Statistics Flashcards
Define a nonparametric test. (name of a common nonparametric test)
They are tests used to look at an experiment utilizing ordinal data with 2 or more groups or a study using continuous or discrete data with 2 or more groups.
The Mann-Whitney rank sum test
Describe a paired t-test vs. an unpaired t-test
Both are used for experiments utilizing continuous or discrete data when there are 2 groups present
In a paired t-test, each subject in an experiment has 2 measurements taken (i.e. one before getting an experimental drug and one after)
In an unpaired t-test, measurements are taken from 2 separate groups of patients (sample size may or may not be equal in the 2 groups)
Describe when an analysis of variance test should be done to analyze a study.
It is used for studies using continuous or discrete data with greater than 2 groups (i.e. an experiment that wants to study heart rate at different times after induction using different induction drugs in each group.)
When is a chi-square test best to be used?
For a study that involves dichotomous data (only 2 possible values: benefit vs. no benefit) and the sample groups are from 2 populations
Define “conflict of interest” in reference to physicians. What is the primary source of conflict of interest with anesthesiologists?
Conflict of interest refers to “a set of conditions in which professional judgement concerning a primary interest (such as patient welfare) tends to be unduly influenced by a secondary interest (i.e. financial, personal, and professional = gaining prestige, promotion, and respect).
For the anesthesiologist, the primary source of conflict of interest is production pressure.
Describe the 2 types of error that can occur in an experiement.
1) Beta error: the probability of falsely ACCEPTING the null hypothesis (falsely saying there is no difference) when it should be rejected. If this value is high, then the power of the study is decreased.
* Power = 1 - beta
2) Alpha error: the probability of REJECTING the null hypothesis (falsely saying there is a difference), when the null hypothesis is true.
What is a null hypothesis
Saying there is no difference between 2 (or more) groups
What is the P value and how is it obtained? What does its value say about the experiment?
P value is computed from the observed data as the probability of obtaining this data set if in reality the null hypothesis were true.
If P is less than the alpha value (the level of significance of a study), the result is statistically significant and the null hypothesis should be REJECTED.
Typically the alpha is set to be 0.05, meaning there is only a 5% chance of obtaining a specified data set by chance alone and not due to the experiment itself
Describe detection bias
The investigator over or under pursues a subpopulation and thus the “detection” is skewed because the entire population is not surveyed accurately.
Describe recall bias
Seen when subjects are asked to recall events in the distant past
Describe selection bias
It occurs when the group to be tested is not representative of the overall population
Describe a cohort study
A cohort study design looks at an exposed group and an unexposed group to a particular factor and analyzes the incidence between the two groups over time
How is Absolute Risk Reduction calculated?
Control event rate (CER) - experimental event rate (EER)
What is the minimum microshock current required to elicit ventricular fibrillation?
100 micro amps
Describe the difference between an ordinal variable vs. interval variable vs. a categorical variable
An ordinal variable has an order but the interval between each entity is not constant (i.e. obese, non-obese, thin & the ASA classification system).
With an interval variable, the difference between each data point is a constant interval (i.e. ages 5 yr vs. 15 yrs vs. 25 yrs).
A categorical variable is the direct opposite of interval in that the variables have no ordering (i.e. ondansetron, dexamethasone, droperidol, metoclopromide)