Selection of Inferential Statistics

• Levels of measurement= the type of data we correct in the study (the dependent variable).
• Nominal Data:

  • Is in categories- means it can only be measured by frequency
  • Cannot be ordered (it's either in the category or not- completed a marathon or didn't).
  • Use mode to calculate the measure of central tendency.
• Ordinal Data:

  • Is ranked, you can count and order ordinal data, eg group of runners: 1st, 2nd, 3rd
  • Use median
• Interval Data:

  • Has equal units of intervals throughout the scale
  • Usually on a standardised scale, eg time, temperature, weight, length etc
  • Example: the time in seconds racers finish the run
  • Use mean
• Choosing the correct test:

  1. Is the study looking at a difference or a relationship? ie an experiment or a correlation.
  2. If it's a difference, then what is the experimental design? ie repeated measures or independent groups.
  3. What is the level of measurement produced? ie nominal, ordinal, interval.
• Calculated value= the number you have calculated from your data set after using a statistical test.
• Critical value= the number we compare our calculated value against to see if our results were significant (located in a table of critical values).
• • Level of significance (eg P<0.05)
  • One-tailed or two-tailed (hypothesis directional or non-directional)
  • Number of participants (N), maybe degrees of freedom too (N-1)
  • Our calculated value needs to be greater than or less than the critical value.
• The Sign Test:

  1. One or two-tailed hypothesis & how many PPs were used.
  2. Work out of the change between each condition is positive or negative.
  3. Count up all the + & - (= values get left out which alters number of PPs).
  4. Take the total of the less frequent sign as your calculated s value.
  5. Compare calculated s value to the critical value & conclude (calculated value greater than/ less than critical value).