Lecture 1 Flashcards
(29 cards)
Accuracy
how close average of values are to true value
Precision
how close measured values are to each other
Standard deviation
measure of precision (not accuracy)- measures the variation in averages
Inferential statistics
Draws conclusions about data
Descriptive statistics
Describes a data set- Not trying to draw conclusions
Categorical Variables /Ex:
Have discrete or qualitative values (gender, pattern of shirt, liking of tofu )
-can be nominal, ordinal, or dichotomous
Continuous variables /Ex:
Measured along a continuum (Height, age, years as a nurse, temperature)
-can be interval or ratio
Nominal
No intrinsic order
Ordinal
Have order
Dichotomous
only 2 values
Interval
Numeric value & is measured (temp, age, height)
Ratio
Like interval, but value of 0 indicates there is nothing
What is a researcher trying to determine from inferential statistics?
Are two groups different?
Chance
random variations- bigger sample size (N) will reduce change errors
Bias
not random, caused by systematic variation- bigger sample size (N) will not help with bias
Selection bias
Bias sampling of population
Measurement bias
Poor measurement technique
Analysis bias
Analysis favors one conclusion over another
Confounding
Misinterpretation of accurate variables- bigger N will not fix confounding (bowling alley ex)
POEM
Patient Oriented Evidence that Matters- morbidity and mortality
DOE
Disease Oriented Evidence - most articles are DOE because it is measurable and faster- suggestive, but not conclusive
Clinical Trial
Non-Randomized- placebo vs drug group are assigned
Randomized Controlled Trial
Random assignment to exposure (or drug, etc)
Cohort Study aka Longitudinal
Observational study - subjects with and without exposure are identified and followed forward in time- time consuming, but little bias
