Statistical Tests

• Statistical tests are used to determine whether a set of data differs significantly from outcomes that might be produced from chance
• A significant difference/correlation is one which is unlikely to have occurred due to chance
• P values are the probability of the difference/correlation being down to chance
• A null hypothesis is a statement that says there is no significant difference/correlation between the variables being investigated
• The statistical test will result in either rejecting or accepting the null hypothesis
• A correlation coefficient test is used when looking for a correlation between two different vairables
• A student t test is used when looking for differences between two means
• A chi-squared test is used when involving data about numbers of individuals in categories
• Degrees of freedom in a student t test is calculated by (n1 - 1) + (n2 - 1)
• If the value calculated is higher than the critical value, the result is significant and the null hypothesis is rejected. There is a less than 0.05 probability the difference is due to chance.
• If the value calculated is lower than the critical value, the result is not significant and we accept the null hypothesis. There is a more than 0.05 probability the difference is due to chance.
• A biological conclusion must be made at the end of each statistical test
• Degrees of freedom for chi-squared tests are calculated using number of groups - 1