PSY6-L5

• CORRELATIONAL DESIGNS

  ● Determines the degree of relationship

  between traits, behaviors, or events.

  ● Show relationship between an antecedent

  and behavioral effects
• LINEAR REGRESSION ANALYSIS

  ● Predicting the score of a variable from the

  other variable in the study

  ● The stronger the correlation, the better the

  prediction
• CAUSAL MODELING

  ● Path Analysis

  • Uses beta weights
• CAUSAL MODELING

  Cross-Lagged Panel Design

  Measures the same pair of variables

  at two different points in time

  • Look at the pattern of correlations

  across time for possible direction of

  cause and effect
• QUASI-EXPERIMENTAL DESIGN

  ● Lacks one or more essential part/s of a true

  experiment

  ● Different treatment conditions on already

  pre-existing groups

  ● Antecedents are not directly manipulated,

  and can be naturally occurring event,

  characteristics, or behaviors

  ● Differences among groups

  ● Changes overtime in the same group of

  subjects

  ● Can never obtain a conclusive causal

  relationship
• QUASI-EXPERIMENTAL DESIGN

  ➢ Ex Post Facto Studies

  ➢ Non-Equivalent Groups Design

  ➢ Longitudinal Design

  ➢ Cross-Sectional Design

  ➢ Pre-test – Post-test Design
• EX POST FACTO STUDIES

  ● “After the fact”

  ● Low in the Degree of manipulation

  ● Natural occurrences

  ● Pre-existing differences on the respondents

  ● High in imposition of units

  ● No definite causal relationship can be made
• NON-EQUIVALENT GROUP DESIGNS

  ● Pre-existing differences on respondents

  ● Different group, different treatment

  ● Generally can be low in internal validity

  ● Should be planned very carefully
• LONGITUDINAL DESIGNS

  ● Focus on changes

  ● Influence of time

  ● Human and animal growth

  ● From months to even decades
• CROSS-SECTIONAL DESIGN

  ● Respondents are on the same situation but

  on a different period of it

  ● Requires more subjects

  ● Less powerful statistical treatments to data

  ● Pre-existing conditions in subjects can affect

  the results
• PRE-TEST & POST-TEST DESIGN

  ● Can be used to measure the effect to

  behavior of naturally occurring events

  ● Confounding variables

  ● Practice effects