T-Score Calculator
The formula
What the t-score tells you
The mean-difference signal measured against noise, calibrated for small samples by the t-distribution’s heavier tails.
Calculation: t = (M₁−M₂) ∕ (s_p·√(2∕n)); df = 2n−2 for equal groups. Enter your inputs and every substitution is shown below the result.
Reading the result
At n=30 per group, the critical bar is ~2.04 not 1.96 — the t test’s honesty about small-sample uncertainty is its whole point.
Rules of thumb: α=.05 two-sided; report t, df, and the exact p, not just stars
Limitations and best practice
The pooled-SD two-sample t assumes roughly equal variances; with a 2:1 SD ratio switch to Welch, which is the default for good reason.
This calculator gives an estimate based on standard formulas; it is not professional financial, medical, legal or engineering advice. Verify anything you plan to act on with a qualified professional and your own records.
Frequently asked questions
When does t become z?
Is one-sample or two-sample t more common?
Source & methodology
Student (1908); standard inferential statistics.
T-Score explained: formula, worked example and honest limits
You want to know what the number requires, not just what it is; this page shows both. Searchers land on a T-Score page for three reasons — to plan, to compare options, and to check that somebody else’s arithmetic is honest. This tool serves all three, and everything after the calculator explains the model instead of hiding it.
What makes T-Score worth calculating is not precision but position: it tells you which side of a threshold you are on, and by how much. A rough figure that settles the direction beats an elaborate one that never commits.
Most confusion about T-Score comes from mixing up the input you control with the input you merely observe. The formula treats them identically, but your decisions should not: one deserves measurement, the other deserves a defensible estimate.
Nothing below is stock text pasted around a number: the worked example computes independent t-statistic = -5.03 from this tool’s own defaults, and the sensitivity section ranks the inputs that move the result — Sample 2 mean first, at ≈33.5 % of the headline for a ten-percent error either way in that input. Work down the page and by the end you could re-derive the result in a spreadsheet from memory.
The T-Score formula, explained
One formula, published in full: t = (M₁−M₂) ∕ (s_p·√(2∕n)); df = 2n−2 for equal groups. Anything you can read, you can reproduce in a spreadsheet, and that reproducibility is the whole point of printing it rather than hiding it behind a button.
Sample 1 mean (example value 61) is a measured number in the unit named by the label; the formula does not convert for you. Replace the example with the figure from your own records — the tool recomputes instantly and the printed steps follow your numbers.
Sample 2 mean (example value 87) is a plain measured quantity — the unit shown in the label is the unit the formula expects. The example value is there to demonstrate the mechanics, not to describe your situation — overwrite it.
Pooled SD (example value 20) is a measured number in the unit named by the label; the formula does not convert for you. If you are unsure what belongs in this field, look at the printed steps: the substitution shows exactly how the engine uses it.
Per group size (example value 30) is a numeric input whose unit is fixed by the label; a wrong unit is a wrong answer wearing a plausible disguise. Keep every length field in the same unit; the formula assumes a consistency it cannot check for you.
T-Score with real numbers, step by step
The tool prints its own working; this is the same arithmetic on the example values, reproduced so you can follow it by hand. Starting figures: Sample 1 mean, Sample 2 mean, Pooled SD, Per group size. The endpoint of that chain is -5.03 of independent t-statistic, sitting in the “t(58) = -5.03, two-sided p ≈ <0.0001 — significant at the 5% convention” band.
- SE = 20.0 × √(2/30) = 5.16
- t = (61.0 − 87.0) ÷ 5.16 = -5.03 on 58 df
Standard error of difference — 5.16. This is the number experienced users quote when the headline alone would mislead.
Effect size (Cohen’s d) — -1.30. Read next to the headline, this is the figure that shows the shape of the trade-off.
Critical t at α=.05 (≈1.96+) — 2.00. This is the number experienced users quote when the headline alone would mislead.
95% CI of difference — -26.00 ± 10.12. If your inputs change, this is usually the first companion figure to move.
Overwrite any input above and every line of this example updates with it — the substitution is live, not a screenshot of somebody else’s numbers.
How far T-Score can reasonably move
Single-point answers invite false confidence, so the same engine was run across a band around the input that matters most here. Each row is a full recalculation of T-Score, not an interpolation.
| Sample 2 mean setting | Independent t-statistic |
|---|---|
| 80 % of the example value | -1.67 |
| 90 % of the example value | -3.35 |
| example inputs | -5.03 |
| 110 % of the example value | -6.72 |
| 120 % of the example value | -8.40 |
That spread — -8.40 at the low end, -1.67 at the high — is what a careful read of T-Score looks like: a band you can state out loud, with the reason for each edge. Reproduce it in the form above by editing Sample 2 mean alone; the substitution steps will follow your figures exactly.
What actually moves your T-Score
This list was computed for this specific engine just now: each numeric input was nudged ±10 % while every other field held its example value, and the independent t-statistic was recalculated from scratch. Read it as a map of where T-Score is fragile — drawn for the example inputs, which is the map you want when deciding what to measure carefully.
- Sample 2 mean: ten percent up gives -6.72, ten percent down gives -3.35 — ≈33.5 % of the independent t-statistic either way
- Sample 1 mean +10 % → T-Score = -3.85 · −10 % → -6.22 (the independent t-statistic shifts ≈23.5 %)
- Pooled SD: ten percent up gives -4.58, ten percent down gives -5.59 — ≈11.1 % of the independent t-statistic either way
- Per group size: ten percent up gives -5.28, ten percent down gives -4.78 — ≈5.1 % of the independent t-statistic either way
Sample 2 mean is the dominant lever here: ten percent either way moves the result by about 33.5 %, from -6.72 to -3.35. Spend your verification effort there first, and treat Sample 1 mean as the second priority (≈23.5 % per ten percent). The rest of the inputs earn only a rounding-error of attention, and that asymmetry is the practical payoff of measuring rather than asserting.
Reading the result: bands, zones and thresholds
Nobody acts on a bare figure — they act on a figure with a comparison attached. On the example inputs the band reads “t(58) = -5.03, two-sided p ≈ <0.0001 — significant at the 5% convention”; your own figures place you on a different rung of the same ladder. The sensitivity list shows how far a modest input error moves you along that ladder: for Sample 2 mean, ±10 % covers -6.72 to -3.35. The practical reading rules: ask what would have to change to move you one band. If the answer is “a small, plausible input difference”, the band itself is not the finding — the uncertainty is.
In Science & Research, T-Score is one instrument in an orchestra: on its own it is a number, next to Standard error of difference and Effect size (Cohen’s d) it becomes a diagnosis. The tools linked above exist so the headline never has to carry that weight alone.
Why T-Score results go wrong — and how to prevent it
The pooled-SD two-sample t assumes roughly equal variances; with a 2:1 SD ratio switch to Welch, which is the default for good reason. This calculator gives an estimate based on standard formulas; it is not professional financial, medical, legal or engineering advice. Verify anything you plan to act on with a qualified professional and your own records.
Reporting the headline and quietly discarding Standard error of difference. The companions are where the shape of the trade-off survives the rounding.
This model has its own particular trap: Sample 2 mean carries ≈33.5 % of the response per ten percent of error, so effort spent elsewhere is largely decorative.
Another is comparing across contexts: a T-Score computed with one set of conventions should never be lined up against a figure computed with another. Check the inputs behind both before ranking them.
Before you trust your T-Score: a short checklist
- Re-run the calculation with a ±10 % stress on the two most sensitive inputs.
- Note the date and the inputs beside the number so the figure can be audited later.
- Check Sample 2 mean before anything else: it is the measured dominant input, ≈33.5 % of T-Score per ten percent.
- Confirm the example values are plausible for your own situation before replacing them.
- Check that every field the formula needs actually arrived from a document, not from recollection.
- Repeat the ±10 % test on the inputs that matter and see whether the decision — not just the number — still holds.
Step by step: getting a trustworthy T-Score out of this tool
Using a calculator well sounds trivial, but the difference between a decision-grade figure and a throwaway one lives entirely in the routine. The routine for T-Score takes a minute to learn and applies to every input above.
Step 1 — start from evidence. The inputs behind T-Score are sample 1 mean, sample 2 mean, pooled sd, per group size. Write them down with their source next to each one; if you cannot name where a figure came from, that is the input most likely to be wrong. When two sources disagree, it is almost never the arithmetic that differs.
Step 2 — baseline before exploration. Get one defensible Independent t-statistic on screen before you start nudging numbers. Without a baseline, every variation looks interesting and none of them means anything. If the verdict changes with a ten-percent nudge, report the nudge alongside the number.
Step 3 — play with the inputs on purpose. Ten percent up, ten percent down, one field at a time. Within a few minutes you will know this calculation better than most people who quote its results. Numbers are cheap to produce and expensive to retract; the extra minute is the cheapest insurance available.
Step 4 — ask what would have to be true. Take the result you want and work backwards through the inputs until you find the assumption that produces it; then decide honestly whether that assumption is defensible. Say the answer out loud before you save it; if it sounds wrong, the model did not change, your understanding did.
Step 5 — save the story with the figure. A number without its inputs and its date is unusable within a quarter. Record -5.03, the assumptions behind it, and today’s date; your future self will be auditing a stranger’s arithmetic. Estimates improve most when the worst case is written down in ink.
Who actually uses T-Score — and when
Science & Research arguments end at arithmetic or they do not end. T-Score is the arithmetic: -5.03 on the example inputs, reproducible by anyone who types the same numbers. The input you reach for first is usually the one you measured longest ago.
The companion outputs — Standard error of difference and Effect size (Cohen’s d) — exist for the second stage of that screen: the headline sizes the prize, while the companions describe the trade-off you are accepting to get it. Say the answer out loud before you save it; if it sounds wrong, the model did not change, your understanding did.
Re-running this page costs nothing and catches the slow drifts that never announce themselves. In Science & Research the expensive mistakes are rarely sudden; they are small changes in Sample 1 mean that nobody recalculated for a year. Compare the calculation to the rough guess you arrived with, and ask why they differ.
Quick answers: what searchers ask about T-Score
What affects T-Score the most?
The short answer: Sample 2 mean. The longer answer is the sensitivity table on this page, where every numeric input was probed by ten percent and the independent t-statistic recalculated — Sample 2 mean moved it furthest, by roughly 33.5 % (-6.72 to -3.35).
Which matters more: sample 1 mean or sample 2 mean?
For T-Score, Sample 2 mean wins the head-to-head (≈33.5 % against ≈23.5 % per ten-percent change). Knowing which input dominates is the difference between careful work and merely slow work.
What is a typical T-Score?
There is no honest universal average for T-Score, but there is a reference point: the defaults on this page produce -5.03 (t(58) = -5.03, two-sided p ≈ <0.0001 — significant at the 5% convention). Replace the inputs with your own and compare — that comparison is the useful meaning of “typical”.
How do I calculate T-Score by hand?
Use the stated model — t = (M₁−M₂) ∕ (s_p·√(2∕n)); df = 2n−2 for equal groups — and substitute in the order the tool prints (it starts from SE = 20.0 × √(2/30) = 5.16). On the example inputs the hand route lands on -5.03, exactly what the calculator shows. A spreadsheet reproduces it in two or three cells; doing it once by hand is what makes the automated answer trustworthy.
What happens to T-Score if sample 2 mean changes by 10 %?
The engine was re-run with that field moved by ten percent while everything else stayed fixed: the independent t-statistic goes across -6.72 to -3.35, a swing of roughly 33.5 % around the -5.03 baseline. That ratio is the practical definition of how much this input matters.
Is the T-Score result of -5.03 good or bad?
Neither, on its own — the label “t(58) = -5.03, two-sided p ≈ <0.0001 — significant at the 5% convention” is the band this figure falls in for the example inputs, and your own inputs will place you elsewhere on the same ladder. Judge the number against your history and your alternatives, and remember that a ten-percent error in Sample 2 mean alone moves it by ≈33.5 %.
Can T-Score be negative or empty?
Yes, and that is informative rather than broken: a negative T-Score usually means the inputs encode an outflow or a shortfall, and the verdict band above labels exactly that situation instead of hiding it.
What affects T-Score the most?
Sensitivity depends on where you start, but the inputs Sample 1 mean and Sample 2 mean typically move T-Score further than the rest because they enter the formula multiplicatively. If you can only measure two things well, measure those and hold the others at conservative defaults.
When does t become z?
Effectively past n≈120 per group; the distributions are identical above a few hundred.
Is one-sample or two-sample t more common?
Two-sample dominates practice; one-sample needs a meaningful external reference mean to interpret against.
Can this tool see the numbers I type?
It cannot. There is no submission step: the calculation happens in your browser and the result disappears when you close the page. Treat the fields above as a scratchpad rather than a form.
The vocabulary of T-Score
- Standard error of difference
- 5.16 for the T-Score worked case; it sits next to the main result because it carries context the headline leaves out.
- Effect size (Cohen’s d)
- -1.30 on the example inputs — a companion figure printed beside the headline of T-Score so that no result is read in isolation.
- Critical t at α=.05 (≈1.96+)
- 2.00 in the default T-Score calculation — one of the secondary figures an experienced reader checks before trusting the primary number.
- Independent t-statistic
- The primary output of this tool: -5.03 for the default inputs, computed exactly from t = (M₁−M₂) ∕ (s_p·√(2∕n)); df = 2n−2 for equal groups.
- Worked example
- Every substitution in order, from the example inputs to -5.03 — the audit trail for the headline figure.
- Verdict band
- The interpretation attached to a result (“t(58) = -5.03, two-sided p ≈ <0.0001 — significant at the 5% convention” on these inputs), which turns a bare number into a position on a scale.
- Sensitivity (±10 %)
- How far the T-Score moves when a single input is nudged ten percent either way; the ranked list on this page is computed live from this engine.
Where the T-Score model comes from
The model is published arithmetic implemented exactly as written: t = (M₁−M₂) ∕ (s_p·√(2∕n)); df = 2n−2 for equal groups. Source of record: Student (1908); standard inferential statistics. Where a professional version of this calculation differs, it differs in inputs and jurisdiction-specific coefficients, not in the algebra — which is why re-checking your inputs beats hunting for a second calculator.
Treat the output as a precise answer to the question you typed: exact arithmetic, transparent formula, published source, no hidden rounding. Nothing is smoothed, weighted or adjusted behind the scenes; where judgement enters, this page says so and names the input responsible.
Next steps after the calculation
Work backwards from your target: adjust the inputs until the result matches the score you need, and those inputs become the study plan.
The audience here is students and candidates planning toward a target score. Everything you need to defend the result — formula, substitution, sensitivity, source — is printed above, so the only remaining variable is the honesty of your inputs.