Lecture Notes: Introduction to Memory Version 3 by Amanda Montoya

About the Speaker

• Name: Amanda Montoya
• Position: Assistant Professor of Quantitative Psychology at UCLA

Purpose of the Video Series

• Created for beta testers of Memory Version 3.
• Aims to provide a fun, interesting way to learn about Memory through video format instead of just documentation.

Video Series Overview

1. General Introduction
   • Overview of Memory and changes in Version 3.
2. Installation Guides
   • How to install Memory in SPSS, SAS, or using the dialog box.
3. Model Templates
   • Explanation of model templates, expanded since Version 2.
4. Moderated Mediation Models
   • New feature in Version 3.
5. Conditional Effects Plots
   • Detailed explanation not commonly found in papers.
6. Wrap-up Video
   • Additional details and resources if certain aspects are not of interest.

Introduction to Memory

• Definition: Macro for SPSS and SAS for mediation and moderation in repeated measures designs.
• Name Origin: Mediation and Moderation for Repeated Measures designs.
• Functionality:
  • Allows mediation and moderation analyses in repeated measures.
  • Works with continuous mediators and continuous outcome variables.
  • Not suitable for dichotomous outcomes (a student is working on this).
  • Utilizes different scores to model the effect of repeated measures factors on outcomes.
• Appropriateness:
  • Memory is used when the causal variable is a repeated measures factor.
  • Not suitable for cases with measured X variables.

What's New in Version 3

• Development History:
  • Version 1: Mediation.
  • Version 2: Added Moderation.
  • Version 3: Adds Moderated Mediation.
• New Additions:
  • Models 4-18: Incorporate one moderator into simple or parallel mediation models.
  • Smarter mean centering:
    • Center = 0: No centering.
    • Center = 1: Centers all moderators, including dichotomous ones.
    • Center = 2: Centers only non-dichotomous moderators.

Changes in Version 3

• Bias Corrected Bootstrap Confidence Interval:
  • Removed due to inflated Type I error rates.
  • Default is now Percentile Bootstrap Confidence Interval.
  • Monte Carlo Confidence Interval remains available.
• Standard Error Calculation Error:
  • Discovered error in normal standard error estimates for serial mediation models.
  • Affected users of the normal = 1 option.
  • Corrections made; parallel mediation models unaffected.

This concludes the introduction video. Further details will be available in the subsequent videos.