Transcript AP Psychology Chapter Two

Chapter 2 Research Methods
ETHICS IN RESEARCH
APA GUIDELINES:
 Research must be approved by ethics
board or institutional review board (IRB)
 If research doesn’t meet APA criteria, the
researcher will be asked to revise and
resubmit work.

Animal Research-APA
1.
2.
3.
4.
Clear Scientific Purpose
Humane Care
Legal Acquisition
Minimal Suffering in Experiment
Human Research
1.
2.
3.
4.
5.
6.
Informed Consent
Deception
Coercion
Anonymity
Risk
Debriefing
Regardless
of the method
used, all research is based
on the Scientific Method of
Psychology
Scientific method(s)
 Scientific
means systematic,
testable, and objective.
Essential Question
What are the three
main principles that
guide the Scientific
Method of
Psychology?
Simplified Scientific Method
1 – Theory
Step 2 – Hypotheses
Step 3 – Research and
Observation
Step
Step 1
Theories
organize known
facts and summarizes
current research in the field.
What do we already know?
Step 2
A
hypothesis is then created
as a testable prediction based
on what is currently known and
what we want to find out.
What do we want to know
more about?
Step 3
 Research
or observation or
experiments are generated to
collect data, which then goes into
evaluating the hypothesis, which
may or may not add to the existing
theory. What did we find out?
What are some
examples of
research methods?
I. Naturalistic Observation
Study
behavior in its
natural context.
Spontaneous behavior in a
subject’s natural
environment. No
interaction with the subject.
 IE.
If you want to study the
interactive behavior of a specific
breed of gorillas, you would need to
go to where the gorillas live in
nature (not a zoo). You would need
to observe them without their
knowledge, and without
manipulating anything.
II. Case Study


A case study is one type of
observational data collection technique
in which one individual is studied indepth in order to identify behavioral,
emotional, and/or cognitive qualities
that are universally true, on average, of
others. Case studies often include faceto-face interviews, paper and pencil
tests, and more.
Why use a case study?
IE.
I want to know why Bart
killed thirty-five people over a
twenty-year period of time. I
will examine the police files,
observe and interview Bart,
talk to his and the victims
families, etc.
III. Survey
Questionnaires/
interviews. Getting a
large amount of
information from a large
group of people.
When Creating A Survey
Questions need precise answers
 Language and wording must be simple


IE. 77% of New Yorkers were interested in
plants and trees, but only 39% were
interested in botany; 48% were interested in
fossils, but only 39% were interested in
paleontology; 42% were interested in rocks
and minerals, but 53% were interested in
Geology
Wording Is Key
Example: “abortion” or “Pro-Choice”
 “unborn child” or “fetus”
 Wording must be as neutral as possible
 Groups conducting research should be
neutral-avoid confirmation bias

When Creating A Survey
Ask questions that won’t embarrass or
humiliate
 Responders will lie if there is a perceived
punishment or reward– Anonymity is key
 Don’t ask morally ambiguous questions – keep
it simple and to the point
 Who the interviewer is will affect the
responders answers-gender, affiliation, etc.expectations?

When Creating A Survey

Shortly phrased questions.

IE. As you know, the term Holocaust usually
refers to the killing of millions of Jews in
Nazi death camps during WWII. Does it
seem possible or does it seem impossible to
you that the Nazi extermination of the Jew
never happened?

1 out of 5 Gallup poll responders said that the
Holocaust never happened due to the phrasing of the
question
When Creating A Survey
Hot
 IE.
Topics/Key Words:
Do you favor an
amendment prohibiting
abortions? >50% opposed OR
Do you favor an amendment
protecting the life of an unborn
child? <30% opposed it
When Creating A Survey
Limited Answer Options
 Order of Questions – easier to more
difficult works best
 Fright Terms – avoid using terms with
big repercussions
 IE. Problem V. Crisis, Passed away V.
Dead, Dealt With V. Punish

False Consensus Effect
 Tendency
to overestimate the
extent to which others share
our beliefs and behaviors.
Skews the reports by jumping
to large conclusions that fit
into our pre-conceived ideas.
IV. Experiment
 An
investigation seeking to
understand relations of cause and
effect. The experimenter changes a
variable (cause), and in turn
changes another variable (effect).
At the same time the experimenter
hopes to hold all of the other
variables constant so that they can
attribute any changes to the
manipulation.
How do I create a
valid and reliable
experiment?
Step 1: Choose a Hypothesis
A
hypothesis expresses a
relationship between two variables.
 IE.
My hypothesis is that watching
violent television programs makes
people more aggressive.
Step Two: Choose Variables
 Variables
are things that are
measured, controlled, or
manipulated in research.
The
independent variable is
the manipulated variable
.
IE.
Watching television
violence is the independent
variable because I can adjust
what shows are viewed, for
how long, by whom, etc.
 The
dependent variable is
measured for change.
 IE.
Measuring the change in
aggression levels is the dependent
variable in our experiment because
it changes based on what is viewed,
for how long, etc.
Step Three: Operationalize

When you operationalize your variables,
you are explaining how you will measure
them.

IE. The operational definition of the
independent variable (what defines a violent
show?) would be shows that have scenes of
fighting, bloodshed, use of weapons, injury,
kicks, punches, etc.
 IE. The operational definition of the
dependent variable (what constitutes an
increase in aggressive behavior?) would be an
increase in agitation or tenseness, increased
vocal volume, threats of bodily harm, kicks,
punches, throwing objects, etc.
Step Four: Identify Potential Extraneous
Variables/Confounding Variables

It is important to make sure that
during the experiment as many other
factors that are NOT part of the
therapy are NOT included. Any factor
or variable that causes an effect (or
potential affects) other than the
variable being studied is considered
an extraneous variable.

IE. An extraneous variable in our
experiment would be a phone call from a
solicitor during a program, the viewer
receiving mail including a poor report
card, a viewer stubbing their toe during a
show, alcohol abuse, etc. All of these
could increase aggressiveness, but are
not related to viewing violent television.
Step Five: Identify Who You Will Be
Testing
The individuals on which the research
will be conducted are called subjects (or
participants).
 A small group of subjects are drawn from
a larger potential population.


IE. Our subjects will be drawn from the
overall population of 12th grade males at
Phoenixville Area High School.
Step Six: How Do We Decide Who
Will Be Subjects, and Who Won’t?
 Since
we can’t realistically test all
12th grade males at Phoenixville
Area High School, we have to
create a representative sample of
the population so that we can
generalize our findings to the whole
group.
Method 1: Rigorous Control Design
 Designing
an experiment with
specific, hand-picked groups in
mind.
 IE. Only testing males, 18 years
old, in AP Psychology.
Method 2: Sample Design
A
sample is a representation of
the entire population.
 A random sample allows that
every member of an overall
population has an equal chance
to be in the sample.
 IE. Drawing names from a hat.
Method 3: Stratified Sample
Subdivide the population into at least two
different subpopulations that share the
same characteristics, then draw a
random sample from each group.
 IE. Surveying views on Equal Rights.
Split your human population into men
and women, and then randomly draw
eight men’s and women’s names.

Method 4: Systematic Sample
Select a starting point from your
population and then select every ?th
participant.
 IE. Merck corporation wants test the
effectiveness of a new aspirin on their
100,000 employees. Get a roster of
employees, start at #1, and then choose
every 100th name on the list.

Method 5: Cluster Sampling
Divide your population into multiple
subgroups, randomly choose a subgroup
to test, and then test the entire
population of that subgroup.
 IE. Split the country up into geographic
regions (East Coast, Midwest, etc.).
Randomly choose a region to test, and
then test every person in those states.

Method 6: Convenience Sampling
Use
a population that is
readily available.
IE. Test your neighbors,
your family, your coworkers, a passer-by on
the street.pd1 stop
Step Seven: Assignment
 Once
you have chosen your
subjects to study, you must
assign them to one of two
groups; those that will be
manipulated, and those that
won’t.
Group 1: Experimental Group
The
experimental group
receives the independent
variable and is
manipulated throughout
the experiment.
Group 2: Control Group
The
control group
does not receive
the independent
variable.
 IE.
In our television violence
experiment, those in the
experiment group will watch
varying degrees of violent
program, for varying lengths of
time, etc., and their changes in
levels of aggression measured.
 IE.
In our television violence
experiment, the control group
will be shown a variety of nonviolent programming in order to
create a baseline to compare
the experiment group against.
Steps to follow for an experiment








1. choose hypothesis
2. choose variables
3. operationalize
4. identify confounding variables
5. identify subjects/population to be tested
6. Who will be tested? Sampling
7. Assignment
8. Address other issues/bias/flaws
Method 1: Random Assignment
 Random
assignment means that the
subjects have an equal chance of
being placed into each group. If we
allow subjects to choose their own
group, we may have a subjectrelevant confounding variable.
Subject-Relevant Confounding
Variables
A
subject-relevant confounding
variable would allow those people
that liked violent movies or were
prone to violence already to choose
to be in the experimental group. We
therefore could not accurately find
that viewing violence led to
aggression.
To
help avoid this
confounding variable, we
prescribe a single-blind
design. The subjects do not
know whether they have
been randomly placed in the
control or experiment group.
Method 2: Group Matching
 When
assigning members to
the experiment or control
group, it is important that the
characteristics of both groups
need to be as similar as
possible.
 IE.
After rigorously or randomly
determining our subjects, as many
white, black, tall, short, overweight,
slim members should be in the
control group as there are in the
experiment group.-easy if it is a
large group. Difficult if small
Step Eight: Address Other Potential
Issues With Experiments


Situation-relevant confounding variables
refer to making sure that the situations that
the experiment and control groups are
placed in are exactly the same. We must
have equivalent environments.
IE. We cannot have those watching violent
films in a large auditorium, and those
viewing sitcoms in a cramped, warm, loud,
smelly, or otherwise unpleasant
atmosphere.
Experimenter Bias

Experimenter Bias occurs when the
experimenter unconsciously treats members of
the control and experiment groups differently,
which increases the chances of confirming
their hypothesis.

IE. The experimenter gives soda to the control
group, and beer to the violent viewers. The
experimenter speaks more abruptly with the
violence crowd (inciting them?).

To help avoid this type of
confounding variable, we employ
a double-blind design, where
neither the subjects nor the
researcher may know which is
the control or the experiment
group. A third-party has the
appropriate records so that the
date can be analyzed later.
 The
Hawthorne Effect refers to
the fact that some subjects will
alter their behaviors simply
because they know that they are
part of an experiment,
regardless of what is being done
to them.
 The
Placebo Effect refers to the
phenomenon that a patient's
symptoms can be alleviated by
an otherwise ineffective
treatment, apparently because
the individual expects or
believes that it will work.
 Hindsight
Bias is the tendency to
believe, once the outcome is
already known of course, that you
would have foreseen it…that even
though it's over and you know the
outcome, you knew it all along.
 Overconfidence
occurs when we
tend to think that we know more
than we do. -generalizations when
reporting results, forcing results
into preconceived hypothesis to
say “I told you so”, instead of
letting the results speak for
themselves, etc.
REMEMBER !!!
Only experimental data can
conclusively demonstrate
causal relations between
variables
(A causes B to happen).
V. Correlation Study
A
correlation study is a
statistical measure of the
relationship between two or
more variables without
assigning a cause and effect
relationship.
Correlation Study
You
cannot conclude a
cause and effect relationship
from this statistical analysis;
you can only imply or predict
a strong or a weak
relationship between
variables !!!
 IE.
Student scores on the SAT are
collected, as are senior year GPA’s.
We want to see if a high GPA
correlates to a high SAT score. We
can’t say one causes the other, but
we can imply that students who
have high/low GPA’s score high/low
on SAT’s. Can we use GPA as a
predictor of SAT performance?

-1 to +1

stop
Once I have
completed my
research, how do I
present my findings ?
Skip to 93 graphing
data
start
Descriptive Statistics
Descriptive
Statistics
describe a set of data.
If I want to report findings from an
observation or a survey, I may want to
use a frequency distribution.
 A frequency distribution may be a simple
chart, a list, or a graph.
 A graph of information always plots the
frequency along the y-axis, and the
subject of the graph along the x-axis.

Measures of Central Tendency

Measures of central tendency provide
statistics that indicate the average or typical
score in the distribution. There are three
measures of central tendency:
Mean
Median
Mode
Mean

The mean is the arithmetic average of all
the scores in the distribution. It is
calculated by adding all the scores in the
distribution and then dividing this sum by
the number of scores.
Median

The median is the middle score of the
distribution, the point that divides a rankordered distribution into halves
containing an equal number of scores.
Thus 50% of the scores lie below the
median and 50% lie above the median.
Mode

The mode is simply the score in the
distribution that occurs most frequently.
The Bell curve- “normal range”
Graphing Measures of Central
Tendency
 When
graphing the mean,
median and mode of a
distribution, roughly speaking, a
distribution has positive skew if
the right tail is longer and
negative skew if the left tail is
longer.
Positively Skewed

This distribution has a positive skew.
Note that the mean is larger than the
median.

IE. In a neighborhood of relatively low
incomes, a few millionaires move in.
Those few high salaries will inflate the
mean (average), but the median will
remain relatively low.
Negatively Skewed

This distribution has a negative skew.
The median is larger than the mean.

IE. In a particular well-to-do
neighborhood, a few low-income
residents move in. The overall average
income will drop a bit, but the median will
remain relatively high.
Measures of Variability

Measures of variability show how
spread out the distribution of scores
is from the mean, or how much
dispersion or scatter exists in the
distribution. If there is a large degree of
dispersion, that is, if the scores are very
dissimilar, we say the distribution has a
large or high variability, or variance. If
the scores are very similar, there is a
small degree of dispersion and a small
variance.
Measures of Variability
Range
Standard
Deviation
Range

The range is simply the numerical
difference between the highest and
lowest scores in the distribution.
Standard Deviation

The measure of variability used most
often in research is the standard
deviation, a statistic that indicates the
average distance of the scores from the
mean of the distribution.
stop

IE. Our class took Unit Exam 2. I
scored a 76%. I want to know how well I
did in relation to the rest of the class to
see whether or not that score was good
or bad. I need to figure out what the
class average was, figure out the
standard deviation from the mean, and
I’ll know how well I did.
Standard Deviation
Graphing Standard Deviation
Find the mean of your distribution set.
 Calculate the SD on your calculator.
 The mean is set at “0”.
 +1 and -1 are your SD above and below
the mean.

IE. Your mean is 56 with a SD of 6. +1
would be 62, and -1 would be 50.
 Calculate +- 2 and +- 3 in the same manner.

Graphing Standard Deviation

What does this tell us?
If the mean of a set of class scores on a unit
exam was 72, with a SD of 8, 68% of
students scored between a 64 and an 80.
Your score of a 76 would be close to being
better than 68% of the rest of the class.
 Approximately 95% of the class scored
between a 56 and an 88. Your score of a
50 would indicated that roughly 96% of the
class did better than you on the test.

Graphing Data

Scatterplot – A graphed cluster of dots
that represent the values of two
variables.
The
SLOPE of the points
suggests whether there is
a positive, negative, or
non-existent relationship
between two variables.

POSITIVE CORRELATION – as one set
of scores increases, so does the other

NEGATIVE CORRELATION – as one
set of scores goes up, another set goes
down

NO CORRELATION
How
closely the dots are to
each other along the line
indicates the strength or
weakness of the
correlation as well
 No
correlation relates to a score of
0.00
 A positive correlation ranges from
0.00 to 1.00
 A negative correlation ranges from
0.00 to -1.00 f
 Difference between positive and
negative?

Worksheet/activity
No more notes please!!