An arithmetic sequence's common difference is the constant amount you add to get from one term to the next, and you find it by subtracting any term from the term right after it: d = a(n+1) − a(n). Because the value is constant across the whole sequence, any neighboring pair gives the same answer — 7, 11, 15, 19 has a common difference of 4, then 3, then 4. If two neighboring pairs disagree, the sequence is not arithmetic, and no single common difference exists. The one condition that trips people up: subtracting in the wrong order flips the sign, so always pick a consistent direction before you start.
The One Subtraction That Defines the Common Difference
Every arithmetic sequence follows the rule that a(n+1) − a(n) equals the same number for every index n. That number is d, the common difference. In the sequence 5, 12, 19, 26, subtracting the first term from the second gives 7; subtracting the second from the third gives 7 again. According to the standard treatment in OpenStax's College Algebra (Chapter 9, "Sequences, Probability and Counting Theory"), this constant-difference property is the defining feature of arithmetic sequences, and it is the test you apply to any list of numbers before doing anything else.
Pick the direction once and record it. If you write d = later term − earlier term, every subtraction runs left to right, and the sign means something: positive d means the sequence climbs, negative d means it falls. The same sequence read right to left (26, 19, 12, 5) has d = −7. Neither answer is wrong; they answer different questions about the same numbers.
The operation itself is ordinary integer or decimal subtraction — no special notation is needed, which is why this skill sits early in most algebra courses. What matters is that you compare adjacent terms. Subtracting the first term from the fourth term gives 3d, not d, and that confusion is the most common arithmetic error on this topic.
Memorize the shape of the definition, not a worked example: d = a(n+1) − a(n) for any n ≥ 1. Once that is in place, the method below is just careful bookkeeping.
Step-by-Step: Subtracting Neighbors to Get d
Here is the procedure in the order you actually perform it, assuming a list of terms and no formula given.
1: Write the terms in order and label them a1, a2, a3, and so on, so you never lose track of which number is which. 2: Subtract the first term from the second: a2 − a1. 3: Subtract the second term from the third: a3 − a2. 4: Compare the two results. If they match, that value is d. 5: Check at least one more pair (a4 − a3) before you commit to the answer on a test or in a spreadsheet formula. 6: If any pair disagrees, stop and reclassify the sequence.
A concrete run: for 3, 8, 13, 18, you get 8 − 3 = 5, 13 − 8 = 5, and 18 − 13 = 5, so d = 5. For 20, 17, 14, 11, you get 17 − 20 = −3, 14 − 17 = −3, so d = −3. The sequence 81, 64, 47, 30, 13 gives 64 − 81 = −17 and 47 − 64 = −17, confirming d = −17 — the same −17 you would find in the classic "take 17 away each time" pattern.
Failure handling matters more than the happy path. Suppose a list reads 2, 6, 18, 54. The first subtraction gives 4, the second gives 12. They disagree, so there is no common difference; that list is geometric (each term times 3), not arithmetic. Do not average 4 and 12 to force a number — averaging produces a value that fails on the next pair. If a sequence contains a fraction like 4, 9/2, 5, 11/2, convert everything to the same denominator first (8/2, 9/2, 10/2, 11/2), then subtract; the answer is 1/2.
Reading d Straight Off a Formula and a Function
When the sequence is given as a formula rather than a list, you do not need to subtract at all — the coefficient on n is d. For the explicit formula a(n) = 4 + 3(n − 1), expand it to 3n + 1; the multiplier 3 is the common difference. The general explicit form a(n) = a1 + d(n − 1) encodes d directly, a convention documented across standard algebra references such as Khan Academy's arithmetic sequences unit and OpenStax's College Algebra.
Recursive notation tells you just as plainly. The statement a(n) = a(n − 1) + 6 with a(1) = 2 hands you d = 6 in the "+6" term. Recursive form always names the step size explicitly, which is why some teachers introduce it first.
Linear functions wear the same costume. If a function rule is f(x) = 11 − 2x, the slope −2 is the common difference when x runs 1, 2, 3, 4: f(1) = 9, f(2) = 7, f(3) = 5, and 7 − 9 = −2. Slope and common difference are the same idea in two notations, and every arithmetic sequence graphs as a straight line with slope d when plotted against term number.
One warning: the coefficient is d only after the formula is in the a1 + d(n − 1) shape. A form like a(n) = 2(3n + 5) hides the step size until you distribute — 6n + 10, so d = 6 — which is exactly why a quick expansion is worth the ten seconds.
Missing and Fractional Terms: Working Backward Safely
When a term is missing — say 6, 10, __, 18 — use the known pairs on either side before filling the blank. Here 10 − 6 = 4, so the missing term is 14, and the check 18 − 14 = 4 confirms it. This is the safest pattern: establish d from a complete pair, apply it, then verify against the next complete pair.
If the blank sits at an end, regress instead. The list __, 15, 22, 29 has d = 7 (22 − 15), so the term before 15 is 8. Working backward means subtracting d instead of adding it, and the sign rules apply exactly as they do going forward.
Non-integer differences are common and deserve a template. For 1/3, 5/6, 4/3, 11/6, put everything over sixths: 2/6, 5/6, 8/6, 11/6. Each step is 3/6, so d = 1/2. Find a common denominator first, subtract, then simplify — doing it in the other order invites arithmetic mistakes. Negative fractions follow the same route: 0.5, 0.25, 0, −0.25 subtracts to −0.25 per step, and it checks three times.
Two-term problems need one more idea. Given only a3 and a7, divide the gap by the number of steps: a7 − a3 spans 4 steps, so d = (a7 − a3) ÷ 4. If a3 = 10 and a7 = 30, d = 5, and the sequence runs 0, 5, 10, 15, 20, 25, 30. Count steps between terms, not terms themselves — an off-by-one here is the second most common error after sign confusion.
When There Is No Common Difference: Sorting False Patterns
A disagreement between two neighboring pairs is decisive. Consider 1, 4, 9, 16: 4 − 1 = 3, 9 − 4 = 5, so the list fails immediately — it is the square numbers. The Fibonacci-style run 1, 1, 2, 3, 5 gives 0, then 1, then 1; no constant step exists.
Some near-misses fool people. The list 2, 4, 8, 16 gives 2, then 4 — a doubling pattern, not a constant step. A list that wobbles, like 5, 9, 13, 16, 20, is arithmetic for three pairs and then breaks at 16 − 13 = 3 versus 20 − 16 = 4. That single break disqualifies it.
Per a convention used across secondary-school curricula, the tell for geometric sequences is a constant ratio instead: each term divided by the previous one gives the same number. Testing both operations takes seconds and settles the classification.
Rechecking your own subtraction is the cheapest defense. If the two results differ by exactly 1 or 2, suspect a mental arithmetic slip before concluding the sequence is non-arithmetic; redo the subtraction on paper. False negatives from sloppy subtraction are more common than genuinely irregular sequences in textbook exercises.
How to Find the Common Difference in an Arithmetic Sequence: The Formula in Action
The nth-term formula does the heavy lifting once d is known: a(n) = a1 + d(n − 1). Take a1 = 7 and d = 4. Then a(10) = 7 + 4(9) = 43, and a(50) = 7 + 4(49) = 203. Notice the pattern: each jump of 10 in n adds 40 to the term, because 10 steps of 4 is 40 — a relationship you can use to sanity-check any answer.
The formula also reverses. Given a(1) = 3 and a(1) = 3, supposing a(20) = 79, then 79 = 3 + 19d, so 19d = 76 and d = 4. Solving for d this way is the standard route when neither endpoint term's neighbors are listed.
Table the relationships so the algebra stays visible:
| What you know | Rule | Worked result |
|---|---|---|
| Two adjacent terms | d = a(n+1) − a(n) | 31 − 26 = 5 |
| First term and 12th term, a1 = 4, a12 = 81 | d = (a12 − a1) ÷ 11 | 77 ÷ 11 = 7 |
| Explicit formula 4 + 3(n − 1) | coefficient on n | d = 3 |
| Recursive a(n) = a(n−1) + 6 | the added constant | d = 6 |
| Terms 6 and 18, four steps apart | d = (18 − 6) ÷ 4 | d = 3 |
| Third and seventh term, a3 = 10, a7 = 30 | d = (a7 − a3) ÷ 4 | d = 5 |
Read the third row carefully: a12 − a1 spans 11 steps, not 12, and that off-by-one is the row most likely to be marked wrong on a quiz. The last row shows the same rule generalized — subtract the indices to get the step count, then divide.
The table also settles a common question. An explicit formula's coefficient equals d only when the formula is fully simplified; a form like a(n) = 5 + 2(n − 3) simplifies to 2n − 1, and the coefficient 2 is d, while the constant −1 is not a term of the sequence at all.
Spotting the Sequence in the Wild: Tables, Graphs, and Paychecks
A table of term number against value hides an arithmetic sequence well until you subtract down the value column. In a table where term 1 maps to 9, term 2 to 7, term 3 to 5, the column differences are −2 each time, so d = −2 and the explicit rule is f(x) = 11 − 2x. Checking the formula against the table's first cell is the fastest verification path.
Graphs give it away visually: an arithmetic sequence plots as evenly spaced points along a straight line, and the vertical rise between consecutive points is d. Spacing that widens — as in 1, 4, 9, 16 plotted as points — is the visual signature of a non-arithmetic sequence.
Real-world phrasing is where students lose marks. A job paying $80 on day one with a $5 raise each day is an arithmetic sequence with a1 = 80 and d = 5, so day 10 pays 80 + 5(9) = $125. A car worth $25,000 that loses $2,000 annually gives d = −2,000, and after 6 years it is worth 25,000 − 2,000(6) = $13,000. Store both: dollars added is a positive d, depreciation is a negative d, and the formula does not care which.
Watch for sequences where the rate is a percentage. A $1,000 balance growing 4% a year adds $40, then $41.60 — the increments change, so the balance is not arithmetic. That single observation prevents a whole class of misapplied formulas.
Pitfalls That Cost Marks: Sign Order, Index Counts, and Mixed Units
The wrong-order subtraction is the biggest single cause of lost points. For 26, 19, 12, 5, computing 26 − 19 gives +7 while 19 − 26 gives −7; only one matches the direction you declared. Decide the direction in writing before the first subtraction and the ambiguity disappears.
The second pitfall is miscounting indices between terms. Subtracting a1 from a4 spans three steps, so dividing by 3 gives d; dividing by 4 gives a number that fails on every subsequent pair. A quick way to remember: the number of subtractions you can perform between two terms equals the difference of their indices.
Mixed units break more real-world problems than either algebraic error. A sequence alternating 1.5 hours and 90 minutes contains identical values written two ways; convert everything to one unit before subtracting or you will compute d = −0.4 or some other artifact. Similarly, a list mixing dollars and cents — 2.5, 3.75, 5.00 — needs consistent decimal places so the step reads as 1.25, not 1.2.
Rounding is the fourth trap. If a sequence carries d = 1/3 and you record 0.33, the fourth term drifts by 0.01 from the true value; over twenty terms it drifts by 0.07. Use exact fractions during working and round only the final answer.
FAQ
What is the common difference of 7, 15, 23, 31?
Subtracting each term from the next gives 8 every time, so d = 8. The nth term follows a(n) = 7 + 8(n − 1), which means the 25th term is 7 + 8(24) = 199. Three consistent pairs is enough evidence to conclude the sequence is arithmetic. If the fourth pair disagreed, the list would need reclassification.
Can the common difference be zero?
Yes — a constant sequence such as 4, 4, 4, 4 is arithmetic with d = 0. It satisfies the definition perfectly because every subtraction yields the same number, zero. Textbooks sometimes exclude such sequences for tidiness, but the standard definition in OpenStax's College Algebra admits them. Recognizing d = 0 prevents you from misfiling a constant list as non-arithmetic.
How is an arithmetic sequence different from a geometric one?
An arithmetic sequence has a constant difference between terms, while a geometric sequence has a constant ratio. In 2, 6, 18, 54 the differences are 4 and 12 (not constant), but each term is three times the last, so the ratio 3 is constant. Test both operations on the first two pairs; whichever is constant names the sequence.
Why do I get 3 when the answer says 4, for 3, 7, 11, 15?
You likely subtracted the fourth term from the third (11 − 15 = −4) or pulled one term out of position. Subtract in a single declared direction: 7 − 3 = 4, 11 − 7 = 4, 15 − 11 = 4. A result that differs by exactly 1 often signals an indexing slip rather than a sign error. Recount the labels before redoing the subtraction.
What if the first pair gives 5 and the second gives 6?
Then the sequence has no common difference and the task changes to classification. Averaging 5 and 6 to get 5.5 produces a number that fails on the third pair, so it is never the correct move. Look for a constant ratio or a second-level pattern instead.
How do I find d from only two terms, like a2 = 9 and a5 = 24?
Subtract the indices to find the step count — 5 − 2 = 3 steps — then divide the term gap by that count: (24 − 9) ÷ 3 = 5. So d = 5, a1 = 4, and the sequence runs 4, 9, 14, 19, 24. Dividing by 4 instead of 3 is the standard error here, so always subtract indices rather than count terms.
Is there a quicker way than subtracting every pair?
Subtract the first two terms, then verify with one later pair near the end of the list; two well-separated checks catch nearly all errors. On a classroom quiz, a full pairwise check is usually unnecessary and a single end-of-list verification is faster. In a spreadsheet, the formula =B3-B2 copied down the column flags any mismatch instantly.
Can an explicit formula have a hidden common difference?
Yes, whenever the formula is not in simplified a1 + d(n − 1) form. The rule a(n) = 5 + 2(n − 3) expands to 2n − 1, so d = 2 even though the constant "5" appears nowhere in the slope role. Distribute first, then read the coefficient on n.