Experimental Design

PSYC 641 - Spring 2025

Welcome Back!

Updates/Check-in

Computational setup (don’t think we’ll need)

# load packages
library(tidyverse)   # for data wrangling and visualization
library(scales)      # for pretty axis labels
library(knitr)       # for pretty tables
library(patchwork)   # for laying out plots

# set default theme and larger font size for ggplot2
ggplot2::theme_set(ggplot2::theme_minimal(base_size = 20))

Experimental Designs

  • Identified as the “Gold Standard” within research. Why is it considered the “gold standard”??

  • The ability to draw causal connections between variables

    • e.g., Diffusion of Responsibility (Darley & Latane, 1968)

Stress Experiment

On a scale of 1 (low levels) - 10 (high levels), take note of your current level of stress

Stress Experiment Follow-up

On a scale of 1 (low levels) - 10 (high levels), take note of your current level of stress

  • Did these cute animal pictures cause you to be more/less stressed?

Causation

How can we determine causation? i.e., one variable causes another to change

Shadish, Cook & Campbell (2002)

Drawing from John Stuart Mill’s philosophy

Three Criteria for Causal Relationships:

  1. Temporal precedence (cause precedes effect)

  2. Covariation (cause and effect are correlated)

  3. Elimination of plausible alternative explanations (ruling out confounding variables)

Counterfactual Reasoning

  • Hypothetical situation where manipulated outcomes are compared to the outcomes without a manipulation (i.e., time travel)

  • How does this work?

    • Group of individuals get a treatment

    • We measure their outcomes

    • We rewind time and do not give the group the treatment

    • We measure these outcomes

    • Compare!

Design

Random Assignment –> IV –> DV

Simple!

  • What if we can’t randomly assign? What are things that we can’t randomly assign?

Quasi-experimental Design

  • But that is next week. Stay on track and don’t get distracted

Random Assignment

Every participant has an equal likelihood of being in each of the groups

  • Coin Flip

  • Computer Algorithm

Will need to check for group differences on variables of interest and other meaningful characteristics.

  • What statistical test would we use?

Randomized Control Trial (RCT)

  • We have a random sample ✅

  • We have random assignment ✅

  • But what are the groups that we are assigning?

    • What is the research question, and what do we want to compare it against?

RCT

The comparison/control group means a lot when it comes to the conclusions that we can establish

  • Treatment vs. …

    • Waitlist Control?

    • Treatment as usual?

    • Other treatment?

    • Placebo?

    • Nothing?

Placebo Effect

This has been an emerging area of research that is utterly fascinating

On occasion, the placebo shows just as much of an effect as the actual treatment

Example: Treating Depression (link)

Example

Investigate the bidirectional relationships between sleep patterns, social media use, and depressive symptoms in adolescents over a two-year period.

  • Measures:
-   Sleep: Actigraphy (1 week per assessment) + Pittsburgh Sleep Quality Index

-   Social Media: Screen time logs + Social Media Addiction Scale

-   Depression: Beck Depression Inventory

-   Covariates: Pubertal status, SES, parental depression

Draw it on the board

  • What stats can we do?

Next Class…

We will have a class activity where you will evaluate studies and provide additional input

Please review the following studies:

  • Bobo Doll Experiment

  • Zimbardo Prison Study

  • Milgram Obedience Study

  • “Monster” Study

    Mundane Realism vs. Psychological Realism