Lecture 1 - Statsitics, Data & Graphics Flashcards
population:
complete set of units (e.g. all individuals of a species; all cars in the world) that possess some common characteristic
sample:
a smaller (but hopefully representative) set of units from a population; used to find the true properties about that population
what do statistics help with?
statistics help us to find signal amongst noise (random variation)
parameter:
from Ancient Greek meaning “beyond measurement”
population parameters and their units:
mean: µ
variance: σ2
standard deviation: σ
variable:
(characteristic, property) measured on units (e.g. individuals, species): e.g. size, hormone concentration, age, number of genes
what do we try to predict or explain with variables and data? we
we try to predict or explain a response variable from one or more explanatory variable (s)
graph needed when: categorical data and one variable
(pie chart) bar graph
graph needed when: categorical data with 2 or more variables
(grouped) bar graph or stacked histogram
graph needed when: numerical data and one variable
histogram (distribution of numerical variable), bar graph or box-and-whisker-plot
graph needed when: numerical data with two or more variables
scatterplot
what should each sample be?
each sample should be random and unbiased
random sample:
each member of a population has an equal and independent chance of being selected
bias:
a systematic discrepancy between the estimates and the true population characteristic
how does measuring usually take place in an experiment?
we normally cannot measure whole populations (all individuals, units etc.) but just a sample
