How to Find a Z-Score From a Percentile (Step-by-Step)
The short answer: finding a z-score from a percentile means running the z-table in reverse. Instead of looking up a z to get an area, you search the table's body for the area (percentile ÷ 100) and read the z off the edges. The 95th percentile gives z = 1.645. The 90th gives z = 1.28. The 50th gives z = 0.
Here's the full method, because exam questions love this direction.
The 4 steps
Step 1 — Turn the percentile into a decimal area. The 95th percentile means 95% of the distribution lies at or below your value, so the left-tail area is 0.9500. Just divide by 100.
Step 2 — Find that area inside the z-table. Scan the body of the table (the grid of probabilities) for the value closest to yours. This feels backwards the first time — you're searching the middle of the table, not the edges. If you're shaky on table layout, our walkthrough on how to read a z-table shows the anatomy first.
Step 3 — Read the row and column. The row gives the z-score's whole-and-tenths part (1.6), and the column gives the hundredths digit. For 0.9500, the closest table entries sit at columns 0.04 and 0.05 in row 1.6, so the accepted value splits the difference: z = 1.645.
Step 4 — Attach the sign. Percentile above 50 → positive z. Below 50 → negative z. Exactly 50 → z = 0.
Or skip the paper entirely: the percentile to z score calculator does the inverse lookup instantly and shows the shaded area.
Worked example 1: the 95th percentile
Area = 0.9500. Searching the z score table body, 0.9500 sits at row 1.6, between columns 0.04 (0.9495) and 0.05 (0.9505). The textbook answer splits the difference: z = 1.645. Above the 50th percentile, so positive. Done — 95% of values fall below 1.645 standard deviations above the mean.
Worked example 2: the 25th percentile (the mirror trick)
Area = 0.2500, which lives in the negative half of the table — awkward if your printed table only shows positive z. Use symmetry instead:
- Mirror it: 100 − 25 = 75th percentile, area 0.7500.
- Look up 0.7500 → z = 0.674.
- Flip the sign: z = −0.674.
Check it makes sense: the 25th percentile is below the mean, so the z-score must be negative, and it's less than one standard deviation away, so a size under 1 is right. Always run that two-second sanity check — sign errors are the number-one way this method loses marks (see our rundown of z-score vs percentile confusions).
Worked example 3: turning a percentile cutoff into a real score
This is where the method earns its keep. A scholarship requires the top 10% on an entrance exam. Exam scores: mean 70, standard deviation 8. What raw score qualifies?
- Top 10% = 90th percentile → z = 1.28.
- Reverse the z formula: x = μ + zσ = 70 + (1.28 × 8) = 70 + 10.24 = 80.24.
So a student needs about 80 or above. Same machinery powers the critical z score calculator used for confidence levels and significance cutoffs — 90%, 95%, and 99% confidence all start life as percentile lookups.
Translating exam phrasing into lookups
Half the battle is recognizing that a question is asking for this at all. Exam writers rarely say "find the z-score for this percentile" in so many words. They say things like:
- "The top 10% of applicants are accepted." Top 10% means the cutoff sits at the 90th percentile → z = 1.28.
- "The bottom quarter of the class must retake the module." Bottom 25% = the 25th percentile → mirror-trick lookup → z = −0.674.
- "The middle 95% of measurements are within tolerance." Middle 95% leaves 2.5% in each tail → the boundaries are z = ±1.96, not 1.645.
- "Scores above the 84th percentile earn a merit mark." No lookup needed if you know your anchors: the 84.13th percentile is z = 1.00 exactly.
Train yourself to convert the words into a percentile first, then run the four steps. Most wrong answers in this topic aren't lookup errors — they're translation errors, where someone looked up the 95th percentile when the question described the middle 95%.
The percentiles worth memorizing
| Percentile | Z-score |
|---|---|
| 50th | 0 |
| 75th | 0.674 |
| 84.13th | 1.00 |
| 90th | 1.28 |
| 95th | 1.645 |
| 97.5th | 1.96 |
| 99th | 2.326 |
Notice 1.96 again: it's the 97.5th percentile, not the 95th. It shows up everywhere because two-tailed 95% tests leave 2.5% in each tail. If a problem says "95th percentile," answer 1.645. If it says "95% in the middle," answer ±1.96. That distinction alone is worth a few exam marks.
Frequently asked questions
What is the z-score for the 95th percentile?
The z-score for the 95th percentile is 1.645. That means 95% of values in a normal distribution fall at or below 1.645 standard deviations above the mean, and 5% fall above it. Don't confuse it with 1.96 — that's the 97.5th percentile, used for two-tailed 95% cutoffs.
What is the z-score for the 90th percentile?
The z-score for the 90th percentile is about 1.28. More precisely it's 1.2816, which most tables round to 1.28. Scholarship and admissions cutoffs phrased as "top 10%" use this value: convert it back with the mean and standard deviation to get the actual score a student needs.
How do I find the z-score for a percentile below 50?
Find the z-score for the mirror percentile above 50, then flip the sign. For the 25th percentile, look up the 75th (100 − 25), which gives z = 0.674, then negate it: −0.674. The bell curve's symmetry guarantees this works for every percentile under 50.
Why isn't my percentile listed exactly in the z-table?
Tables only print probabilities to four decimal places in fixed steps, so most percentiles fall between two entries. Pick the closest value — for the 95th percentile, 0.9500 sits right at z = 1.645 in most tables. Coursework accepts the nearest entry; software interpolates for extra precision.