Questions: Gaussian Elimination and Row Reduction

Question 1

While row-reducing a matrix, a student multiplies the second row by −3. Their partner warns them this might have changed the solution set. Who is correct?

• A. The partner — multiplying a row by a constant changes the equation it represents and could alter the solutions
• B. The student — multiplying a row by any nonzero constant is an elementary row operation that preserves the solution set exactly
• C. Both — the solutions are preserved only if you also multiply a corresponding column by −3
• D. Neither — you can only safely multiply a row by 1 or −1 without affecting solutions

Answer: (B)

Multiplying a row by a nonzero scalar is one of the three elementary row operations, and all three preserve the solution set. Multiplying the equation 2x + y = 4 by −3 gives −6x − 3y = −12, which has exactly the same solutions. The solution set is unchanged because the operation produces an equivalent equation. This is the foundational guarantee that makes Gaussian elimination trustworthy: you can transform the matrix as aggressively as you want using these operations without fear of introducing or losing solutions.

Question 2

After row-reducing a system, a student finds the row [0 0 0 | 5] in the augmented matrix. What does this row indicate about the system?

• A. The system has infinitely many solutions — this row contributes no constraint
• B. There is an arithmetic error; such a row cannot appear in a consistent system
• C. The system has no solution — this row encodes the contradiction 0 = 5
• D. The system has exactly one solution, which can be found by ignoring this row

Answer: (C)

The row [0 0 0 | 5] translates back to the equation 0·x₁ + 0·x₂ + 0·x₃ = 5, which simplifies to 0 = 5 — a contradiction. No values of the variables can satisfy this, so the system is inconsistent and has no solution. This is the standard indicator of an inconsistent system in row reduction. Contrast this with [0 0 0 | 0], which translates to 0 = 0 — a tautology that contributes no information and typically signals infinitely many solutions (a free variable).

Question 3

The reason Gaussian elimination is mathematically valid — that you can freely transform the augmented matrix without worrying about changing the solutions — is that each elementary row operation is reversible and preserves the solution set.

• T. True
• F. False

Answer: True

This is the core guarantee that underlies the entire algorithm. Each elementary row operation (row swap, scalar multiplication, row addition) is invertible: the reverse operation is also an elementary row operation. Because each step can be undone and does not alter which (x₁, x₂, ...) values satisfy the system, any solution to the original system is a solution to the transformed system and vice versa. This means no matter how many row operations you apply, the solution set is identical to that of the original system — which is why you can read the answer off the final echelon form.

Question 4

Gaussian elimination is fundamentally a different method from the substitution and elimination techniques taught in algebra — it is a matrix-based approach, not an equation-based one.

• T. True
• F. False

Answer: False

Gaussian elimination is the same process as algebraic substitution and elimination, systematically organized through the augmented matrix notation. Adding a multiple of one row to another is exactly 'multiply this equation by a constant and add it to another equation.' The augmented matrix just strips away the variable names (which carry no information beyond their column positions) and makes the coefficient structure visible. The algorithm does not change; only the notation does. This is why understanding elimination by hand is the conceptual prerequisite — the matrix is a cleaner bookkeeping device for operations you already know.

Question 5

What does the augmented matrix notation contribute to Gaussian elimination that writing out full equations does not? Why is separating coefficients from variable names useful?

Model answer: The augmented matrix removes everything that doesn't matter — the variable names — and displays only the information that does: the coefficients and constants. This makes the structure of the system visible as a pattern of numbers in a grid, and it makes the row operations mechanical and easy to verify. You can see immediately where zeros are, where pivots need to be created, and what the staircase shape looks like. Variable names would just clutter this view without adding information.

When you write out equations, you must track which variable each coefficient belongs to in every operation, which invites errors and obscures the pattern. The augmented matrix makes explicit that a linear system is really just a structured array of numbers — the variables are implicit in the column positions. This separation also enables the theory: you can talk about properties of the matrix (rank, pivot positions, row space) without reference to any specific variable names, which is what makes linear algebra a general framework rather than a collection of techniques for specific systems.