Research methods #1

Research methods #1

1. Statistical tests

   • Chi-squared test
     • This is used when:
       • The data is nominal
       • It is a test of association
       • The data are independent
     • Alpha p in psychology = p<0.05 (less than or equal to)
       • If value is less than alpha p (e.g. 0.005) its statistacally significant & we can reject our Null hypothesis (HO)
   • Mann-Whitney U
     • This is used when:
       • Testing for a difference
       • There is an independent groups design
       • The data are ordinal or interval/ration
   • Spearman's Rho
     • This is used when:
       • Testing for a correlation
       • The data is ordinal, interval or ratio
       • There are no or very few tied ranks
   • Wilcoxon Matched Pairs test
     • This is used when:
       • Testing for a difference
       • There is a repeated measures design
       • The data is ordinal, interval or ration
   • Key words (Levels of data)
     • Ordinal data
       • Info or scores are put in order. Whenever data is ranked, it is ordinal.
     • Nominal data
       • Info is put into categories or just named
     • Interval data
       • This is obtained when there are equal intervals on a measurement scale.
     • Ratio data
       • Similar to interval in that there are equal intervals across the scale but in addition there is a true zero point.

2. Types of errors

   • Type 2
     • Error that occurs when the null hypothesis is retained but is false. (Anything that is greater than the alpha p value)
   • Type 1
     • AKA 'False positive'. Error that occurs when the null hypothesis is rejected but it still should have been retained

3. Types of hypothesis

   • Directional hypothesis (HA)
     • A directional prediction of what is expected to happenv (AKA 'One tailed test')
       • Example - 30mins of sunlight will increase mood scores.
   • Non- directional
     • The difference between two variables exist. But doesn't state the direction of that differennce. (AKA. 'Two tailed test')
       • Example - 30mins of sun will affect mood scores.
   • Null hypothesis (HO)
     • The null hypothesis states that there is no relationship or difference between two sets of data.