Biostats Midterm

parameters

is a numerical value that describes a characteristic of an entire population.

what we are interested in measuring

estimate

our best guess at a parameter within a sample

ex. averages, %s, proportions, counts

null hypothesis

the “no difference” hypothesis

sampling error

the chance difference between an estimate and the population parameter being estimated

precision

do you get the same/similar measurements each time you sample

bias

systematic discrepancy between estimates and the true population characteristic

accuracy

are you measuring what you actually want to measure

random sampling

each unit has an equal and independent chance of being included

sample of convenience

collection of easily available individuals

undesirable, avoid if possible

volunteer bias

Volunteers are not representative people

• intereseted in topic

• more willing to be subjected to treatment

• financial rewards\

• have more time (limits demographic)

categorical variables

describe membership in a group

• ex. languages, experimental vs control, sex chromosomes, colour, rank etc.

nominal

• Type of categorical variables

• No inherent order

• ex. colour, group, chromosomes

Ordinal

• type of categorical variable

• groups can be ordered in a meaningful way

  • ex. age, size, status

numerical variables

measurements are quantitative

ex. body temp, height, age etc.

continuous data

any value within a reasonable range

no negative numbers typically

discrete

indivisible units

tends to be integers

ex. cant have ½ a human

probability distribution

show the distribution of a variable in the whole population

theoretical distributions approximate real life frequency distributions

frequency

number of observations having a particular value (count)

frequency distribution

how often each value of the variable occurs in the sample

experimental pros

• control over sample

• helpful for causation

experimental con

can be unrealistic / unrepresentative or unethical

observational pro

representative of sample/ true population

observational con

cant control sample

difficult to prove causation

relative frequency

proportion of observations

• proportions of different categories sum to 1

categorical data graphs

bar graphs

must start at 0

numerical data graphs

histogram

frequency tables

histogram

uses the area of rectangular bars to display the frequency distribution or relative frequency distribution of a numerical value

abundance interval

• a bin that groups samples into categories of abundance

uniform histogram

bell shaped histogram

asymmetric (skewed)

bimodal histogram

left skewed

right skewed