Two students apply different sequences of row operations to the same matrix and arrive at two different-looking row echelon forms. Student A says both are correct; Student B says only one can be correct. Who is right?
AStudent B — there is only one valid REF for any matrix, just as there is only one RREF
BStudent A — many different REFs are possible for the same matrix, since REF is not unique
CBoth students are wrong — different row operation sequences always produce identical-looking results
DStudent B — row operations must be applied in a fixed canonical order to produce valid REF
REF is NOT unique — different valid sequences of row operations can produce different-looking REFs from the same matrix. Student A is correct. This is precisely why RREF is valuable: RREF IS unique. Every matrix has exactly one RREF regardless of the row operation sequence used. This uniqueness makes RREF a canonical form that unambiguously represents the solution structure.
Question 2 Multiple Choice
A matrix with 4 columns is brought to RREF and has exactly 3 pivot columns. How many free variables does the corresponding linear system have?
A0 — every variable is determined by a pivot
B1 — there is one non-pivot column
C3 — one free variable per pivot
D4 — the number of free variables equals the total number of columns
Free variables correspond to non-pivot columns. With 4 columns and 3 pivot columns, there is 1 non-pivot column, giving 1 free variable. The rank is 3, the nullity is 1, and rank + nullity = 4 = number of columns — the rank-nullity relationship made directly visible in RREF.
Question 3 True / False
Every matrix has a unique reduced row echelon form, regardless of which sequence of row operations was used to compute it.
TTrue
FFalse
Answer: True
Uniqueness of RREF is a theorem, not just a convention. No matter what valid row operations you perform, you will always arrive at the same RREF. This makes RREF a canonical form: two matrices have the same RREF if and only if they represent equivalent linear systems with the same solution set. This uniqueness is what distinguishes RREF from REF, which is NOT unique.
Question 4 True / False
Row echelon form (REF) is also unique — any two valid REFs of the same matrix should look identical.
TTrue
FFalse
Answer: False
REF is NOT unique. You can scale a pivot row by any nonzero constant and still have a valid REF; different elimination paths leave different patterns of nonzero entries above the staircase. Multiple valid REFs can represent the same system. RREF, by contrast, is unique — the additional requirements (pivots equal to 1, all entries above pivots equal to 0) are precisely what force uniqueness.
Question 5 Short Answer
What does RREF directly reveal about a linear system's solution structure that REF does not, and why is this useful?
Think about your answer, then reveal below.
Model answer: RREF makes the solution structure completely transparent without back-substitution. Pivot columns correspond to basic variables (determined uniquely once free variables are assigned); non-pivot columns correspond to free variables (which take any value). The number of pivots is the rank; the number of free variables is the nullity; rank + nullity = n is directly visible. From REF, you can obtain the same information but must work backward through back-substitution. From RREF, you simply assign parameters to free variables and read off basic variables directly from the pivot rows.
The key insight is that RREF is a canonical form that completely exposes solution structure, while REF is merely a form that makes back-substitution tractable. Because RREF is unique, it also determines whether two matrices are row-equivalent — they share the same RREF if and only if they represent systems with identical solution sets.