choosing a statistical test

Question 1

statistical testing

Answer 1

• used to determine whether a difference or correlation found is significant
• outcome of this has implications for whether we accept or reject the null hypothesis

three factors used to decide which test to use -
- whether researcher is looking for difference or correlation
- in the case of a difference, which experimental design is being used
- level of measurement

Question 2

decision 1 - difference or correlation

Answer 2

• should be obvious from wording of hypothesis whether researcher is looking for a difference or a correlation

Question 3

decision 2 - experimental design

Answer 3

• this is not an issue if looking for a correlation
• repeated measures and matched pairs designs are classed as related designs, as in repeated measures, all participants do all conditions, and in matched pairs, participants are matched on a similar quality, making them related
• participants in conditions for independent groups designs are different, making it unrelated

Question 4

decision 3 - levels of measurement

Answer 4

nominal data -
- data represented in categories
- discrete because one item can only appear in one category

ordinal data -
- data is ordered in some way
- does not have equal intervals between each unit
- lacks precision because it is based on subjective opinion
- sometimes referred to as unsafe data, so it is not used as part of statistical testing
- raw scores are converted to ranks and then these are used in the calculation

interval data -
- based on numerical scales that include units of equal, precisely defined size
- better than ordinal data because more detail is preserved
- most precise and sophisticated form of data and is necessary criterion for parametric tests