Sei sulla pagina 1di 4

Lecture #2

Three Strategies to Test Theories:


-Descriptive
-Correlational
-Experimental

Descriptive
(systematic observation)

Three Types:

1) Case Study (oldest descriptive method)


-Study one or more individuals in great depth
-e.g.Sigmond Freud
-studied people with psychological disorders in Vienna
-extrapolated on personality in general using info. from these people
-Problem: the person or people you select may be so abnormal that data based on
them is misleading
-Today: case studies are used with brain damage patients
-e.g. how strake victims speak provides information on the role of the
brain in language - from this we learned that language is mostly in
the left hemisphere and there are two separate areas involved, one
for speaking and one for understanding

2) The Survey
-Asks many people to report their behavior or opinions
-e.g. election surveys, sex on campus by the “Times”
-can be used to compare cultures
-Trying to make a statement about a large group of people
-You can’t usually ask all of the people you are interested in (called the
target population), so you get a subset of the population
-this subset is called a sample
-For the results to be reasonable, the sample must be random, meaning that
every individual in the population has an equal chance of being picked for
the sample
-If the sample is not random, you can get erroneous conclusions
-e.g. book Women in Love only used the data from women in
national women’s organizations and chose to respond to the
survey (~4%) and found that 70% of the respondents had had an affair
-when the sample was made more representative, the figure was
10%

3) Naturalistic Observation
-You go to the national environment and you watch and record behavior (you do
not interfere)
-e.g. Jane Goodall learned about chimp society by observing them for years
-e.g. you can learn about how children play by watching a kindergarten class
-Film of naturalistic observation:
-possible relationship between children who are leaders and good liars
-leaders determined by observing the number of times the children
engage in certain behaviors
-ability in lying determined by a committee which viewed a video tape of
the children lying about the taste of juice with the sound turned off
(nonverbal behaviors)
-Conclusion: the better a child was at lying, the more of a leader they were
-does the same relationship hold for adults?
-lying measured in the same way as the children
-leadership determined by observing them in a cooperation task
-For males, leaders are between liars, but there is no relationship
between the two traits for women.

2.) Correlation Method:


-A statistical measure that evaluates the degree of association/ relationship
between two variables
-allows the prediction of one variable based on the other
-e.g. in the video from last class, leadership skills increased as lying skill
increased for both children and men but not women
-e.g. possible correlation between aggressive behavior on the playground and
children’s exposure to violent TV.
Scatter plot - a graph of the relationship between plots two variables.

# of aggressive acts
| * *<-Chastity
| * *
| * *
| * *
| * *
|__*<-Butch__________________
exposure to violent TV

-You can draw a line through the points to show the trend
Positive correlation – two sets of scores increase or decrease together.

Correlation coefficients:
-can be computed from a set of scores
-precisely quantifies the degree of association between to variables
-2 parts:
-the number indicates the strength of the relationship
-a perfect relationship = 1.0
-no relationship is = 0
-the higher the number, the stronger the relationship.
-the plus or minus sign indicates the direction of relationship
+ (positive), the two scores increase or decrease together
- (negative, inverse relationship), as one score increases, the other
decreases

Important: correlation does not imply causation


-2 things may be correlated without one causing the other
-a third variable may be causing a correlation between two other variables.
-e.g. rate of skin cancer is positively correlated with the amount of time
spent wearing a hat because fair skinned people have high rates of
both of these variables
-e.g. cancer rates in adults are negatively correlated with liking for Coco
Puffs because older adults have higher cancer rates and less
preference for Coco Puffs than younger adults

Illusory correlation – the perception of relationship where none exists


-e.g. assuming that getting a call from an old friend just as you were thinking
about them was related to the fact that you were thinking about them; you probably forgot
about all the times you thought about them and they didn’t call

Important: neither correlation nor descriptive method can tell you whether one variable
causes another

3.) Experimental method:


-used to determine cause and effect relationship
-you manipulate a presumed cause and observe and document its effect on another
variable
e.g. Violent TV -expose one group of kids to violent TV and another to
non-violent TV, then compare the number of violent acts the kids
in the different groups perform

For an experiment you need to:


-take a random (representative ) sample from the a target population (e.g. first
graders)
-randomly assign kids in the sample to either the experimental (violent TV) or the
control (Nonviolent TV) group - The kids in the groups did not have a choice as to which
group they end up in.
-Type of TV - the independent variable (independent of the other variables - the
cause)
- the number of aggressive acts on the playground - the dependant variable (it
depends on the type of TV - the effect)

Summary of experimental design:


-Independent variable – the experimental factor you manipulate
-Dependent variable – the behavior measured; the factor that might be affected by
changes in the independent variable
-Experimental condition – the condition that exposes subjects to one version of
the independent variable
-Control condition – a condition identical to the experimental one; except the
independent variable has a different value, such as a zero
-Random assignment - assigning subjects to conditions by chance; thus
minimizing preexisting differences between those in the different conditions

Potrebbero piacerti anche