Factorial Design Explained: Full vs Fractional Factorial (With Free Tool)
Factorial designs are the workhorses of experimental research. They let you study multiple factors simultaneously and discover interactions that single-factor experiments miss entirely.
This tutorial walks through full and fractional factorials with a live interaction plot you can modify and run.
What You Will Learn
0.Live Code: Interaction Plot
1.What is a Factorial Design?
2.Full Factorial (2^k)
3.Fractional Factorial
4.Interaction Effects
5.Visualizing Results
0. Live Code: Interaction Plot
A 2x2 factorial interaction plot - the key visualization for detecting factor interactions. Non-parallel lines mean the effect of one factor depends on the other.
Preparing preview
Running once automatically on first load
Learn by Experimenting
This is a safe playground for learning! Try changing:
- • Colors: Modify color values to see different palettes
- • Numbers: Adjust sizes, positions, or data ranges
- • Labels: Update titles, axis names, or legends
Edit the code, run it, then open the full data visualization tool to continue with your own dataset.
1. What is a Factorial Design?
A factorial design investigates the effect of two or more factors by testing all possible combinations of their levels. Unlike one-factor-at-a-time (OFAT) experiments, factorial designs reveal interactions - when the effect of one factor depends on the level of another.
Why Not OFAT?
One-factor-at-a-time experiments require more runs to achieve the same precision and completely miss interactions. Factorial designs are statistically more efficient in every way.
2. Full Factorial Design (2^k)
In a 2^k design, you have k factors, each at 2 levels (typically "low" and "high"). The total number of experimental runs is 2^k.
| Run | Temp | Pressure | Catalyst |
|---|---|---|---|
| 1 | Low | Low | Low |
| 2 | High | Low | Low |
| 3 | Low | High | Low |
| 4 | High | High | Low |
| 5 | Low | Low | High |
| 6 | High | Low | High |
| 7 | Low | High | High |
| 8 | High | High | High |
3 factors x 2 levels = 2^3 = 8 runs. Each row is one experimental condition.
3. Fractional Factorial Design
When you have 5+ factors, full factorial becomes expensive (2^5 = 32 runs, 2^6 = 64 runs). Fractional factorial designs use a strategically selected subset of runs to estimate main effects and some interactions.
2^(5-1)
16 instead of 32
Loses some 3-way interactions
2^(6-2)
16 instead of 64
Loses 2-way confounding
2^(7-3)
16 instead of 128
Resolution III - screening only
Create Your Design Free
Use our Experiment Planner to generate factorial designs automatically. Get randomization, blocking, and export your design matrix.
4. Understanding Interaction Effects
The key advantage of factorial designs is detecting interactions. An interaction occurs when the effect of factor A depends on the level of factor B.
For example: Temperature might have a small effect at low pressure, but a large effect at high pressure. This is a Temperature x Pressure interaction. In the interaction plot above, you can see this as non-parallel lines.
Reading Interaction Plots
Parallel lines = no interaction (factors are independent). Non-parallel lines = interaction present. Crossing lines = strong interaction (factor effects reverse depending on other factor levels).
5. Visualizing Factorial Results
After running your experiment, upload results to Plotivy and ask for visualizations. Example prompt: "Create an interaction plot showing Temperature x Catalyst effect on Yield".
Bar Charts
Main effects comparison
Line Graphs
Interaction plots
Heatmaps
Response surfaces
Box Plots
Replicate variability
Chart gallery
Chart types for factorial analysis
Start from these chart types commonly used in DOE analysis.
Bar Chart
Compares categorical data using rectangular bars with heights proportional to values.
Sample code / prompt
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from scipy import stats
# Generate performance scores for 5 treatment groups
np.random.seed(42)
groups = ['Control', 'Treatment A', 'Treatment B', 'Treatment C', 'Treatment D']
n_samples = 30
Line Graph
Displays data points connected by straight line segments to show trends over time.
Sample code / prompt
import matplotlib.pyplot as plt
import numpy as np
# Generate temperature data for 3 major US cities over 12 months
months = ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
nyc = [30, 32, 40, 52, 65, 75, 82, 81, 74, 63, 50, 38]
miami = [65, 66, 70, 76, 82, 87, 90, 90, 87, 80, 72, 66]
chicago = [25, 27, 35, 48, 62, 72, 80, 79, 71, 60, 45, 32]
# Create figure with enhanced styling
Heatmap
Represents data values as colors in a two-dimensional matrix format.
Sample code / prompt
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
import numpy as np
# Create correlation matrix for financial metrics
metrics = ['Revenue', 'Profit', 'Expenses', 'ROI', 'Customers', 'AOV', 'Marketing', 'Employees']
correlation_data = np.array([
[1.00, 0.85, -0.45, 0.72, 0.88, 0.65, 0.72, 0.55],
[0.85, 1.00, -0.78, 0.92, 0.75, 0.58, 0.63, 0.48],.png&w=1920&q=75)
Box and Whisker Plot
Displays data distribution using quartiles, median, and outliers in a standardized format.
Sample code / prompt
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from scipy import stats
# Generate gene expression data for 4 genotypes
np.random.seed(42)
genotypes = ['WT', 'KO1', 'KO2', 'Mutant']
n_per_group = 20
Error Bars
Graphical representations of the variability of data indicating error or uncertainty in measurements.
Sample code / prompt
import numpy as np
import matplotlib.pyplot as plt
from scipy import stats
# Generate bacterial growth data with replicates
np.random.seed(42)
time_points = np.array([0, 4, 8, 12, 18, 24])
mean_values = np.array([10, 25, 80, 250, 600, 800])
# Generate 5 replicates per time point with noiseFrom Design to Visualization
Design your experiment with the Planner, then upload results to create publication-ready interaction plots automatically.
Found this helpful? Share it with your network.
Experimental Physicist & Photonics Researcher
Hands-on experience in silicon photonics, semiconductor fabrication (DRIE/ICP-RIE), optical simulation, and data-driven analysis. Built Plotivy to help researchers focus on discoveries instead of data struggles.
More about the authorVisualize your own data
Apply the techniques from this article to your own datasets. Upload CSV, Excel, or paste data directly.