Psychology Research Methods Lecture

Longitudinal Designs

• The residual longitudinal regression design allows testing for both stability and change, but has limitations, including only testing for uni-directional effects.
• The residual longitudinal regression design only measures the dependent variable at two time points, and the independent variable is only measured at one time point.
• The dependent variable is tested as the criterion, and the independent variable is not tested as a potential outcome.

Cross-Lagged Design

• The cross-lagged design is a more advanced approach that allows testing for bi-directional effects.
• This design combines two residual longitudinal regressions into one analysis.
• The stability of the constructors is tested against each other, and each variable is used to predict the change in the other variable.
• This approach has six parameters: correlations between the variables at time one, correlations between the variables at time two, and the tests of stability for both variables.
• The cross-lagged design is a powerful approach that can investigate bi-directional relationships.

Example Study

• A study of 1000 adolescents followed over two years examined the relationship between binge drinking and sharing alcohol content on social networking sites.
• The study used a cross-lagged design to investigate the direction of the relationship between sharing alcohol content and binge drinking.
• The results showed significant bidirectional effects, with binge drinking predicting changes in sharing alcohol content and vice versa.

Assumptions of Longitudinal Analysis

• The first assumption is individual stability, which assumes that there is no systematic difference in the stability and change between groups.
• The second assumption is consistent measurement, which assumes that the measurement is the same across time periods.
• The third assumption is synchronicity, which assumes that the administration of the questionnaires is at the same interval between time periods for all participants.
• The fourth assumption is timeframe, which assumes that the time between measurements is sufficient to capture the relationship between variables.
• Failure to consider these assumptions can lead to incorrect conclusions.

Limitations of Longitudinal Analysis

• Longitudinal analysis can be expensive and time-consuming.
• The design may not be able to capture causality, especially if there are third variables that are not accounted for.
• The analysis may be affected by omitted variables, which can have a substantial impact on the interpretation of the results.

Cross-Lagged Models

• In cross-lagged models, both measures serve as both independent and dependent variables.
• There are two or more dependent variables (DVs) in this analysis.
• The technique combines two residualized longitudinal regressions into the same analysis.
• The stability of the construct is statistically removed, allowing each variable to predict the change in the other variable.
• This technique allows for an investigation of bi-directional effects.

Parameters of a Two-Wave Cross-Lagged Model

• 6 parameters:
  • Correlation between variables at time 1
  • Correlation between variables at time 2
  • Stability of Variable 1
  • Stability of Variable 2
  • Cross-lag path 1: Variable 2 at T1 predicting change in Variable 1
  • Cross-lag path 2: Variable 1 at T1 predicting change in Variable 2

Example of a Cross-Lagged Model

• Studied the longitudinal relationship between binge drinking and sharing alcohol content on social networking sites (SNS) among 1,006 adolescents over two years.
• Found a bi-directional relationship between sharing alcohol references on SNS and binge drinking.

Overview of Longitudinal Regression

• Longitudinal regression models:
  • Accounts for stability and change
  • Explores uni- and bi-directional relationships
  • Predicts change

Assumptions of Longitudinal Analysis

• Inter-individual stability:
  • Assumes no systematic differences in stability and change between participants
  • Violated when one group of participants changes faster or slower relative to the other participants
• Consistent measurement:
  • Assumes the same measurement when using repeated measures
  • Items administered to participants need to be exactly the same across time periods
• Synchronicity:
  • Assumes the administration of the questionnaire occurs with the same interval between time periods for all participants

Considerations of Longitudinal Analysis

• Timeframe:
  • Needs to be considered to find a longitudinal effect
  • Length of time between measurements may be too short or too long for a variable to impact another
• Other variables (third variable effect):
  • Important to ensure that important variables are not omitted
  • Related to a causal process or pathway