A table shows a proportional relationship when one constant multiplier turns every input value into its matching output value. Divide each output by its input and you get the same number every time, with one exception that matters: the pair (0, 0) is allowed and cannot be used as a test pair. If the ratios differ, or the relationship only works for some rows, it is not proportional.
The Division Test That Decides Every Row
Take the second column value and divide it by the first column value, row by row. Proportional means every quotient matches. For a table pairing 2 with 6, 5 with 15, and 9 with 27, the division gives 6 Γ· 2 = 3, 15 Γ· 5 = 3, and 27 Γ· 9 = 3, so the constant of proportionality is 3.
The constant has a name and a formula. If the table's columns are x and y, proportionality means y = kx, where k equals y divided by x for any pair with x not zero. The Common Core State Standards list this as 7.RP.A.2b: students must identify the constant of proportionality in tables, graphs, equations, diagrams, and verbal descriptions of proportional relationships.
Column order decides which ratio you build. Divide the dependent quantity by the independent quantity, meaning the value that reacts divided by the value that drives it. If a table lists distance in the first column and time in the second, the constant describes distance per unit of time only if time drives distance, which is the usual reading.
Keep the units attached while dividing. A table giving 4 liters per 2 bottles and 10 liters per 5 bottles yields 2 liters per bottle both times. Dropping the units is how people end up comparing a rate of 2 with a value of 2 that measures something completely different.
Reading the Headers Before You Divide
Headers tell you which quantity is independent, so start there rather than with the numbers. A column labeled "hours worked" driving a column labeled "pay in dollars" means you divide pay by hours to get dollars per hour. Reverse the division and you describe hours per dollar, which answers a question nobody asked.
Watch for tables where the labeled columns hide a third quantity. A recipe table listing flour in cups against cookies produced implies servings or batches between them. The constant you get is cups per cookie, which is valid but rarely the useful rate. Check whether the useful relationship sits one step away.
Unit labels also flag mismatches that ratios cannot fix. If one column reads minutes and another reads hours, convert before dividing, or your quotients will differ by a factor of 60 and you will wrongly call a proportional table non-proportional. In a table of 30 minutes to 2 hours and 60 minutes to 4 hours, dividing directly gives 0.0667 and 0.0667 only if you keep hours over minutes consistently.
Scanned worksheets and copied tables lose headers more often than rows. Confirm both column labels exist before trusting any quotient. When a header is missing, compute the ratio both ways and inspect the context for which direction produces a sensible unit rate.
When the Pair (0, 0) Is Evidence, Not an Exception
A proportional relationship must pass through the origin, so a table row of (0, 0) supports proportionality without helping you prove it. Zero divided by zero is undefined in ordinary arithmetic, which means that row can never produce a quotient. Use any other row to find the constant.
The origin rule gives you a fast rejection test. A table where time 0 pairs with a starting distance of 5 shows a y-intercept, so the relationship is linear but not proportional. This is the classic taxi-fare table: $3.50 at 0 miles, then a fixed rate per mile. The fare is linear, but the $3.50 base fee breaks proportionality, and dividing fare by miles gives different answers every row.
Tables that skip the zero row still obey the rule. If a table of calories against servings starts at 1 serving and each quotient equals 250, the implied graph crosses the origin. You can state that the table is consistent with proportionality and that the relationship would pass through (0, 0) if extended.
A zero in only one column is a red flag rather than a neutral pair. Input 0 with output 8 means the relationship starts at 8, so it fails. Output 0 with a nonzero input, such as 4 hours to 0 dollars, means the rate is zero, which is proportional only if every row shows zero.
A Worked Check on a Table That Fails
Consider a table pairing 3 with 12, 5 with 20, and 7 with 26. The first two rows give 12 Γ· 3 = 4 and 20 Γ· 5 = 4, and it is tempting to stop. The third row gives 26 Γ· 7, which is about 3.71, so the table fails. One failing row is enough; you do not average the quotients.
This pattern shows up in tables with a one-time add-on. Drop the 26 to 28 and the table becomes 4, 4, 4 and passes. The difference between 26 and 28 is the added fee, not an error, and tables built from real pricing often hide that fee in the last row when a discount or surcharge kicks in.
Averaging the quotients is the most common mistake in this test. Averaging 4, 4, and 3.71 gives about 3.90, a number that appears nowhere in the table and describes no row. The constant of proportionality is a value every pair must share, so a single mismatch settles it.
Check the failing row against the constant before blaming the numbers. If most rows share k = 4 and one output sits 2 above 4 times its input, describe the relationship as proportional plus a fixed surcharge of 2 and note that it holds only for the rows before the surcharge applies. That description is honest and more useful than calling the whole table broken.
Tables That Mix Proportional and Non-Proportional Rows
Real datasets often include a header row, a total row, or a tax row that belongs to a different relationship. A receipt table with item prices proportional to quantity followed by a tax line is proportional in the item rows and not proportional once the tax row joins. Separate the rows before testing.
A totals row always fails unless the table is empty. Summing inputs and outputs doubles the scale without changing the ratio for a proportional relationship, but the total row's own quotient still matches if you sum correctly: total output divided by total input equals k when every row shares k. When it does not match, some row inside the table is already non-proportional.
Spotting the intruder row is faster than testing all rows. Compute the ratio for the rows that share the same units, mark any row with different units or a fee label, and test around it. In a table of delivery costs by package weight, the flat handling row breaks proportionality and the shipping rows alone usually hold.
Document what you excluded. A teacher or grader will accept "rows 2 through 4 are proportional with k = 4; row 5 adds a fixed $3 fee" more readily than a bare yes or no. The distinction between the test failing and the table being untestable as a whole is worth writing down.
Cross-Multiplying to Catch Rounding Errors
Division can mislead when quotients land on long decimals. Cross-multiplying compares the products of opposite corners within a pair and stays exact for whole numbers. For the pair (3, 12) and (5, 20), check 3 Γ 20 = 60 and 5 Γ 12 = 60; equal products mean equal ratios.
Use cross-multiplication when a table's numbers are close but not obviously matched. Pairs (8, 20) and (10, 24.9) are hard to judge by quotient, and 8 Γ 24.9 = 199.2 against 10 Γ 20 = 200 shows the discrepancy of 0.8. Cross products agree only when the ratios are identical, so one unequal product disqualifies the table.
Keep cross-multiplication for verification rather than discovery. It needs two rows, and two matching rows prove nothing about a third. Its strength is confirming whether a discrepancy is real or an artifact of rounding in a decimal quotient, which is precisely the case where division displays like 3.3333 in a calculator.
Set a tolerance when the data are measured rather than counted. If a table was produced from laboratory measurements, quotients equal to 2.01, 1.99, and 2.00 are proportional within measurement error. State the tolerance you used; unnamed tolerance is the difference between a math answer and a defensible one.
Comparing the Test Methods Side by Side
Each method has a job. The comparison below assumes the same table is tested four ways, pairing inputs 4, 6, and 9 with outputs 10, 15, and 22.5.
| Method | Rule used | Result on this table | Best for | Weakness |
|---|---|---|---|---|
| Quotient of each row | y Γ· x matches every row | 2.5, 2.5, 2.5 β proportional | Classroom checks and finding k | Rounds long decimals |
| Cross products of two rows | xβΒ·yβ equals xβΒ·yβ | 4 Γ 15 = 60 and 6 Γ 10 = 60 | Spotting rounding gaps | Two rows prove nothing |
| Graph check at the origin | Line through (0, 0) | Passes origin with slope 2.5 | Visual confirmation | Requires a plot |
| Equation fit | y = kx for one k | y = 2.5x fits all rows | Communicating the answer | Zero rows need a footnote |
The graph check appears in the standards alongside tables. 7.RP.A.2a requires students to decide whether quantities are proportional from graphs that pass through the origin with a straight line. Use it as a second opinion rather than the primary test.
Order the methods for speed. Run the quotient test first because it finds k, use cross products only when a quotient looks off, and finish with the equation so your answer carries a formula. A table answer of "proportional, k = 2.5, equation y = 2.5x" covers every version of the question.
Decimal k values are common and not a defect. A rate of $0.45 per ounce produces quotients like 0.45 and 0.4500001 when the table provider divided and re-typed. Equal to the stated precision is the standard; the source table should declare its precision.
Choosing Units So the Ratio Means Something
The same proportional table supports different constants depending on direction. A table pairing gigabytes with seconds can report gigabytes per second or seconds per gigabyte, both constant, and both proportional. Direction changes the story, not the verdict.
Pick the direction that produces a familiar unit. Pay per hour beats hours per dollar; cost per unit beats units per dollar; miles per gallon beats gallons per mile. For small numbers, gallons per 100 miles is the industry habit, and the constant becomes a rate with a clear use.
Note where k equals 1. A table of centimeters against millimeters is proportional with k = 10 one way and k = 0.1 the other, and both are correct. When the constant is exactly 1, the two quantities are the same measurement in different units, which is worth saying aloud because it explains why every pair matches exactly.
Swap in a unit of a different size and the constant changes by a known factor. A rate of 1.6 kilometers per mile becomes 2.6 knots per horsepower if you switch one column's unit, and the proportional verdict stays. Units are a display choice; proportionality is structural.
Decimal Ratios, Uneven Columns, and Screening Before You Test
Screen for proportionality before doing arithmetic. Ask three questions: are there rows with 0 in exactly one column, do the columns have comparable units, and is there a row that looks like a fee or total? One yes does not always disqualify, but it flags where the test must go.
Decimals, fractions, and mixed representations do not stop a table from being proportional. A table pairing 0.5 with 1.5 and 2.5 with 7.5 has a constant of 3 despite the decimal inputs. Convert fractions like 18.5 and 7.4 to whole numbers when they contain nonterminating decimals, since a constant like 7.4 divides cleanly into 18.5 only as 2.5.
Some tables include a message about missing data, which explains rows that belong to different clusters. If a 2024 report from a statistics agency or an app export has blank cells, that usually means a non-proportional cluster rather than a proportional table. Treat blank cells as separate clusters until proven otherwise.
Check whether the first column increases by equal steps. Equal steps with equal second-column steps usually signal a proportional relationship, but the pattern is a clue rather than a test. A table stepping by 0.25 in the input and 0.75 in the output has k = 3 if the steps always match, and any skipped or doubled step breaks the streak.
Why Independent Work Beat Memorized Tables
Dependence on a memorized answer looks like competence in a worksheet and fails on a spreadsheet. A learner who can decide proportionality from any two rows plus one origin check can audit an unfamiliar export without a teacher. 7.RP.A.2b's demand to identify the constant in tables, graphs, equations, and diagrams is really a demand for that flexible skill.
The zero row teaches why proportional is not a synonym for linear. A linear table with a nonzero y-intercept fails the proportional test, and that distinction lets you predict the graph's starting height from the table alone. Linear tables can slope at any angle; proportional tables all run through a single point.
Reporting the result takes one sentence. "This table is proportional with a constant of 2.5; the equation is y = 2.5x" is the entire answer, and adding the units makes it complete. Any additional commentary is optional.
The skill transfers to the data a person handles in work. Cost tables, conversion charts, and time logs all invite the same division, and catching the fee row before it poisons the rate is worth more than a perfect score on a worksheet. The test is small; the habit is durable.
FAQ
Does a table with matching ratios always mean the quantities are proportional?
Almost, but you must also admit the zero pair. A matching ratio in every nonzero row plus a consistent pass-through of the origin is sufficient. If the table shows the input 0 pairing with a nonzero output, the relationship is not proportional.
Why does my table with a starting fee fail the proportional test?
A proportional relationship has no fixed add-on, so any starting fee at zero input breaks the test. In the taxi example, $3.50 at 0 miles is a y-intercept, and the fare becomes $3.50 plus a rate per mile. That formula is linear, not proportional.
What if my table only has two rows?
Two rows with matching ratios suggest proportionality but cannot confirm it. A third row could differ by a surcharge or a broken step and you would never see it. Report the table as consistent with k from the available data and check a third pair before answering yes.
Do I divide the second column by the first column every time?
Divide the output column by the input column, meaning the dependent quantity by the independent one. The headers decide which is which, and reversing them still yields a constant, just one measuring the inverse rate. Any row with a zero in the independent column is skipped.
How should I round k when the quotients differ in the last digit?
Round to the precision the table declares and call the table proportional within that precision. A table showing 0.45 and 0.4500001 quotients is consistent to three decimals. If no precision is stated, report the discrepancy and treat the table as unverified rather than choosing a comfortable number.
Can a table show proportionality without 0 appearing anywhere?
Yes. The origin row is allowed to be absent, and the table is still consistent with proportionality if every present row shares a constant. State that the relationship would pass through (0, 0) if extended, which is what a table-based claim of proportionality implies.
Is a table with a decreasing second column proportional?
Constant decrease is fine. A table pairing 2 with 10 and 5 with 4 yields 5 and 0.8, which differ, so that table is not proportional. A table of 2 with 10 and 4 with 5 yields 5 and 1.25, also not proportional. Only an identical quotient row after row qualifies.