public:t-701-rem4:ultimate_note_on_statistical_tests
Differences
This shows you the differences between two versions of the page.
| Next revision | Previous revision | ||
| public:t-701-rem4:ultimate_note_on_statistical_tests [2007/11/17 19:37] – created helgi | public:t-701-rem4:ultimate_note_on_statistical_tests [2024/04/29 13:33] (current) – external edit 127.0.0.1 | ||
|---|---|---|---|
| Line 1: | Line 1: | ||
| ====== The ultimate note on statistical tests ====== | ====== The ultimate note on statistical tests ====== | ||
| - | ** | + | |
| (Ultimate because it is most likely the last one before the exam) | (Ultimate because it is most likely the last one before the exam) | ||
| Line 9: | Line 9: | ||
| * F-test (not really the details) in anova tables for linear models for testing the whole model | * F-test (not really the details) in anova tables for linear models for testing the whole model | ||
| - | * Shapiro test for normality | + | * Shapiro test for normality |
| Line 16: | Line 16: | ||
| **We have a random sample from a (often infinite) population.** The population is either explicit or hypothetical. | **We have a random sample from a (often infinite) population.** The population is either explicit or hypothetical. | ||
| - | **We have a precise hypothesis called the null hypothesis (H0) to be tested**. Loosely speaking, "being tested" | + | **We have a precise hypothesis |
| Often, H0 states that the effect of some independent variable on another dependent variable is zero (thus null-hypothesis). The independent variable can by either a grouping (gender of a living being, type of machine, ..., called qualitative variables, classifications, | Often, H0 states that the effect of some independent variable on another dependent variable is zero (thus null-hypothesis). The independent variable can by either a grouping (gender of a living being, type of machine, ..., called qualitative variables, classifications, | ||
| Line 40: | Line 40: | ||
| The only rule of thumb we have for directly interpreting test statistics is for the t-value: If the absolute value of t is smaller than 1.5, we never reject H0 (p-value much higher than 0.05), if it is larger than 4 we always do (p-value very small), if it is in between we need to choose a threshold for rejection and look at the p-value. | The only rule of thumb we have for directly interpreting test statistics is for the t-value: If the absolute value of t is smaller than 1.5, we never reject H0 (p-value much higher than 0.05), if it is larger than 4 we always do (p-value very small), if it is in between we need to choose a threshold for rejection and look at the p-value. | ||
| + | {{public: | ||
/var/www/cadia.ru.is/wiki/data/attic/public/t-701-rem4/ultimate_note_on_statistical_tests.1195328227.txt.gz · Last modified: 2024/04/29 13:32 (external edit)