Table Analysis: SORT First, Think Second

By EMPOWERgmat on August 6, 2024

Picture a smart person sitting in front of a Table Analysis question. Eight rows, five columns, three statements. Their finger is on the screen, sliding down a column of numbers, lips moving, trying to find the middle value. Forty seconds later they lose their place and start over. The whole time, there is a button at the top of that column that would have lined every number up from smallest to largest in one click.

I watch some version of this every week. It is not a math problem. It is a process problem. And it is completely fixable.

What Table Analysis actually is

Table Analysis is one of the five question types in the Data Insights section of the GMAT Focus Edition. You get a table of data with a short note above it. Each column can be sorted. Below the table sit three statements, and for each one you pick between two choices, like Yes or No, or True or False.

That's it. No answer choices A through E. No hidden trick in the format. Just a table, a sort button and three calls to make.

The Data Insights section gives you 20 questions in 45 minutes, per GMAC's page on the exam format. That works out to about 2 minutes and 15 seconds per question on AVERAGE. A Table Analysis question done the right way should come in well under that, which buys you time for the wordier questions in the same section.

The order that wins: statement, column, sort, call

Here's the wager. I'll bet that when you open a Table Analysis question, you start by reading the table. Every row. Every column. You try to 'understand the data' before you know what you're being asked.

That is backwards. The table is not a story. It is a reference. You don't read a phone book cover to cover before you look up one name.

Instead, do this every time:

Skim the headers and the note. Ten seconds. What is in each column? What are the units? Is revenue in dollars or thousands of dollars?
Read ONE statement. Not all three. One.
Name the column it needs. Write it on your notepad. "Bikes sold, median."
Sort that column. Click it.
Make the call. Write Y or N on your pad. Move on.

Stop reading for a second and think about step 3. Why write the column name down? Because the moment you sort, the rows jump around, and if you're holding the question in your head, you will lose it. Notes keep you from re-reading. Re-reading is where the time goes.

A jumbled row of blocks of different heights beside the same blocks lined up from shortest to tallest, with the two middle blocks glowing. — Sorted, the middle finds you. Unsorted, you hunt for it.

Try one

Try it

A regional bike shop chain tracked its eight stores for one quarter. The table shows each store's region, the number of bikes it sold, its revenue in thousands of dollars, and the number of bikes customers returned. On the test, you could sort by any column.

Store	Region	Bikes sold	Revenue ($000)	Returns
Ashford	North	412	286	19
Belmont	South	275	198	22
Carver	East	530	341	15
Dalton	West	198	151	12
Easton	North	366	240	30
Fulton	South	489	312	18
Granby	East	305	221	9
Hollis	West	451	297	27

For each statement, select Yes if the statement is true based on the information in the table. Otherwise, select No.

The median number of bikes sold per store is greater than 400. (Yes / No)
The store with the fewest returns sold fewer bikes than at least half of the eight stores. (Yes / No)
The store with the highest revenue also had the highest revenue per bike sold. (Yes / No)

Show the answer

Statement 1: No.

Column needed: Bikes sold. Sort it. The order becomes 198, 275, 305, 366, 412, 451, 489, 530.

The median is the middle value of a list put in order. With eight stores there is no single middle, so the median is the average of the 4th and 5th values: (366 + 412) / 2 = 778 / 2 = 389. Is 389 greater than 400? No.

The trap: grabbing 412 because it 'looks like the middle,' or looking at the unsorted column and picking the 4th or 5th row as listed. Ashford, the 1st row, is 412 and Easton, the 5th row, is 366, so the unsorted order points you nowhere good. Sort first and the answer is sitting right there.

Statement 2: Yes.

Column needed: Returns. Sort it. Fewest returns is Granby with 9. Note on your pad: "Granby, 305 bikes."

Now sort Bikes sold. Stores with more than 305 bikes: Easton 366, Ashford 412, Hollis 451, Fulton 489, Carver 530. That's 5 stores. Half of 8 is 4. Granby sold fewer bikes than 5 stores, and 5 is at least 4. Yes.

The trap: answering about the wrong store. People sort Returns, see Easton at the top of a descending sort, and run with the MOST returns instead of the fewest. Write the store name down before you sort the next column.

Statement 3: No.

Column needed: Revenue, then a quick ratio. Sort Revenue. Highest is Carver at 341 (thousand dollars) on 530 bikes.

Here's the move that saves you. To prove this statement false, you don't need to calculate eight ratios. You need ONE store that beats Carver. Estimate Carver: two thirds of 530 is about 353, and 341 is less than that, so Carver earns less than 2/3 of a thousand dollars per bike.

Now look for a store with high revenue compared to its bikes. Dalton: 151 on 198 bikes. Three quarters of 198 is 148.5, and 151 is more than that, so Dalton earns more than 3/4 per bike. Since 3/4 is bigger than 2/3, Dalton beats Carver. The statement is false. No.

The trap: assuming the biggest store must be the most efficient. Big totals and big ratios are different things. The test LOVES putting them in the same statement.

Final calls: No, Yes, No.

The mistakes I see every week

1. Reading the whole table before reading any statement. You spend a minute 'getting familiar' and then have to look everything up again anyway.

INSTEAD: Skim the headers and units for about ten seconds, then go straight to statement 1. Let the statement tell you which column matters.

2. Scanning a column with your eyes instead of sorting it. Finding the max, the min or the median by eye in an unsorted list is slow and error prone. One skipped row and your answer is wrong.

INSTEAD: If a statement mentions highest, lowest, most, fewest, median, rank or 'at least half,' sort. Every time. The click costs you one second.

3. Calculating every ratio when one counterexample will do. A statement that says 'the store with X also has Y' is false the moment you find a single store that breaks it.

INSTEAD: Ask yourself what would make the statement false, then hunt for that one row. Use easy benchmarks like 1/2, 2/3 and 3/4 instead of long division.

4. Doing it all in your head. This is the one that hurts the most, because the people who do it think they're being fast. They sort a column, the rows shuffle, they forget which store they were tracking, and they sort again. And again.

INSTEAD: Write the store name and the number you need on your pad before you click anything else. Three short notes per question. That's not slow. Re-sorting three times is slow.

A hand writing a short note on a notepad while a table on a screen beside it shuffles its rows. — The table moves. Your notes don't.

Why this is a thinking test, not a math test

Look back at the example. The hardest math in it was adding two numbers and halving them. Everything else was DECIDING: which column, which direction, what would make this false. That's the same skill you need for Multi-Source Reasoning, where you hunt for the one tab that answers the question instead of reading them all, and for any chart, where you read the axes before anything else.

The people who struggle with Table Analysis are rarely weak at math. They're doing the right math the slow way. And slow on one question means rushed on the next, and rushed is where the silly mistakes come from. You can review and change up to 3 answers in a section, but that's a safety net, not a plan. Get it right the first time.

So here's what you're going to do today. Pull up five Table Analysis questions from GMAC's official practice materials. Before you answer any statement, write down the column it needs and whether you'll sort high to low or low to high. Then sort, call it and write Y or N. Time each question. When you're done, look at your pad. If any question has no notes on it, I'll bet that's the one that took the longest. Redo it the right way tomorrow.