Statistical Testing: The Sign Test

• Does finding a difference in mean between conditions guarantee significance?

  No, it does not guarantee significance
• What could cause a difference in word count?

  It could have occurred by chance
• What is the purpose of statistical testing?

  To find likelihood of chance occurrence
• What test is used to find differences without correlation?

  The Sign Test
• What design must be used for the Sign Test?

  Repeated measures design
• What type of data is required for the Sign Test?

  Nominal data
• What is the significance level commonly used in studies?

  p ≤ 0.05
• Why might researchers use a stricter significance level?

  To be more confident results aren't due to chance
• What does it mean if results are significant?

  Less than 5% chance of occurring by chance
• What can we infer if results are significant?

  The difference was likely caused by the independent variable
• What are the steps to perform the Sign Test?

  1. Write alternative and null hypotheses
  2. Record data and determine the sign
  3. Calculate the value
  4. Find the critical value
  5. Compare calculated value to critical value
  6. Accept or reject hypotheses based on significance
• What should you do if the calculated value is less than or equal to the critical value?

  Then results are significant
• What should you do if results are significant?

  Accept the alternative hypothesis
• What should you do if results are significant regarding the null hypothesis?

  Reject the null hypothesis