Step 2 · Research question
Your RQ drives the method, the structure and the argument. Most weak essays start with a weak question.
By the end of this step
- A focused research question, with the variables named.
A recipe for a CS research question
Stem · what you Change · what you Measure · which Techniques · under which Conditions.
RQ: To what extent does the load factor (0.1 to 0.9) of a hash table affect the mean lookup time and the mean number of probes of separate chaining compared with linear probing, for 100,000 integer keys in Python?
Spot the problem
What's wrong with this RQ? Self-check drill
Diagnose each question. Some are fine.
“Is quicksort faster than bubble sort?”
Make it better
How effective is encryption?
To what extent does AES-256 keep its encryption speed as input volume grows from 1 MB to 1 GB?
Too broad → narrowed to one algorithm, one measure and defined conditions.
Which sorting algorithm is the best?
How does the proportion of pre-sorted elements (0–100%) affect the number of comparisons made by insertion sort compared with merge sort on arrays of 50,000 integers?
“Best” can't be measured → name the metric and the condition you vary.
Pick the best upgrade Quick check
Which rewrite fixes the problem?
Check it against these
- One question, written as a question.
- Leads to analysis, not description: “to what extent”, “how significant”, “how does … affect”.
- The answer is not obvious before you start.
- Every variable is defined and measurable, with a range.
- Narrow enough to answer well in 4,000 words.
- Clearly computer science: answered with algorithmic complexity, performance metrics or computational theory.