BIOSTATS 100A EXAM QUESTIONS AND ACCURATE ANSWERS 2024/2025 |
VERIFIED
statistics - ✔️✔️branch of science that deals with drawing inferences from samples to populations
double-blind - ✔️✔️experimenters and individual both don't know what is given or gave to control BIAS
margin of error - ✔️✔️statistical uncertainty, uncertainty in Confidence Interval estimate (half the width
of interval)
sampling - ✔️✔️sample should be representative of the population, has statistics
population - ✔️✔️has parameters
simple random sampling - ✔️✔️every element of the population has chance of being selected
random sample - ✔️✔️probability sample
stratified randomization - ✔️✔️separate based on covariates (age, gender, etc.), covariates could affect
data, within strata is homogeneous, across strata is heterogeneous, opposite of cluster randomization,
independent study groups
block randomization - ✔️✔️balance measurements in each strata, could be expensive
Hawthorne Effect - ✔️✔️being in an experiment changes your behavior, not placebo
cluster randomization - ✔️✔️sample based on geography, within cluster is heterogeneous, across
clusters is homogeneous, opposite of stratified randomization
systematic random sampling - ✔️✔️starting point is random, beware of periodicity
population frame - ✔️✔️database, ex. registry of voters, DMV
, non-response bias - ✔️✔️only people who want to respond will respond, this affects data
population-based weighting - ✔️✔️scale measurements according to their weights in the population,
ex. more men responded than women to poll, each man's vote is weighted only 4/5 of a vote
non-probability sample - ✔️✔️not every element of population has chance of being selected, ex.
polling, person on the street, clinical research
descriptive statistics - ✔️✔️1. graphical format/tabular
2. numerical
3. frequency distribution
percentiles - ✔️✔️measurement such that X% are at or below
quartiles - ✔️✔️rank statistics
right skew - ✔️✔️median<mean, negative skew value
left skew - ✔️✔️median>mean, positive skew value
probability sample - ✔️✔️type of sample where every element of the population has a change to be
chosen, subject to uncertainty
outliers - ✔️✔️box plot, beyond 75th + 1.5 IQR or 25th - 1.5 IQR
interquartile range - ✔️✔️middle 50% of data
geometric mean - ✔️✔️more representative of central tendency for skewed distribution (take log, find
arithmetic mean of log, exponentiate mean)
VERIFIED
statistics - ✔️✔️branch of science that deals with drawing inferences from samples to populations
double-blind - ✔️✔️experimenters and individual both don't know what is given or gave to control BIAS
margin of error - ✔️✔️statistical uncertainty, uncertainty in Confidence Interval estimate (half the width
of interval)
sampling - ✔️✔️sample should be representative of the population, has statistics
population - ✔️✔️has parameters
simple random sampling - ✔️✔️every element of the population has chance of being selected
random sample - ✔️✔️probability sample
stratified randomization - ✔️✔️separate based on covariates (age, gender, etc.), covariates could affect
data, within strata is homogeneous, across strata is heterogeneous, opposite of cluster randomization,
independent study groups
block randomization - ✔️✔️balance measurements in each strata, could be expensive
Hawthorne Effect - ✔️✔️being in an experiment changes your behavior, not placebo
cluster randomization - ✔️✔️sample based on geography, within cluster is heterogeneous, across
clusters is homogeneous, opposite of stratified randomization
systematic random sampling - ✔️✔️starting point is random, beware of periodicity
population frame - ✔️✔️database, ex. registry of voters, DMV
, non-response bias - ✔️✔️only people who want to respond will respond, this affects data
population-based weighting - ✔️✔️scale measurements according to their weights in the population,
ex. more men responded than women to poll, each man's vote is weighted only 4/5 of a vote
non-probability sample - ✔️✔️not every element of population has chance of being selected, ex.
polling, person on the street, clinical research
descriptive statistics - ✔️✔️1. graphical format/tabular
2. numerical
3. frequency distribution
percentiles - ✔️✔️measurement such that X% are at or below
quartiles - ✔️✔️rank statistics
right skew - ✔️✔️median<mean, negative skew value
left skew - ✔️✔️median>mean, positive skew value
probability sample - ✔️✔️type of sample where every element of the population has a change to be
chosen, subject to uncertainty
outliers - ✔️✔️box plot, beyond 75th + 1.5 IQR or 25th - 1.5 IQR
interquartile range - ✔️✔️middle 50% of data
geometric mean - ✔️✔️more representative of central tendency for skewed distribution (take log, find
arithmetic mean of log, exponentiate mean)
