Envr301.7

Envr301.7

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ls in the middle of an ordered sample  Not greatly affected by outliers o Mode –– Value that occurs most frequently • Statistical Inference 1. Calculate a test statistic on the base of the results in the sample and ¨¨ 2. compare that calculated test value with the known distribution of the test statistic (e.g. t, F, Chi2). Usually, use the variance. • The larger the test statistic, the more likely it is considered “statistically significant”, and the more confident a researcher can be that the effect is not produced by chance. • Hypothesis Testing o Testing a null hypothesis H0 against an alternative hypothesis H1. o The null hypothesis is always the ‘no change, no effect’ statement: could mean that an observed effect is produced by chance (no effect of the independent variable), or that no change in the dependent variable occurs at all, or that the...
View Full Document

This note was uploaded on 03/29/2013 for the course ENVR 301 taught by Professor Jeffreycardille during the Winter '13 term at McGill.

Ask a homework question - tutors are online