accept labels
  • They believe tests (Couple compatibility) test
    they find in magazines
  • They often read a summary of a study on stress
    but dont even ask what procedures were used to
    manipulate or measure stress.
  • 3.

51
Examples of not questioning labels
  • 3. People dont realize that manipulations and
    measures are rarely 100 pure manipulations
    often contain additional perceived and actual
    ingredients (e.g., the effect of a treatment may
    be due to placebo).

For pills, even color affects responses.
Contains more than caffeine
52
Why do peoples conclusions often lack construct
validity?
  • We can only infer that we have measured or
    manipulated mental constructs.
  • These inferences can be wrong. Thus, although the
    evidence may not lie, it can be misleading or
    misinterpreted
  • (Monty Python clip).

53
3 main reasons inferences about mental states are
often wrong
  • Observers are not cautious about labeling
    behavior.
  • Snap judgments about traits in everyday life
  • 2 Examples from psychiatry
  • patient exhibits bizarre writing behavior

54
3 main reasons inferences about mental states are
often wrong
  • Observers are not cautious about labeling
    behavior.
  • Snap judgments about traits in everyday life
  • 2 Examples from psychiatry
  • patient exhibits bizarre writing behavior
    (Psychiatrists report of a grad student
    writing a letter home to his mother. The student
    was a pseudopatient in Rosenhans On being sane
    in insane places study)

55
3 main reasons inferences about mental states are
often wrong
  • Observers are not cautious about labeling
    behavior.
  • Snap judgments about traits in everyday life
  • 2 Examples from psychiatry
  • patient exhibits bizarre writing behavior
    (Psychiatrists report of a grad student
    writing a letter home to his mother. The student
    was a pseudopatient in Rosenhans On being sane
    in insane places study)
  • Patients have reached nirvana, heaven on earth

56
3 main reasons inferences about mental states are
often wrong
  • Observers are not cautious about labeling
    behavior.
  • Snap judgments about traits in everyday life
  • 2 Examples from psychiatry
  • patient exhibits bizarre writing behavior
    (Psychiatrists report of a grad student
    writing a letter home to his mother. The student
    was a pseudopatient in Rosenhans On being sane
    in insane places study)
  • Patients have reached nirvana, heaven on earth
    (1950s report on the effect of a new major
    tranquilizer Patients were asleep)

57
3 main reasons inferences about mental states are
often wrong
  1. Observers are not cautious about labeling
    behavior.
  2. Those being observed may behave in a way that
    masks their true mental state.

58
People may mask their true mental state
  • Social desirability bias not responding like one
    really would, but instead responding in a way
    that would make a better impression.
  • Lying can be a big problem in surveys

59
Placebo effects
  • Definition people will believe they have been
    helped if they expect the treatment to work. They
    are being changed by their expectations of what
    the treatment will do, rather than whats in the
    treatment itself.
  • Examples Pills, doctor visits, psychotherapy,
    subliminal tapes

60
How do scientists avoid making construct validity
errors that people normally make?
  • Ask whether we could be wrong.
  • Collect evidence to see whether we are wrong.
  • Reduce researcher biases by using blind
    techniques
  • Make participants unlikely to lie or act.
  • Make the results anonymous so no motivation to
    lie.
  • Use blind techniques so no ability to give
    desired responses.
  • Make the study interesting and involving, so
    participants react to this psychologically real
    experience rather than play act a role.

61
Review
  • Headline
  • Stress makes Americans aggressive.
  • Validity
  • makes causes internal validity
  • Americans generalizing beyond participants
    studied external validity
  • Stress, Aggression accuracy of these labels
    construct validity

62
Review
Validity Allows us to Is hard to achieve because
Construct Trust variable names Cant see constructs Participant and researcher bias


63
Review
Validity Allows us to Is hard to achieve because
Construct Trust variable names Cant see constructs Participant and researcher bias
External Generalize results Results that hold with one group in one situation may not hold with other groups in other situations.

64
Review
Validity Allows us to Is hard to achieve because
Construct Trust variable names Cant see constructs Participant and researcher bias
External Generalize results Results that hold with one group in one situation may not hold with other groups in other situations.
Internal Make cause-effect statements Even if find that A and B are related, may not know whether A?B, B?A, or some C causes both.
65
Type of Validity Example of Leap Making the leap
Construct Manipulation of construct construct Thinking we see thoughts. Valid measures manipulations, Blind techniques
66
Type of Validity Example of Leap Making the leap
Construct Manipulation of construct construct Thinking we see thoughts. Valid measures manipulations, Blind techniques
External Applies to my participants, so it applies to everyone. Going from lab to life. Replication of study with different groups in different settings, Random sampling
67
Type of Validity Example of Leap Making the leap
Construct Manipulation of construct construct Thinking we see thoughts. Valid measures manipulations, Blind techniques
External Applies to my participants, so it applies to everyone. Going from lab to life. Replication of study with different groups in different settings, Random sampling
Internal Theres a relationship between A and B, A must cause B. Experiments using random assignment
68
Summary Questions to Ask
If there is evidence, is it any good?
Construct validity (if care about names you are
giving the variables)
External validity (if generalizing)
Internal validity (If making cause- effect claims
)
Assignment
69
Group Review
  • Form 3 groups Internal, External, and Construct
    Validity group
  • Decide how you will explain your validity
  • 1st time through Explain your validity to the
    other groupsand why your validity is important.
  • 2nd time through Explain threats to your
    validity.
  • 3rd time through Explain how your validity
    differs from other validities.

70
Assignment
  • Use what you know about the three validities to
    criticize
  • Testimonials
  • Unprofessional polls

71
Chapter 2
Chapter Websites Quick Visuals Quizzes Video
Teacher Chapter outline Description and Explanation Quiz An invalid witch test (from Monty Python)
Student Concept Map Validity Quiz
Chapter Summary Ethics Quiz
About PowerShow.com