choosing a statistical test

• statistical test table

  name the unrelated design tests in the test of difference

  nominal: chi-squared

  ordinal: mann-whitney

  interval: unrelated t-test
• statistical test table

  name the related design tests in the test of difference 


  nominal: sign test

  ordinal: wilcoxon

  interval: related t-test
• statistical test table

  name the tests in the test of correlation

  nominal: chi-squared

  ordinal: spearmans rho

  interval: pearsons R
• difference 


  researcher controls what happens to at least one of the variables
• difference

  researcher looks for a difference in outcomes between the experimental groups
• relationship

  researcher observes what happens naturally without manipulating the variables
• relationship


  you look for a relationship or pattern between things by using correlational designs
• alternative hypothesis H1: It makes a statement that suggests a potential result or an outcome that an investigator or the researcher may expect.
• The null hypothesis H0 states that there is no relationship or difference between the two variables being studied
• if results are significant, null hypothesis must be rejected.

  if results are not significant, null hypothesis must be accepted
• steps to choose the correct statistical test

  difference or correlation?

  if a study of difference identify the experimental design

  level of measurement
• nominal data

  form of categories

  discrete - one item can only appear in one of the categories

  data is frequencies; cannot be arranged into rank order, no scale of measurement (counting and categorising)

  data usually gathered in terms of yes or no answers
• ordinal data

  participants are given a score/rating that ranks highest to lowest/first to last

  data has a natural order

  intervals are not equal

  lacks precision as it is based on subjective opinion e.g how happy are you from 1-10
• interval data

  numerical scales, includes units of equal, precisely defined size e.g height

  gaps between intervals are equal

  objective and shareable, usually collected with devices like stopwatch, thermometer, weighing scales
• a researcher compared the helping behaviour of members of the public. it was counted how many people stopped to help a man dressed in a business suit and how many stopped to help a man dressed as if he was homeless

  diff or relationship?: difference

  operationalised IV and DV: IV: man's clothes DV: how many times people stopped to help

  experimental design: independent groups so it is unrelated

  level of measure: nominal

  statistical test: chi-squared