Statistical Tests

- Used to identify the presence of a relationship between two data sets and the strength of the relationship

- Presented in a table with title of variable and rank within column (1 = largest)

- The difference in ranks is then recorded in a separate column

- The difference is then squared in the final column (d2)

- Value of d2 recorded and multiplied by 6

- 'n' is the number of data sets, for example sites in an investigation

- Values put in to formula

- Result compared to significance table; probability of occurring by chance is predicted

- Relationship between distance inland and percentage vegetation cover

- Confirmed presence of relationship using table of significance

- A comparative test to show differences between data sets and whether these differences occurred by chance; the statistical significance of the difference

- All values ranked together (both columns)

- Lowest value ranked first

- Lowest U value compared to significance table

- U value is calculated; if exceeds U value in critical values table, there is evidence sufficient to reject null hyptothesis

- U value refers to normal distribution

- Does not identify reasons for difference 

- Data must be ordinal (able to be ranked)

- Must have at least 5 samples in data set; no more than 20

- Used to test the differences in pebble roundness between the west and east end of criccieth beach

- Critical value for U for a sample size of n1= 6 and n2= 5

- Calculated U value for our results is 1.5

- This is less than the critical value; we can be 95% certain that there is a statistical difference between pebble roundness on the west and east sides of the beach

- We can therefore reject the hull hypothesis

- This supports our hypothesis that attrition due to long shore drift has occurred, making the pebbles rounder as they are transported from the west to the east

- During LSD material is rolling and grinding over each other, gradually wearing away the pebble and smoothing any sharp, angular edges.