Initial Planning: Hypotheses and Investigation Constraints
Learn how a statistical investigation begins: turning a question into a testable hypothesis, and planning around the real constraints of time, cost, ethics, confidentiality and convenience.
Revise this, the fun way
Play it interactively, earn XP and build a streak, free.
Start revising freeWhat you'll cover
Planning a statistical investigation
Before anyone collects a single number, a good statistician plans. They turn a question into a **testable hypothesis**, decide what data to collect, and think hard about what could get in the way. This module is about that first stage: writing a hypothesis you can actually test, and spotting the **constraints** that shape every real investigation.
What is a hypothesis?
A hypothesis is a clear statement you can test with data. Learn these planning terms.
Is it testable?
Which of these is a testable statistical hypothesis?
- Students who revise more tend to score higher in tests
- Maths is the best subject
- Everyone should revise more
- Revision is interesting
Strong or weak?
A strong hypothesis is specific and testable. A weak one is vague or just an opinion.
Match the constraint
- Time
- Cost
- Ethics
- Confidentiality
- You only have two weeks to finish
- Printing hundreds of surveys is too expensive
- You must not pressure anyone into taking part
- You keep every reply private and unnamed
Constraints to plan around
When you design an investigation, five practical constraints come up again and again: time, cost, ethical issues, confidentiality, and convenience. A good plan names them early and works around them.
Spot the constraints
Which of these are real constraints on designing an investigation? Choose all that apply.
- A tight deadline
- A limited budget
- Keeping answers private
- The hypothesis is interesting
- You enjoy the topic
Complete the plan
A good plan starts with a testable _____. You decide what _____ to collect and how to collect it. You choose a _____ from the population, and you plan around _____ such as time and cost. Finally you think about how to _____ the results.
Handling a problem
In a survey, many people do not reply. This is called non-response. What is a good way to reduce it?
- Send a polite reminder and keep the survey short
- Only count the people who did reply and forget the rest
- Make the survey much longer
- Change the hypothesis after seeing who replied
Find the constraints
An interactive activity.
Match the fix
- The population is hard to identify
- Many people do not respond
- An unexpected result appears
- People worry about privacy
- Define the population clearly and use a sample frame
- Send a reminder and keep the survey short
- Run a small pilot study first to spot surprises
- Promise that answers stay private and unnamed
Order the planning
An interactive activity.
Plan the investigation
An interactive activity.
Your turn
An interactive activity.