intro to statistical testing

• To select an appropriate statistical test you need to know:

  1. The level of measurement/type of data that has been used
  2. The type of test – difference or correlation?
  3. The design used – independent groups/ repeated measures/matched participants?
• nominal data is often referred to as categorical data as it is data which can be separated into discrete categories which do not overlap
• Nominal data is collected by counting the frequency of each category.

  An example of nominal data, in terms of biology, would be hair colour (brown, black, red, etc.)
• ordinal data is measured on a scale with distinct levels but no fixed intervals between them
• ordinal data is data that is presented in rank order (e.g. places in a beauty contest, or ratings for attractiveness). it is also a form of categorical data
• Interval level data is data measured in fixed units with equal distance between points on the scale and can go into minus figures. Uses standardised scales . For example, temperature measured in centigrade.
• interval data is continuous; the value recorded could be any point on the scale used, not limited to a small set of discrete categories
• examples of interval data are time in seconds, weight in grams and temperature in celsius
• Ratio data is similar to interval but you cannot have minus figures. There is an absolute zero point.
• the four levels of measurement are nominal, ordinal, interval and ratio
• there are two types of test- difference or correlation/association
• Is the research hoping to find that women will be more helpful than men or that there will be a difference in what men and women look for in a relationship? If so then you need a TEST OF DIFFERENCE.
• Is the study hoping to find that there is a relationship between level of attractiveness and partner choice?  Do people choose partners of a similar level of attractiveness to themselves? If so you need a TEST OF CORRELATION.
• the design refers to the organisation of the participants in the research- independent groups/ repeated measures/matched pairs?
• experimental design

  Related data means the same group of people undergoing both conditions (repeated measures design/matched pairs). Unrelated means that the research is looking at two different groups of people i.e. control and experimental group, early or late risers etc. (independent groups)
• carrots should come mashed with swede under roast potatoes
• chi squared

  sign test

  correlation(above)

  mann whitney

  wilcoxon

  spearman's rho

  unrelated t-test

  related t-test

  pearson's