Inferential testing

• Operationalised

  the process of defining variables which are vague, abstract and subjective into ones which can be measured objectively.
• Inferential testing

  allows researchers to state whether a result is significant, it is probable that any difference/association/correlation found is not because of chance.
• Null hypothesis:

  there will be no significant difference, correlation or association between variables investigated.
• Alternative hypothesis

  -non-directional (two-tailed) is written when no previous research has been found in that area. There will be a difference/association/correlation between X and Y.

  -directional (one-tailed) is written when there is previous research research done.
• Nominal level

  named or categorical data. Data at the nominal level is discrete and can't be ranked. Only measure of central tendency you can use on nominal data is mode.
• Ordinal level

  PPs individual scores can be ranked and ordered. BUT measuring device does not contain fixed, objective, standardised intervals. The difference between one unit and next one isn't standardised. Measure of central tendency used are mode and median. Can't calculate mean and SD, these measures use all the data in the set of data, must assume there are fixed distances on the scale.
• Interval level

  refers to having measuring device. e.g. time, weight, hight and calories. Measure of central tendency can calculate mode, median and mean. You can calculate range and SD.
• Purpose of inferential test

  -conducted after results have been obtained.

  -Find a difference/correlation/association which is significant means difference/correlation/association in not a result of chance.

  -Use difference significant levels to assess how confident they can be P-0.05 and P-0.01 which shows researcher what hypothesis should be rejected and accepted.
• test of difference tests

  chi square and sign test
• choosing right test for study

  -each test has its own criteria for when it's appropriate to use.

  -study is looking for a difference, association and correlation.

  -If looking for a difference, need to use repeated measures or independent groups
• Choosing the right test for the study

  -Both co-variables must be interval level for persons-R. If one or both is ordinal then use spearman's rank.

  -Test of correlation are used to see if any relationship is found between two co-variable's.

  -Numerical score is given, tells researcher the strength and direction of any relationship. If r value is significant above 0.8 then it isn't because of chance
• Test for association use...

  Chi squared
• Test of correlation use...

  Spearman's Rank/Rho or Pearson's R
• Nominal data use...

  Chi Squared or Sign test
• Ordinal data use...

  Spearman's Rank/Rho, Wilcoxon or Mann Whitney-U
• Interval data use...

  Unrelated-T, Related-T or Pearson's-R
• If study uses independent group designs use...

  Unrelated-T, Mann Whitney-U or Chi squared
• If study uses repeated measures or matched pairs use...

  Related-T, Wilcoxon or Sign test
• If both co-variables are interval level use...

  Pearson's R