-
Other names for "tests"
- assessment
- test
- measure
- instrument
- scale
-
reliability
- a test is reliable if the data is consistent over time when the test is
- repeated
-
-
Test-retest
when someone is administered the same test over time
-
Split-test
Taking parts of a test and comparing those parts to one another
-
Validity
Is the test supposed to test what it is supposed to test?
-
Tests of validity
- content validity
- construct validity
- predictive validity
-
content validity
does the content make sense according to the topic?
-
predictive validity
does it predict a difference in behavior?
-
Theory
A general statement about the relationship between constructs or events.
-
Hypothesis
- Formal prediction about the relationship between two or more variables
- that is logically derived from the theory
- derived from a theory
-
correlational research
- When: used to investigate relationships between variables
- How:
- measure both or all groups involved in study
- limit: does not give a
- cause
- keywords: relationship, related, linked to, associated
-
Correlation coefficient
- between -1 to 1
- the closer to one the correlation is, the stronger
- the relationship is
-
positive correlation
- Both variables travel in the same direction
- 0-1 (0 to .2= no
- correlation, .2 to .4= weak correlation, .4 to .6 =moderate correlation,
- .6 to .8 =strong correlation
- Example: The more time one spends
- studying, the better grades that person will recieve on an exam
-
Negative Correlation
- Both variables that travel in opposite directions
- -0 to -1 (0 to
- -.2= no correlation,
- -.2 to -.4= weak correlation, .-4 to .-6
- =moderate correlation, -.6 to -.8
- =strong correlation
- Example:
- The more time someone spends partying the person is most likely to
- recive a lower score on an exam
-
Experimental Research
- When: when you want to draw cause and effect
- How: manipulate one of
- the variables under controlled conditions
- limit: hard to make
- entirely random and hard to control variables
-
Independent variable
- The variable in a experiment that is being manipulated
- causes direct
- variable
- Example: Different temperatures may cause subjects to
- score higher on test ( temperature is independant variable)
-
Dependent variable
The outcome that is affected by the independent variable
-
Why you may want more than 1 Independent variable
There may be interactions between variables
-
Why you may want more than 1 Dependent variable
The Independant variable may have influenced another dependent variable
-
A true experiment
- A manipulated independent variable while everything else is controlled
- randomly
- assigned groups
- There will be an infinite amount of independent
- variables if groups aren't randomly assigned
-
Experimental Groups
The group that recieves treatment(independent variable)
-
Control Group
The group that doesn't get treatment
-
Sampling Bias
- subjects tested are not chosen by random
- corrected by random
- assignment
-
Confounding bias
- other variables may have affected dependent variable
- corrected by
- randomly assigning groups
-
Placebo effects
- People belive that treatment actually works
- corrected by blind study
-
Experimentor effect
- The experimenter's expectations influence results
- corrected by a
- double blind study
-
Social desirability effect
- subject is not comfortable with test or answers on test
- anonymity or
- non-self report
-
Significance Test
- Tests that determine if group averages were found from real effects or
- just by chance
- Example: ANOVA, Chi-square, or t-test
-
P-value
The number that determines the significance between dependent variables
-
Interpreting P-value
- n = number of subjects in group
- N = total number of subjects in group
- P
- < .05 means that there is a significant difference
- P > .10
- means that there is not a significant difference
- P between .05 and
- .10 means that there is a marginally significant. This usually means
- that you need to get more data and withold judgement
-
Replication
- When an experiment is done by another researcher. If similar results are
- found in future experiments then the experiment is seen as significant
- in terms of a genuine relationship
|
|