VQEG Research Tool

Number of Subjects Calculator

Estimate subjects needed for QoE experiments using power analysis with Bonferroni correction.

Parameters
Expected Standard Deviation 0.8
Expected spread of ratings across subjects. Typical values: 0.8–1.0 for ACR scales.
Desired MOS Difference 1.0
Minimum MOS difference you want to detect. E.g., 1.0 = one quality level on a 5-point ACR scale.
Number of Comparisons 100
Number of t-test comparisons planned. Used for Bonferroni correction to control Type I errors.
Desired Power 0.80
Probability of detecting a true effect (1 − Type II error). Typically 0.80.
Type of Test
Paired: same subjects rate both conditions. Independent: different subject groups.
Results
Minimum Subjects Required
--
subjects
Effect Size (d)
--
Adjusted Alpha
--
Base Alpha
0.05
Correction
Bonferroni

About

Written by: Werner Robitza

Based on the paper: Brunnström, K. and M. Barkowsky, Statistical quality of experience analysis: on planning the sample size and statistical significance testing. Journal of Electronic Imaging, 2018. 27(5): p. 11, DOI: 10.1117/1.JEI.27.5.053013

Copyright (c) Kjell Brunnström, Werner Robitza

License: MIT

More info about this app…