Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
36 Cards in this Set
- Front
- Back
|
Independent variable
|
AKA "predictor" variable; variable believed to cause or influence the dependent variable; purpose: to describe
|
|
|
dependent variable
|
AKA "outcome" or "response" variable; variable of interest; hypothesized to depend on or be caused by another variable; purpose: to be described
|
|
|
Nominal
|
Category; No necessary relationship among categories; Numbers may be assigned to the categories as labels for computer storage, but the choice of numbers is totally arbitrary; Ex: Sex, race, marital status, diagnosis, presence or absence of a sign or symptom
|
|
|
Ordinal
|
Classification by ordered categories low to high, but not able to say how much higher or lower - not equal spacing between items; Numbers assigned to objects according to their relative standing to each other, assignment is not arbitrary; Goes beyond just categorizing, actually rank orders attributes; Ex: Ratings of pt. satisfaction (very satisfied, satisfied, neutral, not satisfied, very dissatisfied); categories of income (low, middle, high); attitudes (strongly agree, agree, no opinion, disagree, strongly disagree); stage of tumor
|
|
|
Interval
|
Distances between any 2 numbers on scale of known & equal size; Not only categorizes & ranks objects, but also orders objects both by the magnitude of the numbers & the relative size of differences between pairs of objects; Does not have an absolute zero point that indicates complete absence of the variable being measured; the zero point is arbitrary; Because the units are in equal intervals, can add & subtract across an interval scale.; Ex: IQ, temperature on the Fahrenheit scale (arbitrary zero point)
|
|
|
Ratio
|
Like interval, but has absolute zero that indicates total absence of the attribute; Because the zero point is not arbitrary, can also multiply & divide, as well as add & subtract, across a ratio scale.; Ex: Age, height, weight, blood pressure
|
|
|
In deciding whether to treat a variable as ordinal or interval consider _(3)_
|
measurement perspective; the # of categories that comprise an ordinal scale; the concept of meaningfulness
|
|
|
Define Continuous Variable
|
Infinite # of possible values that fall between any 2 observed values; divisible into an infinite # of fractional parts; Can theoretically assume all possible values along a continuum; Result from measuring, rather than counting (as for numerical discrete variables). Ex: Ht, wt, age
|
|
|
Define Numerical Discrete/ Catagorical Discrete
|
categories into which the values of the variables fall are not qualitative, but quantitative; can't take on intermediate levels; integer. Ex: class attendance per day, # of episodes of angina during a 1 month period, # of live births (can't have a fraction of a student, angina attack, or live birth) No FRACTIONS
|
|
|
Population vs. Sample
|
Population: collection of all individuals of interest, all members of a defined group; Sample: subset of a population selected for study
|
|
|
chi square
|
X = ___________
o=bserved frequency e= expected frequency E= (# in row) x (# in column) ------------------------- total |
|
|
degrees of freedom
|
df = (r-1)(c-1)
r= row c= column extent to which values are free to vary |
|
|
Measures of Association
Odds Ratio |
odds= calculated by dividing the # of times an event happens by the # of times it does not happen
OR + ratio of 2 odds OR = ad -- bc |
|
|
sensitivity
|
TP
------------ TP + FN TP = test positive FN = False negative The proportion of those who truly have the characteristic who are correctly classified as having it by the test - measures how well a test detects the characteristic it is testing for - the lower the FN rate, the higher the sensitivity |
|
|
Specificity
|
TN
---------------- TN + FP TN= true negative FP= False positive The proportion of those who truly do not have the characteristic who are correctly classified as not having it by the new test - measures how well a test detects absence of characterisitc; how well it correctly identifies those who do not have the characteristic - the lower the FP the higher the specificity |
|
|
Positive Predictive Value
|
TP
-------- TP + FP + test = results who have the disease |
|
|
Negative Predictive Value
|
TN
-------- TN + FN - test = free of disease |
|
|
Confidence Interval
|
is an interval estimate of a population parameter. Instead of estimating the parameter by a single value, an interval of likely estimates is given. How likely the estimates are is determined by the confidence coefficient. The more likely it is for the interval to contain the parameter, the wider the interval will be.Confidence intervals are used to indicate the reliability of an estimate
|
|
|
If the CI includes the number 1,
|
then the association described by the OR is not statistically significant
|
|
|
correlation
|
used to describe the relationship between two variables
|
|
|
simple linear regression
|
used when goal is to PREDICT value of the dependent variable for a given value of the independent variable
|
|
|
Pearson Correlation
|
both values of I/R
|
|
|
Relationship between X and Y
|
Direction = + or -
Strength = measures how well the data fits on a straight line r = correlation co-efficient r + -1 to 1 1= perfect + -1 = perfect - 0 = no correlation |
|
|
r value
|
0 - .25 = little or no value
.26 - .50 = low or fair .51 - .75 = moderate .75 and up = excellent corr. |
|
|
meaningfulness of r
coefficient of determination |
2
R = can range from 0 - 1 |
|
|
regression equation
|
Y = a + b X
Y= predicted value of variable Y a= intercept constant b= slope of line x= actual value of variable x |
|
|
regression line
|
a line drawn through a scatterplot of two variables
|
|
|
determining the Y intercept and Slope
|
Y = a + bX
|
|
|
Slope
|
SS xy
b = -------- SS xx |
|
|
Correlation Coefficient
|
r=
|
|
|
Comparing Correlation and Regression
|
1. a corrlation coefficient tells you that some sort of relationship exists, but doesn't tell you much more.
2. correlation is scale independent, but regression is not. 3. The slope of the regression line has the same sign (+ or -) as the correlation coeffient (r) 4. If the correlation coefficient is statistically significant, the regression equation will also be statistically significant. |
|
|
simple linear regression
bivariate |
determine if a linear relationship exists between IV and DV
-predict value of DV based on given value of IV Variable IV= 1 I/R DV = 1 I/R Ex: predict the length of a child at birht from the birth weight of the child |
|
|
multiple regression
|
extension of simple linear regression in which there is >1 IV
should improve prediction of the value of the DV- shoul dbe able to predict Y with greater accuracy if have more information - used to understand the effects of 2 or more IVs on a DV - Variables 1 I/R DV - > 1 any level IV (aka predictor variable) |
|
|
power analysis
|
method for reducing risk of type II error
|
|
|
type I error
|
in reality there is no difference, but you conclude that there is
researcher attains statisical significantly falsely a= probability of making a Type I error |
|
|
type II error
|
in reality there is a difference, but you conclude that there is not
researcher reports no significant difference between groups when in actuality there is B - probability of making a type II error |
