Good vs Bad Survey Questions

Good vs Bad Survey Questions

par HUF04 Nguyễn Ngọc Quỳnh Như,

One Well-written Survey Question:

"On a scale of 1 to 5, how much do you agree with the following statement: Using Google Translate helps me use more diverse vocabulary in...

suite...

One Well-written Survey Question:

"On a scale of 1 to 5, how much do you agree with the following statement: Using Google Translate helps me use more diverse vocabulary in my English essays?"

(1 = Strongly Disagree, 5 = Strongly Agree)

--> Specific: It focuses on one single aspect of writing quality: vocabulary.

--> Measurable: Using a Likert scale (1–5) allows you to turn the students' feelings into data for your Quantitative analysis.

One Poorly Constructed Survey Question:

"Do you think Google Translate is fast and helps you write better essays without making many mistakes?"

--> It asks about three things at once: Is it fast? Does it help you write better? it is fast but makes many mistakes, they won't know how to answer.

-->The wording "without making many mistakes" suggests that Google Translate is already perfect, which might influence the student to say "Yes."

-->"Better essays" is too broad. It's hard to measure what "better" means to different students.

Good vs Bad Survey Questions

par HUF04 Nguyễn Đăng Hải,
This is a very sharp analysis of survey design! You have correctly identified the "Double-Barreled" and "Leading" flaws in the poor example. By focusing on "Lexical ...

suite...

This is a very sharp analysis of survey design! You have correctly identified the "Double-Barreled" and "Leading" flaws in the poor example. By focusing on "Lexical Diversity" in the first question, you are aligning your survey perfectly with your research aim of measuring "writing quality."

Here are a specific suggestion to further improve these for your final thesis:

Refine the Scale Labels
For your well-written question, ensure that your Likert scale is balanced.

Suggestion: Instead of just "1 to 5," always clearly define the midpoint.

Example: (1) Strongly Disagree, (2) Disagree, (3) Neutral, (4) Agree, (5) Strongly Agree. This prevents "forced choice" and gives you more honest data for your quantitative analysis.