Biostatistics (Chapters 1, 2 , and 9)

population

everyone or everything we’re interested in learning about

descriptive statistics

learn about the random sample

variable

anything of interest that is measured

qualitative

descriptive and nonnumerical

nominal

categorical

quantitative

numerical

ordinal

order

interval

can only be + / -

can’t act like 0

ratio

can be */÷/+/-

can act like 0

discrete

can list all possible outcomes

continuous

not discrete (contains continuum)

random sample

every individual in the population has an = chance of being in the sample

simple random sample

1. every individual gets a number

2. randomly selected #s

stratifies random sample

representative sample

1. divide population into a few large groups

2. within each group, take a simple random sample

cluster random sample

easier logistically

1. divide the population into many small groups

2. randomly select small groups

everyone in the chosen groups are in the sample

systemic random sample

adequate representation

1. sort the population by something important

2. pick a random start

3. every kth individual is in the sample

convenience sample

1. not random

2. destroys the study

misleading and biased

good experimental design

1. random sample

2. control group

3. random assignment (w/o risk that you’re observing the effects of your own assignment process)

4. double blind

5. replication (min. 10x)

6. readability

7. blocking - all groups are well represented in the study

*case study is not an experiment

statistical inference

conclusion about a population on the basis of the information contained in a sample from that population

problems with experiments

1. placebo effect

2. bias

3. fraud

4. confounding variables

5. willful ignorance

n

size of sample

statistics

an area of study concerned with collecting and describing data as well as making statistical inferences

measurement

the assignment of numbers to objects or events according to a set of rules

N

population size

IQR

interquartile range

C.V.

coefficient of variation

k

number of class intervals

μ

population mean

R

range

computation is advantageous, but it only takes into account two values,r so poor measure of dispersion

s

standard deviation

s²

sample variance

σ²

population variance

relative frequency

frequency/sample sizeh

histogram shapes

mound shaped

uniform

skewed right or left (named after the side it’s light on)

*can use multiple words to describe

*at least 7 intervals

*trial & error (too many gives jagged, too few can’t see)

*no overlap

*no gaps

*cover the whole range

*same length

measures of spread

more spread out - further from the center

average square distance from the center - sample variant (s²)

sample standard deviation = s

chebycheu’s inequality

at least 75% of the sample is between x bar - 2s and x bar + 2s

at least 89% of the sample is between x bar - 3s and x bar + 3s

outlier

a data point which is several (2 to 3.5) s away from x bar

percentiles