First 365 words of the document:
Geographical skills theory
Testing for relationships and correlations
There are three types of correlation or relationship, these are; positive
correlation (when one value increases so does the other), negative correlation (as
one value decreases so does the other) and no correlation (there is no link between
the two values, so they do not influence each other). Statistical tests are used to
find out the type of correlation but also the strength of that correlation. To be
confident of a significant relationship you must be at least 95% or above certain
that the null hypothesis can be rejected.
Used when both sets of data can be easily ranked, a quick and easy measure of
correlation is needed and where exact values may be uncertain. This calculation
can indicate if two sets of data are related. A scatter graph can be drawn first to
give a rough idea f the expected result. These can show you if the correlation is
positive or negative.
Firstly data must be ranked, and must be ranked in an appropriate way to suit
your null and alternative hypotheses so from either smallest to largest or vice
The result will lie between +1 (a prefect positive correlation) and -1 (a perfect
negative correlation) with 0 representing no correlation. The exact significance of
your result is checked by looking up the result in a table of significance.
However spearman's ranks shouldn't be used when there are a lot (more than 4)
sets of tied ranks; where there are a limited number of data sets as usually at
least 8 are needed and where two data sets are unequal in number.
Chi squared test
This is most useful when your data has been collected in categories or when you
can group it into categories (when your data is nominal). It tests if there is a
difference between the observed pattern and the expected pattern from chance
events. For the test to be valid, the total number of observations should be more