Quizzes in Computer Science Class

Read before you write

Programming teachers have long suspected that many beginners try to write code they cannot yet read. A multinational study led by Raymond Lister tested that suspicion. First-year programming students at universities in several countries answered short multiple-choice questions that asked them to predict what small pieces of code would do, or to pick the line that completed them. Many struggled, and the authors argued that students who cannot reliably trace code are poorly placed to write it.

Tracing questions are cheap to write and quick to answer, which makes them ideal quiz material. A few at the start of every lesson build exactly the skill that debugging depends on.

“What does this code print?”

The single most useful question in a programming class shows a few lines of code and asks for the output. Show the code as a picture, with a large font and a light background, so it survives the projector. Five to ten lines is plenty for a timed question.

The wrong answers should be the outputs students actually get when they slip. In this example the loop runs over range(1, 4), which stops before 4, so the total is 6. A student who counts the 4 as well gets 10, and in this class two of the six did exactly that.

Put the steps in order

An ordering question gives students up to eight items to drag into the right order. In computing that covers a lot of ground: the steps of an algorithm such as a binary search, the stages of the software development cycle, or what happens between typing an address into a browser and seeing the page.

Shuffled lines of code are the best version of all. Programming educators call these Parsons problems, and they have some evidence on their side: Barbara Ericson and colleagues found that students who solved Parsons problems learned about as much as students who fixed or wrote the same code, in much less time. Students have to understand what each line does and which lines depend on others, without the blank page or the syntax errors.

Each item in its right place earns part of the points, so a nearly right program still scores. Give these questions a full minute: reading six lines of code and dragging them is slow work.

Numbers: binary, hex and more

Conversions have one exact answer, which makes them a natural fit for a question where students type a number. They also have a classic wrong answer: reading the binary digits in the wrong direction.

For a quicker round, or a younger class, let students pick the number from tiles instead of typing it, and make the wrong numbers the results of the usual slips.

True or false for computing myths

Computing is full of beliefs that sound right. A true-or-false question brings each one into the open, and a one-line explanation at the reveal deals with it.

• “x = x + 1 is impossible, because no number equals itself plus one.” (False: = assigns a new value; it is not a statement of equality.)
• “A variable keeps the same value for the whole program.” (False.)
• “The computer can run the lines of a program in any order it likes.” (False.)
• “Deleting a file always wipes its data from the disk straight away.” (False: usually the space is only marked as free.)
• “A bigger number always needs more bits.” (False: 9 and 15 both fit in four bits.)

Terms, matching and blanks

Computing has a lot of vocabulary: variable, loop, function, array, compiler, algorithm. A matching question asks students to connect up to eight terms with definitions or examples, which is a fast way to review the language of a unit between harder questions. A fill-in-the-blank question works for definitions too: “A ___ repeats a block of code while a condition is true.”

For drilling the basics, such as conversions and keywords, play the set in Giga Quizzes’ Hacker mode. Students answer at their own pace, the questions come round again when they reach the end, and by the third time a conversion comes round most of the class has it. The guide to every question type shows each format in the editor.