Reaction Time Analysis Using Factorial Experimental Design and ANOVA

Overview

This project investigates how different behaviors and environmental stimuli affect human reaction time using a controlled experimental design. The analysis evaluates the effects of white noise, multitasking, and dominant hand usage on reaction time performance measured in milliseconds.

A $2^3$ factorial randomized block design with replication was used to account for between-participant variation while allowing for the estimation of main effects and interaction effects between factors. Statistical modeling and ANOVA techniques were applied to determine which factors were significantly related with average reaction times and reaction time variability.

Dataset

The dataset was generated experimentally through repeated reaction time tests performed using a mobile reaction time application.

The Experiment:

• Setup:
  • 3 two-level experimental factors
  • 2 participants used as blocking factors
  • 4 replications for each treatment combination
  • Fully randomized run order
• Experimental Factors:
  • Noise: Presence or absence of white noise during testing
  • Multitask: Presence or absence of a simultaneous counting task
  • Dominant_Hand: Whether the participant used their dominant hand
• Blocking Factor:
  • Person: Participant identifier used to account for between-subject variation
• Response:
  • React_time: Reaction time measured in milliseconds

Methods

• Experimental design construction
• Randomized block factorial design
• Replication and randomized run ordering
• Linear regression modeling
• Interaction effect analysis
• Analysis of variance (ANOVA)
• Smaller-the-better optimization analysis

Models

• Linear Regression Models: Used to estimate the relationships between experimental factors and average reaction time/variance of reaction times
• ANOVA Framework: Used to evaluate statistical significance of main effects and interaction effects between experimental factors and the average reaction time/variance of reaction times while accounting for the blocking variable

Results

The analysis identified that cognitive multitasking is statistically associated with slower average reaction times, and none of the experimental factors were meaningfully related to the variability of reaction times.

The findings of the smaller-the-better analysis show that reaction times are fastest under the absence of cognitive multitasking, which makes sense intuitively. Though this was the only experimental relationship found between the experimental factors and the average reaction times, this finding has practical significance. For example, those working in potentially dangerous circumstances should refrain from cognitive multitasking to best prepare themselves for any quick decision-making that may occur.

Technologies

• R
• tidyverse
• MASS
• ANOVA
• Linear Regression
• R Markdown

Structure

• Reaction_Time_Report.Rmd - Full analysis and modeling report/code
• Reaction_Time_Report.pdf - Report-ready document with only essential code
• Requirements.txt - R dependencies
• test1.jpg - Image of experimental conditions included in the report

Notes

This project was completed as part of a graduate course project and adapted for portfolio presentation.