Intro to Biostatistics

What is the the method by which facts or data are used to estimate truth?

Statistics

What is a field that applies statistical methods to answer questions in public health, medicine, or biology?

Biostatistics

Biostatistics provides tools and techniques for ____,____,____, and _____ ____

collecting, summarizing,

analyzing, and interpreting data

What is individual pieces of information recorded & used for analysis?

data

What is a collection of individuals who share at least one common or organizing characteristic?

Population

What is a sample?

subset of the population of interest

What is a measure of computed for the entire population of interest?

Population parameter

What is a measure taken from the sample that is

used to estimate the population parameter?

Sample statistic

What is it called when you use sample-based data to infer conclusions about the population?

Estimation

What is the degree to which characteristics of the sample correspond to the characteristics of the population from which the sample was chosen?

Representativeness

If samples taken are _____ of the population of interest, they will provide good ____ for the population overall, and allow us to make an ____ about the population samples were drawn from.

representative, estimates, inference

What are FOUR examples of how we pick individuals for our study sample?

◦ Convenience sampling

◦ Systematic sampling

◦ Simple random sampling

◦ Stratified random sampling

What is the goal of sampling?

obtaining a representative sampling

What sampling technique subjects selected by an arbitrary and easily performed method?

Convenience sampling

Convenience sampling is prone to ____ ___ and (is/isnt) recommended

selection bias, isnt

What sampling technique shows investigators select subjects at a regular interval selected in advance?

Systematic sampling

T/F Systematic sampling is easier when you cannot access a list of characteristics of your population in advance

True

What sampling method produces unbiased estimates of parameters, and is the best way to ensure good internal validity and reduce impact of confounding variables?

Simple random sampling

What process shows that every individual has an equal chance of being selected?

Random process

What sampling technique has the population is divided into strata or groups based on certain characteristics and then sampled using random sampling?

Stratified random sampling

When would you use a Stratified random sampling method?

Used when a population is diverse and investigators want to make sure every group is

properly represented in the sample

What are some examples of strata?

age, farm, parity, breed

T/F The larger your sample size, the greater the statistical "power" of your study

True

Must choose the sample size needed for your research question prior to ____ your study

beginning

T/F Association means Causation

false, it does not mean

Association does not imply ____

causation

T/F Confounding, bias, or other sources of error

can create misleading relationships, so

multiple lines of evidence are required before

concluding cause and effect

True

What provide the best evidence in favor of causality, but you may not be able

to perform an RCT due to ethical reasons or rare disease/event?

Randomized controlled trials

What TWO kinds of studies can show an association between exposure and outcome. but if you find a statistically significant association between exposure and outcome, you still cannot say that the exposure CAUSED the outcome?

Cohort studies or case control studies

How do you determine causation?

•Establishing an association between the exposure and

outcome is a crucial first step.

•Then consider the entire body of evidence to try and arrive at a reasonable conclusion about the relationship between the exposure and outcome & potential causality.

Wha are the criteria for causality? (8)

•Strength of association between exposure and outcome

•Consistency or reproducibility of results

•Temporality: exposure or risk factor must precede outcome

•Specificity: single cause/single effect

•Biological gradient

•Plausibility/coherence

•Experimental evidence

•Analogy

With the criteria for causality, which is the only criteria that MUST be present for judgement of

causality?

Temporality

What criteria for causality that has dose-response relationship between cause and outcome?

Biological gradient

What criteria for causality shows findings go along with current evidence to date and makes biological sense?

Plausibility/coherence

What are THREE sources of error?

Random error

Bias

Confounding

What occurs when values from your sample (sample statistics) differ from the values (parameters) of the parent population?

Random error

____ ___ is the reason why if we take different

samples and measure them, we might get a slightly different estimate every time.

Random error

What can you do to decrease random error?

Increase sample size!

What SYSTEMATIC ERRORS IN A STUDY THAT LEAD TO FALSE CONCLUSIONS (IMPACT ACCURACY?

Bias

What is a consistent deviation from the truth due to study design, data collection method, or analysis?

Bias

Errors due to bias are ___ and ___

systematic and

repeatable

What do random errors arise from?

random

fluctuations in measurements

What type of bias is a measurement error

in assessment of exposure or outcome?

Information bias

What type of bias differences between

those who participate and those who do not?

Selection bias

What type of information bias "instrument over or underestimates measurement"?

Detection Bias

What type of information bias questions that

influence responses?

Interviewer bias

What type of information bias in a retrospective study, if participants cannot recall details/exposures well?

recall bias

What type of information bias investigators or

participants modify behavior or responses when they are aware of what group they're in?

Performance bias

What type of information bias result of

investigator's previous knowledge of the hypothesis or of an individual's specific status?

Observer bias

What are THREE things that you can do to control information bias?

standardized protocol, blinding, quality control

What is a systematic distortion of the

exposure-outcome relationship due to the way

participants are selected or retained in the study?

Selection bias

What occurs when the method of

selecting participants results in a sample that

differs systematically from the target population?

Sampling bias

What occurs when participants who leave the study differ from those who remain?

Loss to follow-up (attrition bias)

The following is an example of what?

Studies that recruit individuals directly from clinics miss all those who do not seek veterinary care or cannot afford it

Selection bias

What is distortion of the association between an exposure and outcome by an extraneous, third variable called a confounder?

Confounding

What are THREE things that can be used to describe a Confounding variable?

◦ Is correlated (associated) with both variables of interest

◦ Has its own effect on the outcome

◦ Is not in the causal pathway between exposure and outcome

What are THREE things that can hep you to avoid confounding?

• Perform highly controlled experiments: eliminate all other sources of variation between the two groups except the exposure itself

• Randomly assign treatments

• Attempt to identify confounder ahead of time, collect data on it, and include it in your analysis

____ ____ can cause inaccurate (biased) 'estimates of effect' in observational studies, even if the confounding variables are not known.

Confounding variables

Association or causation? (tbh just read)

Dogs that wear flea collars are more likely to have fleas. Do flea collars cause fleas?

No, dogs prone to flea infestations are more likely to have a flea collar placed

reverse causation, if collars are placed after infestation. Or you could consider that underlying flea risk is the confounder here

Example of Association or causation?

A dog develops IMHA within two weeks of receiving a vaccination.

Did the vaccine cause the IMHA? There was a temporal association, right?

Consider selection

bias here. Dogs that are vaccinated are more likely to have owners that have access to and seek veterinary care.

Example of Association or causation?

Pets that are insured have more cancer diagnoses. Does having pet insurance cause higher rates of cancer?

◦ No. Owners that utilize pet insurance are more likely to pursue aggressive diagnostics if their pet is ill.

selection bias because uninsured pets are

less likely to be tested for cancer

In 12 Otterhound dogs (rare breed), 2 develop lymphoma.

Does this mean there is a breed predisposition to lymphoma in Otterhounds?

◦ No Why?

The sample size is extremely small, resulting in substantial random error and imprecise risk estimates. With such a small denominator, clustering could occur by chance.