1. Show that "delta" in Y did not occur until after "delta" in X
2. Show that X and Y are related. The cause and effect must be correlated with each other
3. Show that other explanations for the relationship between X and Y can be ruled out.
4. Plausible causal link between X and Y
Descriptive Research: Find out how things are
Example facial expressions- eyebrow flash
1. Important research method
2. Sometimes first step in research program
3. Often called case studies or 1-shot case studies
Relational Research: Find out how things are in RELATION to other things
Example Word usage and length of word; Social context and eyebrow flash
Can show that X and Y are related (covariation rule)
Contingency Table-
Tabular presentation of all combinations of categories of two variables, which allows the relationships between the two to be examined
Experimental Research-
Find out how things are in relation to other things and how they got that way
Experimental research:
Independent variable and dependent variable-
1. Show that delta in Y did not occur until after delta in X
2. Show that X and Y are related
3. Show that other explanations for the relationship between X and Y can be ruled out
Relational Research-
Find out how things are in relation to other things
Relational research: find out how things are in relation to other things:
Measure the degree of relationship between 2 or more naturally occuring characteristics, behaviors,or variables.
Researcher doesnt change the value of the variables-looks at natural occurances
Thus no I.V.-no manipulation
Predictor Variable(first possible cause) & outcome variable (second: possible effect)
Ex. Gender and helping behavior, Growth in hippocampus & depression, height and shoe size
Scatter Plots-
One way of doing relational research: look at linear correlation between predictor and outcome variable.
Typical statistic used in Pearson
Pearson Correlation Coefficient: Sometimes transform data by performing some standard mathematical operation on each score. Transform data for a number of reasons:
1. to obtain more convenient numbers
2. to meet the assumptions of certain stats
3. toobtain pretty graphs
4. to meet the assumptions of some theory
5. to minimize the influence of extreme scores
6. to allow for comparison across different data sets such as diff. psychological tests or college examinations
Two transformations:
1. Percent (per hundred)
2. Z scores
Z-scores
How far the score is from the mean in standard deviations units
1. measure of relative position
2. Can compare pple on different measures even if scales are different
3. Convert set of scores to z scores: mean = 0.0 and standard deviation = 1.0
The Standard Normal Curve has a mean of ____ and a standard Deviation of___
0.0
1.0
Strongest Correlation:
Both X and Y of each pair exactly same absolute distance from mean in Z score units
Strongest Positive Correlation:
Zx and Zy have same sign and same absolute distance from mean in Zs
Main idea-Pearson Product moment correlation:
Strong r if each member of pair-same relative position on its respective measure
Scaling-
assigning numbers to the magnitude of psychological events
Fechner's Law:
S = K log R
Weber's Law
Sensitivity of sensory system depends on absolute strength of a stimulus.
Difference threshold = constant x intensity of the stimulus
Steven's Power Law
S = k R^{N
}Equal physical ratios are psychologically equal
Scaling: Psychometrics-
Assign numbers to experiences caused by stimuli that vary along a single psychological dimension
Regression is
symbolized by R rather than r
Correlations have suspect internal validity because of:
The third variable problem
truncated range
Unknown direction of causation
The _______ ______ procedure enhances the internal validity
cross-lagged panel
Experiments are internally valid because:
the causal variables and the direction of causation are unknown