How to Find Rate of Change From a Table: Step-by-Step Method With 4 Worked Examples

To find the rate of change from a table, subtract the output values and divide by the difference in input values between two rows: (y₂ − y₁) ÷ (x₂ − x₁). The table must show equal input spacing for a single slope to describe the whole thing. If the x-values jump by the same amount each row and the quotients come out identical, that number is your rate of change. If the quotients differ, you're looking at a nonlinear pattern, and the honest answer is an average rate over the interval you picked. The two-row subtraction works on any table, linear or not.

That's the whole method in miniature. What follows is the version you can use on homework, on a spreadsheet, and on a real data set without getting tripped up by the three mistakes that cause most wrong answers: subtracting in the wrong order, mismatching rows, and assuming linearity that isn't there.

Pick Any Two Rows and Label Them in the Same Order on Both Sides

The rate of change formula is (change in y) ÷ (change in x), and the order you subtract in matters only in that you must subtract the same direction on both sides. If row 2 is your second point, it goes first in both subtractions. Flip both and the sign flips with them; flip only one and you've invented a wrong answer.

Say a table shows a phone plan's monthly cost: at 2 GB of data the bill is $38, and at 5 GB it's $53. Label 5 GB and $53 as point two. Then (53 − 38) ÷ (5 − 2) = 15 ÷ 3 = $5 per gigabyte. Label them the other way and you get −15 ÷ −3, which still equals $5. The arithmetic is forgiving as long as you stay consistent.

What the sign tells you is direction. A positive result means the output grows as the input grows; a negative result means it shrinks. A table of a car's value by age gives (12,000 − 20,000) ÷ (4 − 1) = −$2,667 per year, and the minus sign is the depreciation itself, not an error to correct.

One habit prevents most mistakes here: write the two points down as ordered pairs before touching the formula. (2, 38) and (5, 53) are harder to mix up than loose numbers sitting in a table. Once they're written as pairs, the subtraction order is visible, and you can check it at a glance.

A quick consistency check costs five seconds. If your inputs increase left to right in the table, your denominator should come out positive when you subtract row two minus row one. A negative denominator is a flag that you reversed one side. Fix the order before you divide, not after.

Confirm the Inputs Step Evenly Before You Trust a Single Slope

A constant rate of change requires evenly spaced inputs. A table listing years 2000, 2001, 2002, 2005 has one gap of three years, and any slope you compute across that gap describes a three-year stretch, not a per-year rate. You can still compute a rate — just not one that matches the other rows.

Here's where a lot of students go wrong without noticing. Take a table of a savings account: $500 at month 0, $650 at month 3, $800 at month 6, $1,100 at month 12. The first two gaps give (650 − 500) ÷ 3 = $50 per month and (800 − 650) ÷ 3 = $50 per month. The last gap gives (1,100 − 800) ÷ 6 = $50 per month. Same answer, because the deposits were steady. But if the final entry were $1,250 instead, that last gap yields $75 per month, and the table would not be linear — the account got a raise in deposits somewhere between month 6 and month 12.

So the check is this: compute the quotient for every consecutive pair of rows, not just one. If they all match, you have a linear table and one rate describes everything. If they don't, the table is piecewise or curved, and the correct statement is "the average rate of change between month 6 and month 12 is $75 per month."

Print and online tables often omit units in the header row. Before computing, confirm what each column measures — months, years, dollars, thousands of dollars. A table of revenue in thousands turns a slope of 4 into $4,000 per unit, and that factor of 1,000 is the difference between a passing answer and a failing one.

Run the Four-Row Table: A Worked Example End to End

Take this table of a delivery van's odometer reading by hour:

Step 1: check spacing. Hours step by 1 each row. Good. Step 2: pick rows 1 and 2. (90 − 45) ÷ (2 − 1) = 45 ÷ 1 = 45 miles per hour. Step 3: verify with rows 3 and 4. (180 − 135) ÷ (4 − 3) = 45 ÷ 1 = 45. Step 4: state the answer with units — 45 miles per hour — and note that because every consecutive pair matches, the rate is constant across the whole table.

Now a table that isn't so tidy:

Rows 0 and 2: (180 − 120) ÷ 2 = 30 per week. Rows 2 and 4: (260 − 180) ÷ 2 = 40 per week. Rows 4 and 6: (320 − 260) ÷ 2 = 30 per week. Three different-looking quotients — 30, 40, 30 — mean no single rate fits. The overall average across the whole table is (320 − 120) ÷ 6 ≈ 33.3 subscribers per week, which is a legitimate summary but hides the mid-period surge.

The lesson from the second table is the one worth carrying: a rate of change from a table is always an average over the interval you chose, and the interval has to be named. "Thirty-three subscribers per week over six weeks" is a true statement. "The channel grows 33 subscribers per week" is not, because weeks 2 through 4 ran faster.

For a table where the inputs aren't consecutive integers, the same four steps hold. Sort by x first, then subtract. A table listing test scores by hours studied at 0, 1.5, 3, and 4.5 hours works identically: (78 − 60) ÷ (3 − 1.5) = 12 points per hour.

Use the Same Formula on Any Table With Two Columns of Numbers

Nothing about the method cares whether the table came from a textbook, a lab notebook, or a spreadsheet export. The formula is subtraction on both sides, division once. What changes is how carefully you read the headers.

Watch for tables where the input column runs downward instead of left to right. A chemistry log listing temperature from 80°C down to 20°C still works — you just have to keep the order straight. Pick (80, 12) and (20, 4), then (4 − 12) ÷ (20 − 80) = −8 ÷ −60 ≈ 0.133 units per degree. Both negatives cancel, and the positive result says the reading rises as temperature falls.

Ordinal or categorical headers — small, medium, large — don't give you a rate of change, because the spacing isn't numeric. Convert to a scale first (1, 2, 3) only if the problem says the categories are equally spaced. Otherwise the honest answer is that a rate of change isn't defined for that table.

Time-zone and calendar quirks matter more than people expect. A table keyed to fiscal quarters has uneven day counts — Q1 has 90 or 91 days, Q2 has 91 — so a per-quarter slope on daily revenue will wobble by a percent or two for reasons that have nothing to do with the business. Use the same unit throughout, and if the source mixes monthly and annual figures, convert before subtracting.

When a table includes a totals row, exclude it. Sums and averages are not data points, and including them skews every quotient that touches them. The same goes for "N/A" or blank cells: skip that row and compute between the rows on either side of it, and say so in your answer.

Compute a Slope From a Table in Google Sheets or Excel

Spreadsheets do the subtraction for you, and for long tables that's the difference between a five-minute answer and a forty-minute one. The formula assumes your x-values sit in column A and y-values in column B, starting at row 2.

1: Click an empty cell in row 3 or later, say C3, and type =(B3-B2)/(A3-A2). Press Enter. This gives the rate of change between the first two data rows. 2: Grab the fill handle at the bottom-right corner of C3 and drag down to the last row. Every cell now shows the rate between that row and the one above it. 3: Scan the column. If every value matches, the table is linear and you're done — that number is the rate of change. 4: If the values differ, wrap the whole-range version in one cell: =(B10-B2)/(A10-A2), replacing 10 with your last row. That's the average rate over the full table. 5: For a formal slope, use =SLOPE(B2:B10,A2:A10), which returns the same figure as step 4 and is the function most teachers expect to see.

Failure handling matters here. A #DIV/0! error means two consecutive x-values are identical, so a vertical step has no defined slope — check for a duplicated row or a typo in the input column. A wildly different value in one cell usually means a text-formatted number; Sheets and Excel both treat "1,200" stored as text as zero in arithmetic. Select the column, format it as a number, and recheck.

If you only have two rows of data, skip the fill-down entirely and type the formula with both cell references. If the table lives in a shared sheet and someone else may edit it, put the slope formula in a separate tab so a sort of the source data doesn't break your references.

Read a Negative Rate of Change Without Panicking

Negative slopes are ordinary. They mean the output falls as the input rises, and the formula reports the decline with a minus sign. A table of a laptop's resale value — $900 at year 1, $650 at year 2, $400 at year 3 — gives (650 − 900) ÷ (2 − 1) = −$250 per year, then (400 − 650) ÷ (2 − 1) again −$250. The resale value drops $250 each year on a straight line.

The sign is information, not a mistake. When you write the answer, keep the minus and translate it: "decreases by $250 per year." Dropping the sign changes a depreciation into an appreciation, which is a different claim entirely.

One confusion worth untangling: a negative rate of change is not the same as a negative value. A company losing money but losing less each quarter has negative profit with a positive rate of change. A table showing −$40,000 at Q1, −$30,000 at Q2, −$15,000 at Q3 gives (+10,000) ÷ 1 and (+15,000) ÷ 1 — the losses are shrinking, so the rate of change is positive even though every entry is below zero.

A table can also switch signs mid-stream. Revenue rising for two years then falling produces a positive slope up top and a negative one at the bottom. That's a maximum, and it's the reason graphing the points before concluding anything is worth the thirty seconds it takes.

Spot Nonlinear Tables Before You Write Down a Wrong Slope

A linear table has equal quotients for every pair of rows. A nonlinear one doesn't, and the tell is usually visible in the y-column before you divide: constant first differences mean linear, changing first differences mean not. The x-column, remember, has to be evenly spaced for this check to work.

Take a table of a falling object's distance: 0 seconds and 0 meters, 1 second and 4.9 meters, 2 seconds and 19.6 meters, 3 seconds and 44.1 meters. Differences are 4.9, then 14.7, then 24.5 — growing by 9.8 each step. That's a quadratic, and the 9.8 is gravity in meters per second squared. No single rate of change describes the fall; the speed itself is changing, which is exactly what acceleration means.

A doubling table shows the same nonlinearity with a different shape. A bacteria count of 100, 200, 400, 800 has differences of 100, 200, 400 — no constant quotient, and no slope worth quoting. Anything exponential will fool a quick average.

If your table fails the constant-quotient test, you still owe the reader a number, and the right one is the average rate over the stated interval. Say "the average rate of change between 2 and 3 seconds is 24.5 meters per second" rather than pretending the object held that speed. This is the same arithmetic, but the sentence around it is honest.

Check Your Answer With the Table's Own Numbers

Verification takes one line: rate × (change in x) + starting y should equal the ending y for any pair of rows in a linear table. Use the first row as the start. Slope 45 miles per hour over 3 hours from a starting reading of 45 miles gives 45 + 135 = 180, and the table's row-4 entry is 180. Match.

For the non-linear subscriber table, the check works only interval by interval. From week 2 with 180 subscribers, a rate of 40 per week over 2 weeks gives 180 + 80 = 260, matching the week-4 row. That's exactly why the overall average of 33.3 fails to reproduce the table: it isn't the rate for any single stretch.

A second check catches unit errors. Write the units into the division: (dollars) ÷ (gigabytes) = dollars per gigabyte. If the units reduce to something nonsensical, like dollars per dollar, the columns got swapped somewhere.

Keep the two-row habit even when a spreadsheet gives you the answer. The manual subtraction is what lets you defend the number when someone asks where it came from, and it's what you'll do on a test with no laptop open.

FAQ

Is the rate of change the same as slope?

Yes for straight-line tables — slope and rate of change are the same quantity, (y₂ − y₁) ÷ (x₂ − x₁), and both carry the units of y per unit of x. The phrase "rate of change" is more common in applied problems like speed or cost per unit. "Slope" tends to appear in pure graphing contexts. For a nonlinear table, slope describes only the tangent at a point, while average rate of change covers an interval.

What if the table has no consistent x-spacing?

You can still compute a rate between any two rows, but you must name the interval, because each row-to-row quotient covers a different span. The standard fix is to divide by the actual gap rather than assume 1. A jump from x = 2 to x = 5 divides by 3, not by 1, and skipping that step inflates the answer by a factor of three.

Does a rate of change of zero mean something is wrong?

No. A zero rate means the output is constant across the interval — (the same y-value twice) ÷ (any x-difference) = 0. A flat shelf of prices or a plateau in a growth chart produces exactly this. It's a real answer, and the sentence to write is "no change over that interval."

How do I find the rate of change when the table has only two rows?

Use those two rows directly. Subtract the y-values, subtract the x-values in the same order, and divide. Two points define exactly one rate, so there's nothing to verify against. If the two-row table represents a linear relationship, that quotient is the slope for every interval, including the ones beyond the table's edge.

Can I use the average rate of change instead of finding a single slope?

Yes, when the table is nonlinear and a single slope doesn't exist. Use the top and bottom rows only, and label the answer "average rate of change over [first x, last x]." It's one number that summarizes the whole table, and the statement is true even when every row-to-row quotient differs.

Why do I get a different answer when I use rows 2 and 3 instead of rows 1 and 2?

You get a different answer only if the table isn't linear — and that difference is the finding, not a mistake. If the quotients match, any pair gives the same slope. If they don't, the table is non-linear and each pair's quotient is that interval's average rate. Report the interval alongside the number.

Which row should be "point one" in the formula?

Any row can be point one, as long as you subtract in the same direction on both sides. Convention is to use the earlier input value for clarity, but (y₁ − y₂) ÷ (x₁ − x₂) gives the identical result. Mixing directions — y₂ − y₁ over x₁ − x₂ — is the only order that produces a wrong answer.