Stats Testing

• Significant

  A finding we are confident is not due to chance
• Statistical testing

  A mathematical method for determine whether hypotheses should be rejected - the outcome tells us whether a difference or relationship is significant
• Probability


  The chance of something happening
• Critical value


  The number that acts as a threshold for passing a statistical test
• Significance Level

  The accepted level of risk that our findings are actually due to chance

  • Normally 5% (0.05) in psychology, but could also be 1% (0.01)
• Steps of a Sign Test

  1. Decide whether the test is one-tailed (Directional hypothesis) or two-tailed (Non-directional hypothesis)
  2. For each participant record a ‘+’ if there’s an improvement, or a ‘-‘ if there’s a reduction
  3. Count the number of +’s or -‘s - the smaller value is the calculated S value
  4. Find the critical value of S (Usually in a table) - N is the number of participants that had any difference
  5. If S<= critical value then the result is significant
• Type 1 error

  A false positive - we say our findings are genuine when they are due to chance
• Type 2 Error


  False negative - we say our findings are due to chance when they were actually genuine
• Nominal Measurement


  Data thats allocated into categories by counting frequency of occurrence (Tall, Medium, Short)
• Ordinal Measurement


  Data is placed into rank order (Tallest to shortest, 1=tallest, 5=shortest)
• Interval Measurement


  Measured in fixed units with equal distances between all points (5’3, 5’9)