BEES2041 Lecture Notes - Lecture 3: Limburger, Contingency Table, Cay
Examples of categorical data
-number of males & females in a sample
-presence/ absence of certain type of plant/ animal/ mineral
-human blood types
-colour morphs of plants / animals
-behavioural trials with few outcomes
-counts of soils/ rock types
Do the counts of observations in our categories differ from expected values
-goodness of fit tests
Are there associations between two / more categorical variables
-contingency tables
Goodness of fit tests (GFT)
-test how well observed counts fit to expected counts under a specified model
Chi square goodness of fit test: are males & females equally likely to choose BEES2041
Female: 74(56%) Male: 59(44%)
Goodness of fit tests can be extended to:
As many categories as you like
Any proportions among the expected ratios
Chi square goodness of fit test:
Mendel’s peas, ae flowe olous podued in expeted 3:1 atios?
Chi square goodness of fit test: are there spatial patterns to sediment deposition on coral
cay?
-island divided into four zones
-sediment deposition measured over time
-Rejecting the Ho tells us there is a difference somewhere
-examination of the residuals (observed-expected) can indicate which cells differ most from
the expected
Contingency tables
-test for associations between two/ more categorical variables
GFT showed cheese was more attractive than a control with no cheese
A contingency analysis: are mosquitoes from Liberia more attracted to Limburger cheese
than mosquitoes from Tanzania
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Number of males & females in a sample. Presence/ absence of certain type of plant/ animal/ mineral. Do the counts of observations in our categories differ from expected values. Are there associations between two / more categorical variables. Test how well observed counts fit to expected counts under a specified model. Chi square goodness of fit test: are males & females equally likely to choose bees2041. Goodness of fit tests can be extended to: Rejecting the ho tells us there is a difference somewhere. Examination of the residuals (observed-expected) can indicate which cells differ most from the expected. Test for associations between two/ more categorical variables. Gft showed cheese was more attractive than a control with no cheese. A contingency analysis: are mosquitoes from liberia more attracted to limburger cheese than mosquitoes from tanzania. Tree species with nests compared to availability in local environment. No more than 20% of the cells have expe(cid:272)ted f(cid:396)e(cid:395)uen(cid:272)ies less than 5(cid:894)(cid:862)spa(cid:396)se (cid:863) ta(cid:271)les(cid:895)