Term
Discrete Variables
2 types |
|
Definition
Also known as counting variables
N: nominal = no order (e.g. gender)
O: ordinal = order, but no consistent difference in magnitude change (e.g. trama scale) |
|
|
Term
|
Definition
| NNT= the number of patients who would have to receive treatment for 1 of them to benefit. |
|
|
Term
|
Definition
| NNH= the number of patients who would have to receive the treatment for 1 of them to experience and adverse effect. |
|
|
Term
Nominal Variable
Defintion and examples |
|
Definition
Nominal variable also called "attribute or categorical" variable. A good rule of thumb is that an individual observation of a nominal variable is usually a word, not a number. Examples include sex (possible values are male or female), Genotype (values are AA,Aa, or aa), ankle condition (possible values are normal,sprained, or broken). |
|
|
Term
Ordinal Variable
Definition and examples |
|
Definition
Ordinal variable also called ranked variable, are those for which the individual observation can be put in order from smallest to largest (interval scale) even though the exact values are unknown. There is no consistent difference in magnitude change. Examples, pain scale
0-10, ) 0 being no pain, 10 being the worst possible pain, or Trauma Score. |
|
|
Term
Continuous Variables
2 types |
|
Definition
Also called measuring variables
I: interval = in order with consistent interval difference
R: Ratio = like interval, but zero is the starting point |
|
|
Term
Interval Variable
Definition and examples |
|
Definition
| Interval variable are things you can measure. An individual observation of a measurement variable is always a number. Consistent intervals occure between the measurements. Example, Temperature or ph. |
|
|
Term
Ratio variable
Definition and examples |
|
Definition
|
|
Term
|
Definition
| Median= Mid-most value of a data distribution. Most useful for describing ordinal data. NOT useful to describe nominal data. |
|
|
Term
|
Definition
p≤ 0.05 is significant
(0.05 is a 5% likeley hood an event will occur by chance)
the lower the p value the less likely an event occurred by chance
|
|
|
Term
Tests to use with Nominal
2 samples (independent, parallel)
no confounders |
|
Definition
Chi Square (X2)
Fisher's exact |
|
|
Term
Tests to use with ordinal variables
2 samples (independent, parallel)
no confounders
|
|
Definition
Wilcox Rank Sum
Mann Whitney U |
|
|
Term
Tests for Continuous Variables
(Interval and Ratio)
2 samples (independent, parallel)
no confounders
|
|
Definition
|
|
Term
SEM
(Standard Error of the Mean)
Formula |
|
Definition
SEM=SD/square root of N
SD = Standard Deviation
N = total number of data points |
|
|
Term
NNT
(Number needed to treat)
formula |
|
Definition
NNT = 1/ARR
(ARR = Absolute risk reduction) |
|
|
Term
AAR
(Absolute Risk Reduction)
formula
|
|
Definition
AAR = the arithmetic difference between 2 event rates; varies with the underlying risk of an event in the individual patient
Absolute risk of control - Absolute risk of active group
expressed as a Percentage |
|
|
Term
RRR
(Relative Risk Reduction)
Formula |
|
Definition
RRR =1-RR (relative risk)
or
AAR/event rate in control group
RRR =
Difference in 2 groups/untreated group
expressed as a ratio of 2 percentages |
|
|
Term
SD
(Standard Deviation)
percentages of 1 and 2 STD |
|
Definition
One STD = 68% of population
Two STD = 95% of population
Only meaningful when applied to data that is normally distributed. It is applicable to interval or ratio scale data. |
|
|
Term
|
Definition
| Type I error = can only be found if a statistical difference is found |
|
|
Term
|
Definition
| Type II Error - can only be found when a result is NOT significant |
|
|
Term
Regression Analysis
defintion |
|
Definition
| A predictive model where associations are derived |
|
|
Term
Correlation (r)
defintion |
|
Definition
| correlation (r) quantifies the linear relationship between variables--strength of association |
|
|
Term
Coefficient of Variation
(r2) |
|
Definition
| coefficient of variation explains amount of variation that is explained by r |
|
|
Term
|
Definition
CI tells the magnitued of difference between comparative groups.
A CI that includes zero is not statistically significant (p>0.05)for prospective trials
All values in a CI are statistically indistinguishable |
|
|
Term
|
Definition
Are compared to baseline risk of 1 for comparision
>1 = increased risk
<1 = decreased risk
data range for either cannot include 1 or there is no difference |
|
|
Term
|
Definition
| Mode=is the most commonly obtained value or highest point of a peak on a frequency distribution. Useful to describe nominal data, defining the most prevalent characteristic of a sample. |
|
|
Term
|
Definition
variance = Σ (mean-x1)2/(n-1)
x1 = each individual data point
n = the total number of data points |
|
|
Term
|
Definition
AAR = the arithmetic difference between 2 event rates; varies with the underlying risk of an event in the individual patient
AAR becomes smaller when event rates are low, while RRR often remains constant.
|
|
|
Term
|
Definition
| SD = square root of variance |
|
|
Term
|
Definition
SEM is simply a quantification of the variability of these sample mean values. It is properly used to estimate the precision or reliabilty of a sample, as it relates to the population from which the sample was drawn. It is use to calculate CI NOT to describe sample data variability.
SEM decreases with increases in sample size. |
|
|
