Some of the ways we can collect data are:

• Observations
• Measurements
• Experiments
• Surveys

Types of Bias:

• Response Bias - factors in the sampling method influence the result.
• Non-response Bias - participants are unwilling or unable to participate in the study.
• Sampling Bias - the chosen sample does not accurately represent the population.
  • Household Bias - one part of the population is over represented.

A biased set of survey questions can lead to response and non-response bias.

Surveys should

• be carried out in a fair manner,
• include well written questions that are culturally sensitive, and 
• always receive consent from participants.

The data that is collected from a survey can be influenced by

• the wording of the questions,
• the choices of answer provided, and
• the sample group that is selected to answer the questions.

Effective surveys will avoid:

• Leading Questions: intended to lead the participant towards a certain answer.
  • Should concerned dog owners vaccinate their pets?
    • A better question to ask is: Do you think that dogs should be required to be vaccinated?​​​​​
• Loaded Questions: questions that require people to answer in a particular way or forces them into an answer.
  • Which restaurant makes the best hamburgers?
    • A better question to ask is: What is your favourite restaurant?
• Questions that contain more than one question: 
  • What do you think of the taste and appearance of our waffles?
    • Better questions to ask are: 
      • How do you like the taste of our waffles?
      • Rate their appearance on a scale from \(1\) to \(5\), with \(5\) being exceptional.

Bias can be minimized by paying attention to how we word the questions. Here are some strategies:

1. Avoid providing a limited number of responses: 
   • Poor example: Do you exercise every day? Yes or No
   • Better example: How many days a week do you exercise?  
2. Provide exhaustive, mutually exclusive response options: 
   • Poor example: How many days a week do you exercise?  \(1-3\), \(3-5\), \(5-7\) days per week
   • Better response options:  \(0-1\), \(2-4\), \(5-7\) days per week
3. Ask clear questions: 
   • Poor example: Do you like cake?
   • Better example: Do you like dessert?

We need samples because conducting a census isn't always possible:

• A sample should be representative of the population.
  • You are curious about the favourite sport of students at your school.
    • Biased sample group: asking \(30\) students at a school basketball game
    • Non-biased sample group: randomly choosing \(30\) students in the school to ask