Design resolution classifies how severely effects are confounded (aliased) in a fractional factorial DOE, using Roman numerals (III, IV, V). It matters because a fractional factorial design deliberately runs fewer combinations than a full factorial, which means some effects become mathematically indistinguishable from others, and the resolution tells you exactly which ones. In a Resolution III design, main effects are aliased with two-factor interactions.
However, In a Resolution IV design, main effects are clean, but some two-factor interactions are aliased with each other. In a Resolution V design, both main effects and two-factor interactions are clean, aliased only with higher-order interactions that are rarely significant in practice. The practical rule: use the highest resolution your run budget allows, since higher resolution means less ambiguity in the results.
Quick Reference Table
| Resolution | What’s Aliased | What’s Clean | Typical Use |
| Resolution III | Main effects aliased with 2-factor interactions | Main effects not aliased with each other | Aggressive screening with many factors and very few runs |
| Resolution IV | Some 2-factor interactions aliased with each other; main effects aliased with 3-factor interactions | Main effects clean from all 2-factor interactions | The most common practical choice for screening |
| Resolution V | 2-factor interactions aliased with 3-factor interactions; main effects aliased with 4-factor interactions | Main effects and all 2-factor interactions clean | Characterization/optimization when 2-factor interactions matter |
| Foldover Design | N/A (a follow-up design, not a resolution level) | Removes aliasing between main effects and 2-factor interactions | Run after a Resolution III design when results are ambiguous |
| Full Factorial | Nothing aliased | Everything estimable | When run budget allows testing every combination |
Table of contents
Key Takeaways
- Design resolution tells you which effects you can’t tell apart, not whether your experiment is valid. A Resolution III design is still a legitimate design; it just requires knowing what its confounding limits are before interpreting results.
- Higher resolution generally requires more runs. Resolution III designs use the fewest runs; Resolution V designs require substantially more, which is the core trade-off practitioners are managing.
- Resolution IV is the most common practical compromise. It keeps main effects clean, the results most practitioners care about most, while still using significantly fewer runs than a full factorial or Resolution V design.
- A foldover design can upgrade a Resolution III design to Resolution IV after the fact, without having to discard the original data or start over.
- The resolution is determined by the design’s “defining relation,” specifically the length of its shortest term, which is a detail practitioners rarely need to calculate by hand since DOE software determines it automatically.
- Higher-order interactions (three-factor and above) are rarely practically significant, which is the underlying assumption that makes Resolution III and IV designs useful despite their confounding.
- Choosing a resolution is a trade-off decision, not a fixed rule. The right choice depends on how many factors you’re screening, how many runs you can afford, and how much you already suspect about interaction effects.
What Is Design Resolution?
Design resolution describes how much the effects in a fractional factorial design are aliased (also called confounded) with other effects. When you run a fractional factorial design instead of a full factorial, you deliberately test fewer combinations, which means the design can no longer estimate every effect independently. Some effects get mathematically combined, and the resolution tells you exactly which ones.
Design resolution is written using Roman numerals, typically as a subscript on the design notation (for example, a 2^(7-4) design with resolution III is written 2_III^(7-4)). The resolution of a design is given by the length of the shortest word in the defining relation, the mathematical rule that determines which combinations of factor settings are included in the reduced design.
Public, Onsite, Virtual, and Online Six Sigma Certification Training!
- We are accredited by the IASSC.
- Live Public Training at 52 Sites.
- Live Virtual Training.
- Onsite Training (at your organization).
- Interactive Online (self-paced) training,
Why Does Design Resolution Matter?
Fractional factorial designs exist because testing every possible combination of factors (a full factorial) becomes impractical fast. At times, the number of factors to be investigated in a screening experiment is so large that even running a fractional factorial design is impractical, which is exactly the situation resolution helps manage: it quantifies what you’re giving up in exchange for the runs you’re saving.
The higher the resolution of a design, the less confounding there is. This matters practically because confounded effects cannot be separated statistically.
If factor A is confounded with the three-way interaction BCD, the estimated effect for A is the sum of the effect of A and the effect of BCD, and you cannot determine whether a significant result is because of A, because of BCD, or because of both. Knowing your design’s resolution tells you, before you even run the experiment, exactly which conclusions you will and won’t be able to draw cleanly from the results.
The Three Common Resolutions Explained
Resolution III
In a Resolution III design, main effects are not confounded with other main effects, but they are confounded with at least some two-factor interactions. This is the most aggressive, run-efficient option: a 2^(3-1) design can screen three factors in four runs, a 2^(7-4) design can screen seven factors in eight runs, and a 2^(15-11) design can screen fifteen factors in sixteen runs.
The trade-off is real: if a significant two-factor interaction exists, it will show up entangled with a main effect, and the experiment alone cannot tell you which one is actually driving the result.
Resolution IV
In a Resolution IV design, main effects are not confounded with other main effects or with any two-factor interactions, but two-factor interactions might be confounded with one another. This is the most commonly used practical resolution for screening, since main effects (which are usually the primary interest in an early screening study) come out clean, while the design still uses meaningfully fewer runs than a full factorial.
Resolution V
In Resolution V designs, main effects and all two-factor interactions can be estimated cleanly; two-factor interactions are only confounded with three-factor interactions. This resolution is used when two-factor interactions themselves matter to the analysis, not just the main effects, typically at the characterization or optimization stage of a DOE project rather than the initial screening stage.
How Do You Choose a Design Resolution?
1. Determine How Many Factors You Need to Screen
The number of factors under investigation, combined with your available run budget, largely dictates which resolutions are even feasible. Screening a large number of factors in a small number of runs generally forces a lower resolution.
2. Decide Whether Interactions Are Likely to Matter
If prior knowledge or subject-matter expertise suggests meaningful two-factor interactions are unlikely, a Resolution III or IV design is often a reasonable, efficient choice. If interactions are suspected to matter, Resolution V is worth the additional runs.
3. Choose the Highest Resolution Your Run Budget Allows
You generally want to use a fractional factorial design with the highest possible resolution for the amount of fractionation required. Given a choice, it is usually better to select a design where main effects are confounded with three-way interactions (Resolution IV) rather than one where main effects are confounded with two-way interactions (Resolution III), since three-way interactions are far less likely to be practically significant.
4. Plan for a Follow-Up Foldover If You Start With Resolution III
If run constraints force a Resolution III design, decide in advance whether a foldover run is realistic if the results come back ambiguous. Planning for this possibility avoids a scramble later if a significant but unclear effect turns up.
What Is a Foldover Design?
Folding is the standard technique for reducing aliasing after running a lower-resolution design. Resolution IV designs may be obtained from Resolution III designs by folding on all factors. If you fold on one factor, then all terms involving that factor become free from aliasing with terms that do not involve that factor.
In practice, this means: once results from a Resolution III design are obtained, and if three-factor and higher-order interactions can reasonably be assumed unimportant, the experimenter can decide whether a fold-over design is needed to de-alias the main effects from the two-factor interactions.
This is a genuinely useful piece of practical guidance that most competing glossary pages on this term omit entirely: you are not stuck with the ambiguity of a Resolution III result. A follow-up foldover run can resolve it without discarding the original data.
Real-World Example (Hypothetical)
Problem: A chemical manufacturer wants to screen seven process factors (temperature, pressure, catalyst concentration, mixing speed, and three others) suspected of affecting yield, but a full factorial would require 128 runs, far more than the pilot plant schedule allows.
Analysis: The team selects a 2^(7-4) Resolution III design, requiring only 8 runs, based on the working assumption that interactions between these particular factors are unlikely to be practically significant given prior process knowledge.
Six Sigma approach: The initial 8-run screening identifies three factors with apparently significant main effects, but one result is ambiguous, aliased with a two-factor interaction the team cannot rule out based on the initial 8 runs alone.
Action: Rather than treating the ambiguous result as inconclusive, the team runs a foldover design on the suspect factor, adding a modest number of additional runs to de-alias the main effect from the interaction it was confounded with.
Result (hypothetical): The foldover confirms the effect was genuinely driven by the main factor, not the interaction, giving the team a defensible conclusion without needing to run the full 128-run factorial. This is a hypothetical illustration of how resolution and foldover work together in practice, not a documented case study.
Common Mistakes With Design Resolution
- Choosing a resolution without checking the runs-vs-factors trade-off first. Selecting a target resolution before confirming the run budget can support it leads to redesigning the experiment later.
- Treating a Resolution III result as fully conclusive. Any significant effect in a Resolution III design should be checked against its known aliasing structure before being treated as confirmed.
- Ignoring the aliasing structure entirely. DOE software reports exactly which effects are aliased with which; skipping this output before interpreting results risks misattributing a significant effect.
- Defaulting to Resolution V “to be safe” without checking the run cost. Resolution V designs can require substantially more runs than the project’s screening budget justifies, especially early on when the goal is just identifying important factors.
- Not planning for a foldover option when starting with Resolution III. Teams that don’t budget for a possible follow-up run get stuck when an ambiguous result appears late in the project timeline.
When Should You Use Each Resolution?
Use Resolution III when:
- You are screening a large number of factors with a very limited run budget.
- Prior knowledge strongly suggests interactions are unlikely to be significant.
- You are prepared to run a foldover design if results come back ambiguous.
Use Resolution IV when:
- Main effects are your primary interest and need to be estimated cleanly.
- You want a meaningful reduction in runs compared to a full factorial without accepting Resolution III’s main-effect aliasing risk.
Use Resolution V when:
- Two-factor interactions are suspected to matter and need to be estimated, not just main effects.
- You are past initial screening and into characterization or optimization, where run budget is less constrained.
Frequently Asked Questions (FAQs) on DOE
Q: What is design resolution in DOE?
A: Design resolution is a classification (Resolution III, IV, V) that describes how severely effects are confounded, or aliased, with each other in a fractional factorial design. Higher resolution means less confounding and clearer results.
Q: What is the difference between Resolution III, IV, and V designs?
A: In Resolution III, main effects are aliased with two-factor interactions. In Resolution IV, main effects are clean, but some two-factor interactions are aliased with each other. In Resolution V, both main effects and two-factor interactions are clean, aliased only with higher-order interactions.
Q: How do you choose a design resolution?
A: Choose the highest resolution your run budget can support, based on how many factors you’re screening and whether prior knowledge suggests interactions are likely to matter. Resolution IV is the most common practical compromise for initial screening.
Q: What is a foldover design?
A: A foldover is a follow-up design run after an initial fractional factorial experiment, typically used to upgrade a Resolution III design’s effective resolution by de-aliasing main effects from the two-factor interactions they were confounded with.
Q: Why is Resolution IV often preferred over Resolution III?
A: Resolution IV keeps main effects, usually the primary interest in a screening study, completely free of aliasing with two-factor interactions, while still requiring meaningfully fewer runs than a full factorial or a Resolution V design.
Q: How many runs does a Resolution V design require?
A: It depends on the number of factors, but Resolution V designs generally require substantially more runs than Resolution III or IV designs for the same number of factors, since preserving clean two-factor interaction estimates requires more information from the data.
Final Words
Design resolution is not a measure of whether an experiment is well-designed; it’s a precise statement of what a specific fractional factorial design can and cannot tell you apart.
Choosing the highest resolution your run budget supports, understanding the aliasing structure before interpreting results, and knowing that a foldover design can resolve ambiguity after the fact, are what separate practitioners who use fractional factorial designs confidently from those who get blindsided by a confounded result they didn’t see coming.
Choosing the right design resolution, and knowing what to do when a Resolution III result comes back ambiguous, is exactly the kind of applied DOE skill that separates a defensible screening study from a wasted one.
Six Sigma Development Solutions, Inc. (SSDSI) is IASSC-accredited and 5-star rated on Google Reviews, having certified 5,322+ professionals across 600+ organizations in 52 cities. Our onsite, live virtual, public, and online Black Belt training covers fractional factorial design, aliasing, and foldover techniques in the depth needed to apply them on real projects. Explore SSDSI’s Black Belt certification to build that capability.
About Six Sigma Development Solutions, Inc.
Six Sigma Development Solutions, Inc. offers onsite, public, and virtual Lean Six Sigma certification training. We are an Accredited Training Organization by the IASSC (International Association of Six Sigma Certification). We offer Lean Six Sigma Green Belt, Black Belt, and Yellow Belt, as well as LEAN certifications.
Book a Call and Let us know how we can help meet your training needs.
