Designing investigations Revision Notes for Edexcel GCSE Statistics

Designing investigations

When you're planning to test a hypothesis through data collection, there are several important factors you need to think about carefully. A well-designed investigation will give you reliable, meaningful results that you can trust.

Key considerations for investigation design

Practical considerations

Time and cost factors play a crucial role in your investigation design. You need to think about how long your investigation will take and whether you have the resources to complete it properly. This includes considering how much time you'll need for data collection, analysis, and any equipment or materials required.

Convenience of data collection is another practical factor. It's much easier to collect data locally rather than from distant locations. Consider whether your target population is accessible and whether you can realistically gather the data you need.

infoNote

Planning ahead for practical constraints can save you significant time and resources later. Always consider whether your investigation design is realistic given your available time and budget.

Ethical and legal considerations

Ethical issues must always be at the forefront of your investigation design. You have a responsibility to respect people's dignity and rights throughout your research. This means being honest about your purposes, not deceiving participants, and ensuring no harm comes to anyone involved in your study.

Confidentiality is essential when handling personal data. You must keep all information secure and confidential, ensuring that individual responses cannot be traced back to specific people. This builds trust with your participants and meets legal requirements for data protection.

chatImportant

Ethical considerations are not optional - they are fundamental requirements for any investigation involving human participants. Failing to address ethical issues properly can invalidate your entire study.

Methodological considerations

Identifying your population and sampling method requires careful thought. You need to be clear about who you're studying and how you'll select your sample to ensure it represents your target population accurately.

Planning for non-response is crucial because not everyone you contact will participate in your study. You need to account for this by planning to collect more responses than your minimum requirement.

Conducting a pilot survey helps you test your methods on a smaller scale first. This allows you to identify potential problems and estimate response rates before launching your full investigation.

Handling anomalous results should be planned in advance. You need to decide how you'll identify unusual responses and what you'll do with them - whether to investigate further or exclude them from your analysis.

Planning for non-response

Non-response occurs when people don't return completed questionnaires or refuse to participate in your study. This is a common problem that you need to plan for from the beginning.

Using pilot surveys to estimate response rates

A pilot survey involves sending your questionnaire to a small group first to test how well it works. This smaller trial helps you understand what proportion of people are likely to respond to your full survey.

From your pilot survey, you can calculate the response rate - the percentage of people who actually complete and return your questionnaire. This proportion becomes crucial for planning your main investigation.

Calculating the number of surveys needed

Once you know your response rate from the pilot survey, you can calculate how many surveys to send out to get your desired number of completed responses.

The basic formula works like this:

Response rate from pilot = $\frac{\text{Number who responded}}{\text{Number sent out}}$
For full survey: Response rate = $\frac{\text{Target responses}}{\text{Number to send out}}$
Rearranging: Number to send out = $\frac{\text{Target responses}}{\text{Response rate}}$

lightbulbExample

Worked Example: Calculating Survey Numbers

The situation: Zeedan wants to investigate whether people in the UK prefer tea or coffee. He conducts a pilot survey sending questionnaires to 270 people and receives 180 completed responses back.

Step 1: Calculate the response rate from the pilot survey Response rate = $\frac{180}{270} = \frac{2}{3}$ (or approximately 0.67)

Step 2: Determine how many surveys needed for target responses Zeedan wants at least 400 completed surveys for his main investigation.

Using the proportion: $\frac{180}{270} = \frac{400}{x}$ where $x$ is the number of surveys to send

Step 3: Solve for x Cross multiply: $180 × x = 400 × 270$ $180x = 108,000$ $x = \frac{108,000}{180} = 600$

Answer: Zeedan should send his survey to at least 600 people to get approximately 400 completed responses.

Checking the answer: If he sends to 600 people with a $\frac{2}{3}$ response rate, he'll get $600 × \frac{2}{3} = 400$ responses, which matches his target.

Alternative calculation method

You can also work with the response rate as a fraction:

lightbulbExample

Alternative Method Example

For a survey where you need 300 responses and your pilot survey had a response rate of $\frac{4}{5}$ :

Set up the equation: $\frac{4}{5} = \frac{300}{x}$

Cross multiply: $4x = 5 × 300 = 1,500$ Therefore: $x = \frac{1,500}{4} = 375$

You need to send out 375 surveys to get 300 completed responses.

Important considerations for sensitive topics

When your investigation involves sensitive or personal topics, you need extra care in your design. People may be reluctant to respond honestly or at all to questions about private matters.

chatImportant

Interview vs questionnaire decision: For sensitive topics, you need to weigh up the advantages and disadvantages of different data collection methods. Interviews might provide richer data but could make people uncomfortable, while anonymous questionnaires might get more honest responses but less detailed information.

Using secondary data: Sometimes you can find relevant data that's already been collected, which might be useful for sensitive topics where people are reluctant to share personal information directly.

bookmarkSummary

Key Points to Remember:

Always plan for non-response by calculating how many extra surveys you need to send out
Use pilot surveys to estimate realistic response rates before your main investigation
Consider ethical issues, confidentiality, and practical constraints when designing your investigation
The response rate formula is: $\frac{\text{Number responding}}{\text{Number sent out}}$
To find surveys needed: $\frac{\text{Target responses}}{\text{Response rate}} = \text{Surveys to send}$
Sensitive topics require extra careful consideration of data collection methods

Designing investigations (Edexcel GCSE Statistics): Revision Notes