Then show which entries were included, the number of comparable observations and the gaps. A precise calculation can still have a limited basis.
Define the comparison before calculating
Write down the signal, unit, local observation window and method. For example: measured drinking amount in mL over a complete day.
Keep the dates that belong to Period A and Period B. Do not label a half-day amount as a whole day. If the collection method changes, preserve that change rather than assuming the values remain comparable.
| Check | If the answer is unclear |
|---|---|
| Same signal? | Separate the series |
| Same unit? | Use an explicit, valid conversion or stop |
| Same collection window? | Do not compare as full-window amounts |
| Same measured/estimated context? | Keep separate or explain the limitation |
| Complete enough for this calculation? | Label partial or unknown entries; do not silently fill them |
This is an editorial comparison method, not a clinical protocol or PETICO's internal maintained-reference formula.
Worked example: fictional measured drinking
The following records are fictional, created to make the arithmetic checkable. They are not recommended amounts for a cat or dog.
| Period | Day | Recorded amount | Completeness | Included in primary mean? |
|---|---|---|---|---|
| A | Day 1 | 100 mL | Complete day | Yes |
| A | Day 2 | 110 mL | Complete day | Yes |
| A | Day 3 | 120 mL | Complete day | Yes |
| B | Day 1 | 140 mL | Complete day | Yes |
| B | Day 2 | 150 mL | Complete day | Yes |
| B | Day 3 | 80 mL | Partial day | No; retained separately |
A's mean is (100 + 110 + 120) ÷ 3 = 110 mL.
B's complete-day mean is (140 + 150) ÷ 2 = 145 mL.
The absolute difference, B minus A, is 35 mL.
Relative to A, the difference is 35 ÷ 110 × 100 = 31.82%, rounded to two decimal places.
The included counts are three complete days in A and two in B. The 80 mL partial day remains in the record. It is not converted to an estimated full-day amount and is not included in the primary mean.
Say what the result actually describes
A clear summary is: “For these fictional complete-day observations, B's mean is 35 mL higher than A's, a 31.82% difference relative to A. Coverage is three days versus two; one B entry is partial.”
That describes the records. It does not establish a trend, clinical normality, a cause or a maintained personal reference. Unequal day counts and missing observations remain relevant even when the calculation is correct.
Avoid three common shortcuts
A blank is not zero. A recorded zero can be a real observation. An empty field means there is no numeric observation available for this calculation.
A partial day is not a complete day. Do not multiply a half-day measurement to invent the rest of the day. Keep its amount and window visible.
An average is not automatically “usual.” A selected period can be unusual or poorly covered. Personal reference has a different evidential basis.
What if A's mean is zero?
You can still calculate the absolute difference. A relative percentage using A as the denominator is undefined when A is zero. Display “Percentage unavailable: Period A mean is zero,” not infinity or 100%.
If either period has no eligible complete numeric observations, do not return a mean or difference as zero. Say which period lacks eligible observations.
Keep a reusable record
Download the blank template and preserve raw entries before summarizing them. The worksheet uses this same fictional fixture and method.
If your question is about your pet's condition rather than the arithmetic, bring the observations and their limitations to a veterinarian. A comparison can help organize a question; it cannot answer every question about its cause.