What is a Dixon Q test?
The Dixon Q test is a quick statistical method for identifying a potential outlier in a small sample. It compares the “gap” between a suspicious value and its nearest neighbor against the full data range.
This tool is useful in laboratory work, quality control, calibration checks, and any setting where you have only a handful of repeated measurements and want a principled way to decide whether one value looks inconsistent with the rest.
How this Dixon Q test calculator works
After you enter your values, the calculator sorts them from smallest to largest. It then computes:
For a low-end candidate: Qlow = (x2 - x1) / (xn - x1)
For a high-end candidate: Qhigh = (xn - xn-1) / (xn - x1)
It then compares your Q value to a critical Q value based on sample size (n) and confidence level (90%, 95%, or 99%). If Qcalc > Qcrit, the candidate value is flagged as a possible outlier.
Step-by-step workflow
- Enter your measurements (3 to 30 values).
- Select confidence level (95% is common).
- Choose lowest, highest, or auto-check.
- Click Calculate Dixon Q.
- Review whether a value is statistically flagged.
Interpreting your result correctly
A flagged value means the number is statistically unusual under the Dixon Q criterion. It does not automatically prove the value is “wrong.” You should still use scientific judgment:
- Was there a known instrument issue?
- Was sample prep inconsistent?
- Is there transcription or unit-conversion error?
- Can the measurement be repeated?
When to use (and not use) Dixon Q
Good use cases
- Small datasets where n is between 3 and 30.
- Roughly single-peaked data with one suspected extreme value.
- Experimental replicates from the same process or condition.
Use caution or alternatives when
- You have large sample sizes (consider robust methods or other tests).
- You suspect multiple outliers at once.
- The data comes from mixed populations.
- The distribution is heavily skewed by design.
Worked example
Suppose your measurements are: 2.01, 2.03, 1.99, 2.02, 2.00, 2.04, 2.31. The top value (2.31) appears suspiciously high.
The calculator computes the high-end Q, compares it to the critical value for n = 7 at your chosen confidence level, and tells you whether 2.31 is statistically flagged. If it is flagged, you can investigate root cause and, if justified, report analyses with and without that value.
Frequently asked questions
Does this remove the outlier automatically?
No. It flags a potential outlier. Removal should be documented and scientifically justified.
Can I test both ends at once?
Yes, use auto-check. If both ends appear significant, investigate carefully; Dixon Q is typically applied to one suspected outlier at a time.
Why does the calculator reject my input?
Common reasons are non-numeric entries, fewer than 3 values, more than 30 values, or all values being identical (zero range).
Bottom line
This Dixon Q test calculator gives you a fast, transparent way to screen small datasets for extreme values. It is best used as part of a complete quality-review process, not as a substitute for domain expertise.