A/B & Multivariate Testing

Designing Factorial Experiments

Factorial designs reveal not just what works but why it works. Sharpen your skills in structuring powerful experiments.

A 2^3 factorial experiment tests how many treatment combinations?

8

6

16

4

Two levels for each of three factors results in 2×2×2 = 8 unique combinations.

In factorial terminology, an effect where the impact of one factor depends on another is called an ______ effect.

interaction

random

additive

blocking

Interactions show that factors do not operate independently.

Adding centre points to a factorial design primarily helps detect ______.

heteroscedasticity

multicollinearity

curvature in the response

serial correlation

Centre runs reveal whether the true relationship is non‑linear.

A fractional factorial 2^(k‑p) design is most useful when you want to ______.

model every high‑order interaction

reduce the number of required runs while screening factors

increase aliasing

avoid randomisation

It sacrifices some detail to make large studies feasible.

Design resolution describes the level of ______ between effects in a fractional factorial.

variance

stratification

confounding (aliasing)

power

Higher resolution means fewer important effects are aliased together.

Blocking in factorial experiments is a strategy to handle ______ variables.

independent

nuisance

latent

response

Blocks group runs to filter out variation you don’t care to measure.

Which statistical technique is typically applied to analyse factorial experiment results?

K‑nearest neighbours

Latent Dirichlet allocation

ANOVA

Survival analysis

ANOVA partitions variation attributable to each factor and their interactions.

In marketing email tests, factors might include subject line, send time, and ______.

preview text snippet

email size in KB

subscriber age

SMTP server

Each element that could influence opens or clicks can be a factor.

Aliasing in a fractional factorial design means that two or more effects are ______.

over‑sampled

auto‑correlated

statistically indistinguishable in the analysis

mutually exclusive

You cannot tell which aliased factor caused any observed change.

To visualise interaction effects in a 2‑factor experiment, analysts often use ______ plots.

funnel

scatter matrix

interaction (profile)

Pareto

Parallel lines indicate no interaction; diverging lines suggest one.

Starter

Starter—brush up on factorial notation.

Solid

Solid—you understand factors and interactions.

Expert!

Expert—your factorial designs balance rigour and efficiency.

What's your reaction?

Related Quizzes

1 of 10

Leave A Reply

Your email address will not be published. Required fields are marked *