PSY 2105 Chapter Notes - Chapter 2.2: Random Assignment, Quasi, Repeated Measures Design
27 Apr 2018
School
Department
Course
Professor
CORRELATIONAL RESEARCH
•Variable:
Any factor that can take on different values
E.g.: running speed, intelligence quotients, gender
• Correlation:
A statistical statement as to the degree and direction of relationship between two variables
Positive: Values of one variable change in the same direction (increase or decrease) as the other
variable
Negative: High values of one variable are associated with low values of the other variable
•The correlational coefficient “r” reflects the direction and strength of a relation between two variables
Ranges from -1 through 0 through +1
Negative values reflect a negative relation
Positive values reflect a positive relation
The strength of the relation is indicated by the size of the number: 0.5 is less strong than is 0.99
Correlations do not prove causality
•Example:
A high positive r (0.78) does not imply that watching Sesame Street causes
improved reading – only that the two variables are related
Suggests the need for an experiment
EXPERIMENTAL RESEARCH
•Experiments offer the opportunity to prove causality, i.e., manipulation of one
variable induces change in another variable
• Variables in an experiment:
Independent variable (IV): Is manipulated by the experimenter and is assumed to be a causal factor
Dependent variable (DV): Is measured by the experimenter and is assumed to be controlled by the IV
•Other experimental designs
Sometimes variables can not be experimentally manipulated
Ethical concerns
The nature of some independent variables (such as cultural background)
Quasi-experimental studies:
- Allow researchers to compare groups differing on some important characteristics
-Lack random assignment
-IV – is naturally occurring and you’re putting them in those groups,
Research Type
Method
Description
Usefulness
Drawbacks
Descriptive
Naturalistic
observation
children’s behaviour is
observed in real-life
setting
direct source of info for how
children behave in the
natural setting
presence of observers may
alter the setting and thus the
behaviour
some behaviours may be
difficult to observe in the
natural setting
setting may not be the same
for all children, making it
difficult to compare between
participants
Structural
observation
children’s behaviour is
observed in a structured
lab experiment
controlled lab environment
ensures that the behaviours
of interest will occur
allows for comparisons
across participants
setting is not real-life, which
can reduce the
generalizability of the finding
Interview
children or other
knowledge informants
are asked to provide
verbal reports via
interview or
questionnaire
can provide valuable info
about what informants think
or feel
not all behaviours are
accessible to verbal report
concerns about what the
interviewer wants to hear or
about portraying oneself
positively may influence the
accuracy of reports
Case study
detailed descriptive
study of a single
individual
allows the study of specific
or unusual situations, thus
aiding clinical intervention
can raise questions for
further study using other
research methods
because only one child is
studied, findings may not be
generalizable to other
children
Correlational
Correlational
study
examines how 2+
variables are related
allows researchers to
quantify relationships
between variables and make
predictions about one
variable based on the other
can’t be used to show
casuality
Experimental
Experiment
researcher manipulates
independent variable
and looks for
corresponding changes
in dependent variable
allows researchers to
assess the effects of one
variable on another
it’s not always possible or
ethical to manipulate certain
types of variables
results may not be
generalizable to real-life
settings
Quasi-
Experimental
groups that differ on
some important
characteristic are
compared
allows researchers to
examine variables that can’t
be experimentally
manipulated
can’t be used to show
cause-and-effect
relationships
Document Summary
Any factor that can take on different values. : running speed, intelligence quotients, gender: correlation: A statistical statement as to the degree and direction of relationship between two variables. Positive: values of one variable change in the same direction (increase or decrease) as the other variable. Negative: high values of one variable are associated with low values of the other variable: the correlational coef cient r re ects the direction and strength of a relation between two variables. The strength of the relation is indicated by the size of the number: 0. 5 is less strong than is 0. 99. A high positive r (0. 78) does not imply that watching sesame street causes improved reading only that the two variables are related. Experimental research: experiments offer the opportunity to prove causality, i. e. , manipulation of one variable induces change in another variable, variables in an experiment: Independent variable (iv): is manipulated by the experimenter and is assumed to be a causal factor.
