Research Basics
A company downloads a government census report to learn about household sizes in a city. Is this primary or secondary research?
Secondary — the data already existed and was collected by someone else for a different purpose
A survey question with fixed answer choices, like a 1-5 rating scale, is called ______-ended.
Closed-ended
Which research method uses a moderator to guide a small group discussion?
Focus group
When two things trend together, what is that called?
Correlation
When the people you ask don't represent the group you're trying to learn about, what is that called?
Sample bias
True or False: Quantitative research is the best tool for understanding the deep "why" behind how a customer feels
False — qualitative research explores the "why"; quantitative measures how many and how much
What's wrong with this question: "How often do you enjoy our delicious new fries?"
It's loaded — it assumes the person enjoys the fries and that they're delicious
Which method reveals what people actually do, rather than what they say they do?
Observation/Data analytics
In a survey of 40 students, 30 pick Option A. What percentage picked Option A?
75%
True or False: Surveying 500 people at one sneaker store fixes the sample-bias problem when asking whether "the town" likes sneakers, because 500 is a big sample.
False — a big biased sample just gives a bigger biased number
Name one downside of primary research compared to secondary research.
It costs more time and money to collect
How satisfied are you with the price and the taste of this product?" Which item on the strong-question checklist does this break?
One idea per question — it combines two questions (price and taste) into one, so you can't tell which one the answer is about
Observation studies and data analytics share one big weakness. What is it?
They show what is happening but don't explain WHY
Before claiming that one thing caused another, what should a good researcher do?
Look for other possible explanations (a third factor, the season, coincidence) before claiming causation
Name one thing participants should be told as part of informed consent.
What the research is for, or how their information will be used
A café owner wants a quick, cheap first look at how much people in town typically spend on coffee each year, before spending anything on her own study. Which kind of research should she start with, and why?
Secondary research — it's fast and inexpensive because the data already exists
Rewrite this leading question so it's neutral: "Don't you think the cafeteria should add more vegetarian options?"
"How interested would you be in more vegetarian options in the cafeteria? (1 = not at all, 5 = very)." Must remove the "Don't you think" push and not assume an opinion.
A brand wants honest, in-depth opinions about a sensitive topic and worries people will clam up in front of others. Which method fits best?
One-on-one interviews — deep, honest insight without group pressure
A business's data shows weeks with more social posts also had higher sales. Reword this as a careful correlation statement instead of a causation claim.
"Weeks with more posts were also weeks with higher sales" or "More posts were associated with higher sales" — NOT "posting caused higher sales"
A researcher who wants to prove her product is popular only counts the positive comments and skips the negative ones. Name the bias and give one fix.
Confirmation bias. Fix: deliberately look for evidence that challenges her assumption, include all comments, and be transparent about limitations.
(Scenario): A drink brand wants to know BOTH how many teens would buy a new flavor AND why some of them hesitate. Name two research approaches that together would answer both parts, and say which part each one answers.
A closed-ended survey (primary/quantitative) answers "how many"; interviews or a focus group (primary/qualitative) answer "why they hesitate."
(Scenario): A survey starts with three questions about a frustrating checkout experience, then asks "How would you rate this store overall?" What effect is this likely to have on the overall rating, and how would you fix the survey?
Question order effect — the earlier negative questions prime people to rate the store lower. Fix: ask the overall rating first (or reorder neutrally and group questions logically).
(Scenario): A store owner thinks shoppers walk right past a new display near the entrance. Name the best method, say exactly what you'd watch for, and name one thing it still couldn't tell her.
Observation. Watch for how many people walk past vs. slow down or stop, and at what times of day. It couldn't tell her WHY people ignore the display (that would need a follow-up interview or short survey).
(Scenario): A city notices that bigger fires have more firefighters at the scene AND cause more damage. A council member says, "Sending fewer firefighters would reduce the damage." What research error is this, and what's the real explanation for the pattern?
Confusing correlation with causation. The hidden third factor is fire size — bigger fires draw more firefighters and cause more damage.
(Scenario): A student surveys only his own soccer team about whether the school should fund a new field, promises answers are anonymous, then posts everyone's names and answers on a group chat. Name TWO problems and a fix for each.
(1) Sample bias — soccer players aren't representative of the whole school; fix by surveying a mix of students across grades/activities.
(2) Consent/privacy violation — he broke a confidentiality promise; fix by keeping responses anonymous, as promised, and sharing only combined results.