Transcript Methods of Research PP

Methods of Research
Psychology as a Science
 Uses
the Scientific Method:
appropriately identifies and
frames a researchable event
 Can help identify fallacies in
thinking
 Become critical of psychological
findings; don’t accept
everything
Good Psychological Science
Don’t accept ideas on
faith or authority.
 Reliance on Empirical Evidence:
Needs to be serious evidence based
on careful observation or
experimentation.
 Precision: Research based on a
scientific hypothesis and testing a
specific theory.
 Skepticism:
Good Psychological Science
 Openess:
Willing to tell others
where they got their ideas,
how they were tested, and
what the results were so that
they can be replicated.
 Willingness to make a risky
prediction: Must state ideas in
a way that they must be
testable.
Good Research must have…
 Hypothesis:
A statement about
the relationship between two
variables. Must be testable,
verifiable, and refutable.
 Operational definition: the
exact procedures/description
of the concept being tested
Testable, Verifiable, Refutable, Risky?
If you pick any day
of the year, a
famous person
was born on it.
Experiments
Experiments are the only
research method capable of
showing cause and effect
because experimenter can
manipulate factors and
control others.
Experiments
 Independent
Variable:
Variable manipulated by
the experimenter, everything
else is held constant.
 Dependent Variable:
Measured variable
influenced by the
independent variable.
Experiments
Confounding Variable: Any
variable besides the
independent variable that
could influence the outcome of
the experiment. (Unwanted)
Controls: methods placed in the
experiment to keep
confounding variables to a
minimum
Experiments
 Subject:
the person on whom the
experiment is being done.
 Experimenter: the person
conducting the research (does not
have to be the researcher)
 Confederate: a person who acts as
a subject but is actually helping
the experimenter
Experiments
 Sampling:
a representative group
of a larger population
 Random Assignment: Selection
and assignment of participants of
subjects to groups through
random or chance procedure.
 Placebo: Drug with no medicinal
value, usually a sugar pill, need
not always be a drug, sometimes
a situation. (placebo effect)
Experiments
Control
Group: Group that
does not take part in the
critical part of the
experiment. Serves as a
comparison of results from
the experimental group.
Experimental Group: Group
that receives the treatment.
Experiments
Single-Blind
Study: The
subjects do not know to
which group they belong.
Double-Blind Study: Neither
the experimenter nor the
subjects know to which
group the subjects belong.
Good Research must have…
Theory:
Organized
explanation used to explain
or predict human behavior
after it has been tested
empirically.
• Useful if it effectively organizes a
range of observations, and
anyone can check.
Good Research Must Have…
Validity:
test that
measures what it is set
out to measure.
Reliability: test that yields
consistent results from
one time and place to
another.
Types of Research
Naturalistic
observation—
watching behavior in a
natural setting
Survey
method—using
questionnaires to poll large
groups of people
(interviews)
Types of Research
Experimental method—
investigating behavior
through controlled
experimentation, usually
in a lab setting
(psychological testing)
Types of Experiments
Clinical
method—studying
behavior in clinical settings
(case study)
Correlational method—
measuring behavior to
discover relationships
(testing, longitudinal, or
cross-sectional)
Correlational Studies
Statistical
technique used to
measure the strength and
nature of a relationship
between two variables.
CORRELATION
DOES NOT
PROVE CAUSATION
Correlation does not mean causation!
 Global
warming has increased in
the last 100 years. Pirating and
the pirate lifestyle has
significantly decreased in the
last 100 years.
 The lack of pirates has obviously
caused global warming.
 Are
you kidding me?!
Correlation does not mean causation!
 As
ice cream sales increase,
the rate of drownings increase.
 OBVIOUSLY, ice cream causes
drowning!
 Are
you still buying this?!
Correlation does not mean causation!

Compared to citizens of other countries,
Americans and Brits have
significantly more heart attacks.

Americans and Brits drink red wine.
Therefore, red wine must cause
heart attacks.

Nope, Italians drink more red wine and
suffer fewer heart attacks.
Statistics
Descriptive Statistics:
Statistics that organize
and summarize data,
frequently use graphs or
charts.
Statistics
Arithmetic Mean: the average of a
set of scores. Add all quantities
together and divide by the total
number of scores. Summarizes a
mass of data, does not examine
variation.
Outliers—extreme scores that
change or skew the mean
Statistics: central tendency and variability
 Mode:
The most frequently
occurring score in a set of
data.
 Median: The middle number in a
collection of data.
 Range: Differences between the
lowest and the highest scores
in a distribution of scores.
Measures of Variability
Normal curve—a bellshaped distribution, with a
large number of data in the
middle tapering to lower
scores on either side
Skewness—asymmetry in a
distribution of numbers


Positive skew—the majority of the scores on the left
side of the mean with the tail trailing to the right
(the mean is greater than the median and mode)
Negative skew—the majority of the scores appear on
the right side of the mean with the tail trailing to
the left (the mean is less than the median and
mode)
Inferential Statistics
 Statistics
that tell the
researcher the significance
of the data.
 Show how likely the data were
to have occurred by chance.
 Assist in making a statement
about relationship in
variables.
Statistical Measure
Variance:
How clustered or
spread out individual
scores are around the
mean.
Standard Deviation: The
average distance of scores
around the mean.
Significance Testing
Significance Testing—used to draw
conclusions about whole populations
based upon samples.
 Alternative (Research) hypothesis—
relationship is the result of a real
effect.
 Null hypothesis—is tested to account
for a chance; states that no
relationship exists between variables.

Significance Testing
Type I Error—seeing a statistical
difference when none is present.
(Rejecting the null hypothesis when
the null is true.)
 Type II Error—seeing no difference
when one does exist. (Accepting the
null when it is false.)
 PROBLEMS?

Statistical Measure
 P-Value:
a number from zero to
one that represents the
probability that an event occurred
by chance.
E.g. P=0.05, means 95 times out of 100 (or 95%) the
results will be similar from one test to the next.
 Statistically
Significant: A low
chance of an event occurring by
chance: P<0.05.
Measures of Variability
Z-score—a
number that tells
how many standard
deviations above or below the
mean a score is.
Formula
for computation:
Score – mean/standard deviation
Correlational Studies
 Illusory
Correlation: The
perception of a relationship
where none exists.
 Coefficient of Correlation:
Perfect positive correlation
+1.00, perfect negative
correlation -1.00, no
correlation 0.
Correlational Studies
 Positive
Correlation: high
values for one variable are
associated with high values
for the other variable. Highest
+1.0.
 Negative Correlation: High
values for one variable are
associated with low values for
the other variable. Highest
-1.0.
Types of Data
Nominal: Numbers that are
used to name or categorize.
–
–
–
driver’s license
numbers on sports
uniforms
gender (#1 female;
#2 male)
Types of Data
Ordinal: Numbers represent
serial position: greater or less
than.
–Class rank
–Age
–Baseball standings
Types of Data
Interval Scale: Consistent units
of measurement, equal spacing
between, allows for
mathematical operations.
No true zero point (arbitrary)
Thermometer, temperature
Can’t say 20 degrees is twice as hot
as 10 degrees; ratios don’t work!
Types of Data
Ratio Scale: Same consistent
units of measurement as in the
interval scale but with the added
property of a true zero point.
Compare scores in terms of
ratios.
–Four pounds is twice as heavy as
two.
–Time
–Length
Graphs
Frequency Distribution A
table that divides an entire
range of scores into a
series of classes and then
records the number of
scores that fall into each
class (page A-4)
Graphs
Pie Graph:
Circle in
which all
data is
represented
in the form of
percentages
pie chart
Females
Males
Graphs
Frequency Histogram: Bar graph
with scores on the horizontal axis
and frequencies on the vertical axis.
(page A-4)
Frequency Polygon: Line graph that
has the same horizontal and vertical
axis as the histogram. Each score
marked with a point and then
connected, can plot multiple data
sets. (page A-4)
Some data interpretation
 Find
the mean, median,
mode, and range for the
salaries of the Indians, the
Yankees, all baseball teams,
and the Browns.
 What conclusions and/or
correlations can you draw?