Epidemiology Week 3: Part 2 bias

definition: a feature of the design or conduct of a study that leads to skewed/unrepresentative findings; systematic flaw; intentional or unintentional

bias

three types of bias in this course

• selection

• experimenter

• social desirability

definition: inclusion of indiv or data in a way that is not representative of the population intended to be analyzed

selection bias

self-selected respondents is which bias? (leads to over-representation)

selection

what studies are susceptible to selection bias?

surveys and studies where people drop out

how to eliminate selection bias?

• avoid it by getting a representative sample and obtain a high response rate and a low drop out

• if it is unavoidable list as a limitation

explain clever hans and experimenter bias

horse (Hans) whose owner (von Olsten) claimed it could do math and do intellectual tasks. he was actually picking up on nonverbal cues and body language; involuntary tensing and releasing of questioners. Hans would get it right if he saw the questioner and wrong if he didn’t therefore proving experimenter bias

definition: experimenters thoughts/expectations/lack of impartiality influence data toward a certain result

experimenter bias

definition: investigator subtly communicates their own expectations to participants this skewing findings; subconscious or conscious

experimenter bias

what is the problem w “rate your pain”

the person who asks may signal about an expected or desired answer or mechanism of single blind study

how to eliminate experimenter bias?

double blind study

definition: tendency of people to fill out surveys to answer questions in a way that will be viewed favorably by others

social desirability bias

which bias: people over-report the good and under-report the bad when asked about themselves

social desirability

T or F: anonymity completely eliminates social desirability bias

false

randomized response technique helps eliminate which bias?

social desirability

example of randomized response and coin technique

• tell student if they flip heads to answer Y even if they haven’t cheated

• tell student if they fip tails to answer truthfully

• investigator does not know if heads or tails

• use math to estimate the % who have cheated

  • about half flip tails so subtract 50 yesses then see how many Y vs N and that’s the probability

T or F: a large sample size is needed for the random response technique

true