Term
|
Definition
tail to the right
mean toward the tail
positive skew
mean > median > mode |
|
|
Term
|
Definition
tail to the left
mean toward the tail
negative skew
mean < median < mode |
|
|
Term
|
Definition
Rejection of a true null
False positive test |
|
|
Term
|
Definition
Acceptance of false null
False negative test |
|
|
Term
|
Definition
Categorical data (race)
Chi Square
Discete numbers with no rank order |
|
|
Term
|
Definition
|
|
Term
|
Definition
level of (pain, symptoms)
Discrete numbers with rank order only , no magnitude of diffrence between |
|
|
Term
|
Definition
Continuous and ordered: distance between variables has meaning
mean, sd, corelation, regression |
|
|
Term
|
Definition
Needs a zero point
(blood cell counts) |
|
|
Term
|
Definition
continuous probability distribution that arises when estimating the mean of a normally distributed population in
Measures the differences in the means
Situations where the sample size (less than 30) is small and population standard deviation is unknown
DF = N +1
sample size = 18 |
|
|
Term
| Descriptive numbers for samples |
|
Definition
parameter = population
statistic = sample |
|
|
Term
|
Definition
In general p should be less than 0.1 (small) an the number of trials should be >15 (large).
Need to be independent at a different time
Electrical system example: telephone calls arriving in a system.
Astronomy example: photons arriving at a telescope.
Biology example: the number of mutations on a given strand of DNA. |
|
|
Term
| Linear Regression analysis |
|
Definition
y= bx + a
y = dependent
b= slope
a = intercept
linear regression is used to describe and association as inverse or |
|
|
Term
|
Definition
Events that can not happen together
Probability of A or B happening
P(A) + P (B) |
|
|
Term
|
Definition
events A or B occurring together or in succession
P(A) * P(B) |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
|
Definition
The class intervals are diffrent
each data point must fit into one class
open classes are not allowed |
|
|
Term
|
Definition
need fixed outcomes
can be approximated by the normal distribution. |
|
|
Term
|
Definition
| 100 times the population standards deviation divided by mean |
|
|
Term
| standard score (Z-scores) of normal distribution |
|
Definition
| the number of standard deviations from the mean |
|
|
Term
| Regression Equation estimate |
|
Definition
| explains the dependent variables. The higher the correlation the better the regression equation |
|
|
Term
|
Definition
measures the extent that diffrent observations agree beyond what would be expected by chance alone.
K= Pobserved - Pexpected/ 100% - Pe |
|
|
Term
|
Definition
|
|
Term
|
Definition
Reciver Opperator Test : determine the accuracy of the test.
Y axis = sen
X axis = 1-spec (FPR)
Area under curve proportional to accuracy of the test |
|
|
Term
|
Definition
Used in Normal Distributions : more powerfull
Data must be interval or ratio
Can be used to describe measures of central tendancy |
|
|
Term
|
Definition
non-normally distributed data
or Ordinal scale data |
|
|
Term
|
Definition
Average
used for interval and ratio data |
|
|
Term
|
Definition
50% percentile
not affected by outliers
Usefull for Ordinal data |
|
|
Term
|
Definition
most usefull when two clusters exist.
It is the highest point of the peak.
Usfull for nominal data |
|
|
Term
|
Definition
| range, standard deviation |
|
|
Term
|
Definition
can be influenced by outliers
Applicable to all data sets nomianal, ordinal, interval, ratio |
|
|
Term
|
Definition
range betweeen the 25th and 75th interval.
Directly related to the median
Applicable to ordinal, interval, and range |
|
|
Term
|
Definition
Degree of scatter of individual data points about the sample mean
Usefull for normal distribution
= square root of varriance = sigma |
|
|
Term
|
Definition
= x-xi squared / n-1
= sigma saured
Applicable to interval and ratio data |
|
|
Term
| Standard Error of the mean |
|
Definition
scatter about the mean = SD/ square root of n.
SEM always < SD
As n increases SEM decreases |
|
|
Term
|
Definition
| probablity of making a type 1 error |
|
|
Term
|
Definition
probability of making a type II error
should be less than .2 |
|
|
Term
|
Definition
1-B = probability of not making a type II error. = sensetivity
increase alpha = increase power
increase n = increase power |
|
|
Term
| Operating Charateristic Curve |
|
Definition
| function that relates the dependent variable B that results from independent values of alpha, and n |
|
|
Term
|
Definition
More powerfull than a two tailed test.
Result must fall in either the top or bottom 5% of a distribution |
|
|
Term
|
Definition
t- test (diffence between means)
ANOVA
Multiple Comparison models |
|
|
Term
|
Definition
- Continuous data
- Normal distribution
- Variance same
- not paired (independent)
Paired T-test for dependent |
|
|
Term
|
Definition
a / (a+b) / c (c+d)
OR shifts RR away from the Null. |
|
|
Term
|
Definition
increase prevalence increase PPV
TP / (TP+FP) = a/ a+b |
|
|
Term
|
Definition
Receiver operator Characteristics
Y = sensitivity
X = FPR
The greater the area under the curve the greater the accuracy. |
|
|
Term
|
Definition
FNR = 1- sen = 1-TP
1-FNR = TP / sen |
|
|
Term
|
Definition
FPR = 1-spec = TN
1-FPR = TN/ Spec |
|
|
Term
|
Definition
used if the variables are both categorical/ dichotomous.
DF= r-1
Chi squared pearsons = diffrences in proportions. |
|
|
Term
|
Definition
- Continuous data
- 3 or more groups
- Independent |
|
|
Term
|
Definition
- Continuous data
- 3 groups or more
- Skewed
- Dependent |
|
|
Term
|
Definition
- Continuous data
- 3 groups or more
- Skewed
- Independent |
|
|
Term
|
Definition
- Continuous data
- 2 groups
- Independent |
|
|
Term
|
Definition
-Continuous data
-2 groups
-Dependent / paired |
|
|
Term
|
Definition
| test of multiple comparison for adjusted data |
|
|
Term
|
Definition
| Univariate test to asses the diffrences in proportions. |
|
|
Term
|
Definition
| Multivariate technique used for continuous data |
|
|
Term
|
Definition
| Multivariate technique used for dichotomous or categorical data |
|
|
Term
|
Definition
| test for linear trend w p values |
|
|
Term
| Cox proportional hazard curve |
|
Definition
| adjusted for confounders in the survival analysis curve. |
|
|
Term
|
Definition
| Comparison of Kaplan Meir Curves for treatments. |
|
|
Term
|
Definition
1.) Temporal
2.) Strenght of Ass
3.) Dose
4.) Consistency
5.) Plausibility
6.) Alt explanation
7.) Coherance |
|
|
Term
|
Definition
| even if a population has a normal distribution the sample mean from repeated samples from this distribution approaches the normal distribution the larger the N . Sample size must be greater than 30 |
|
|
Term
|
Definition
| R value give a indication of how closesly correlated |
|
|
Term
|
Definition
|
|
Term
|
Definition
| Chi Squared for dependent |
|
|
