Lect 1 - Biostats

Question 1

Why random sampling

Answer 1

evens out biases

Question 2

why sample

Answer 2

ideally study everyone
but cost too much
therefore use sample that you hope is representative of everyone

Question 3

what is probability sample

Answer 3

chosen so you can determine how likely it is that an individual from population is included in that sample

Question 4

what are the 4 different probability sampling methods

Answer 4

1. simple random sampling - randomly chosen, e.g. calling phone books. every individual has equal chance of being selected
2. systematic sampling - every 3rd person, this spreads samples more evenly, and easier to conduct - but this may not always be random
3. stratified - e.g. by age or gender, so you can have equal numbers - then randomly chose - this gives greater precision, but bit more complex, can over sample, e.g. older males
4. cluster sampling - cluster chosen from population, then sample randomly taken from each cluster. e.g. different streets, then randomly sample from the street - can introduce bias such as socioeconomic status

Question 5

whats non-probability sample

Answer 5

impossible to determine the likelihood that an individual comes from the population

Question 6

what are some non-prob sampling methods

Answer 6

1. convenience sampling - ask med students to participate
2. quota sampling - fill quota
3. purposeful sampling - exemplify
4. snowball sampling

Question 7

2 types of data

Answer 7

1. categorical

2. measurement (continuous)

Question 8

4 scales of measurements

Answer 8

nominal - gender
ordinal - rate of health
interval - temperature (no real 0)
ratio - height (true 0)

Question 9

what are the differences between descriptive and inferential

Answer 9

1. descriptive

2. inferencial

Question 10

advantage/disadvantages of mode

Answer 10

Represents the largest number of people.

Is a real score from the data set

Can be used with nominal data.

May not be particularly representative of the 
entire data set.

Cannot be manipulated algebraically

Not as stable across samples as mean.

Question 11

median

Answer 11

Relatively unaffected by extreme scores (useful 
for skewed distributions).

Can be used with ordinal data.

Cannot be manipulated algebraically.

Not as stable across samples as the mean.

Question 12

mean

Answer 12


Can be manipulated algebraically.

Most stable estimate of central tendency across 
samples 
(i.e. less variance across multiple samples).

Influenced by extreme scores.

Its value might not exist in the data.

Should only be used with interval or ratio data.