lecture 1 variables + CI Flashcards
independent variable
Experimental or predictor variable
Can be manipulated in an experiment
Changed to have an effect on a dependent variable
dependent variable
Outcome variable
waht is the independent and dependent variable:
study being done to determine efficacy of new analgesic at different doses
Independent variable: dose of medication
Dependent variable: change in pain scale
Nominal variable
categorical outcomes with MORE than two possible outcomes; not numeric
- categories in NAME only, w no particular order
i. e. blood types A, B, AB, O - -> there can be a mode but nO mean (avg b/t type A & B)
Binomial (dichotomous) variable
categorical outcomes with TWO distinct possible outcomes
- subset of nominal
i. e gender (yes/no)
Ordinal variable
express rank and order matters (though not the exact value) (pain scale, level of education, restaurant ratings 1-5 stars)
Continuous variable
represent data capable of possessing any value in a given range (BP, temperature, weight)
- NUMBER
i. e. BP 110 to 120 is the same as 120 to 130–> 10 pt difference
Interval variable
continuous spaced with equal intervals or distances; the zero point is not considered meaningful (example: IQ or temperature)
-diff b/w 5F and 4F is same as 60F to 59F
-0 degrees does not mean NO temperature, it can go lower **
when you change the scale in the y axis (starts at 0 then change scale to 25) does it mean results are different?
no. it’s a matter of how data is represented, makes it look different but not
Ratio variables
cant go below zero
difference between interval and ratio
ratio can be calculated bc 0 point DOES matter
mean
the average
- add all values and divide by total
median
middle value
- put in order
is the median influenced by the outliers
negative
is the mean sensitive to outliers
yessss. –> mean wont be very representative
- it skews the result
what does “n” mean in a study
number of observatiions
what is the percentile of median
50%
if the mean is very similar to the median…. this means….
not a lot of outliers
trimmed mean
ignoring the highest and lowest (usually a percentage)
–removing influence of outliers*
mode
value that occurs most commonly in data set
- not useful w continuous variables
- does not always assess the center of a distribution
does the mode assess the center of a distribution
no
error
variability in the data
most of the scatter in biologic and clinical studies is due too…
biologic variation
i.e. aging, diet, mood
bias
anything that skews the data one way or another
-influence study