Correlation quiz

Effect sizes for r²values

"0.01= small effect, 0.09= medium effect,0.25= large effect"

Effect Sizes in Correlation

"1. Weak (.1 to .2) 2. Moderate (.3 to .6) 3. Large (.7 to 1)"

Odds Ratio

"Odds that a case was exposed vs Odds that a control was exposed"

Correlation Coefficient

"Represents the degree and the direction of the relationship between 2 variables - Ranges from -1.00 to +1.00. - The stronger the relationship, the closer to -1 or +1. The weaker the relationship, the closer to zero"

relative risk reduction

"measures the decrease in relative risk due to intervention

Null Hypothesis Significance Testing(NHST)

"• used to determine ""statistical significance"" • based on probability (p(row) < .05) • come up with a hypothesis • then come up with a null hypothesis (opposite) • NHST tell us the probability of our data, given the null hypothesis"

Relative risk

(RR or risk ratio) is the ratio of specified adverse event of treatment to the incidience to those not given treatment

Scatter Plots

a graph represented by ordered pairs(x, y) Allows us to see if relationships exist between two variables.

Null hypothesis (H0)

a hypothesis stating that no relationship or difference exists between two variables. Also called statistical hypothesis.

Moment

Distance from a mean is called a

-1 to + 1

correlation coefficents range

One-tailed p-value

0.025- applicable when the alternative hypothesis requires the significance to be in one tail rather than the other.

Two Tailed p-value

0.05- applicable when the alternative hypothesis did not specifically predict in which size (tail) of the probability distribution of the significance would be detected

Positive Correlation

A correlation where as one variable increases, the other also increases, or as one decreases so does the other. Both variables move in the same direction.

Negative Correlation

A finding that two factors vary systematically in opposite directions, one increasing as the other decreases.

linearity

A measure of how well data in a graph follow a straight line, showing that response is proportional to the quantity of analyte.

Type II error

Accepting the null hypothesis when it is false

Different Perspectives

Advantage of using more than 1 effect size indicator

Steps for calculating pearson r

List all the X values in a column List the values of all the X values squared (X2) in a new column Calculate the ΣX Calculate the ΣX2 List all the Y values in a column List the values of all the Y values squared (Y2) in a new column Calculate the ΣY Calculate the ΣY2 Calculate the value of XY for each subject, list in a new column Calculate the ΣXY Determine your N (the number of paired scores) Plug the appropriate values into the formula

Hypothesis

Null Hypothesis: no correlation/relationship between two variables reject the null: if absolute value of r is equal to or greater than the critical value of r accept null: if absolute value of r is less than the critical value of r

Point-biserial (r pb)

One continuous and one dichotomous variable, such as the correlation of subjects gender with their performance on the SAT-verbal.

Alternative Hypothesis

Other hypotheses that are potentially true and prevent us from 'proving' a single hypothesis true.

Types of correlations

Pearson r (product-moment correlation) (rxy) two continuous variables Spearman Rank Correlation (Spearman rho) (rs) two ranked variables Point-biserial correlation (rpb) one continuous and one dichotomous variable Phi coefficient () two dichotomous variables

Regression line

Rather than offering the deviation of each score from the mean, regression looks at the distance between each score and ___________.

Type I Error

Rejecting the null hypothesis when the null hypothesis is true.

Pearsons r

Standard of linear relationship, possible values -1.0 to + 1.0

Correlation

Statistic for determining the association of two variables

Relative Risk (RR)

The risk of disease or injury in one group compared with another group. The two groups usually differ in terms of one or more key factors (e.g., physical activity levels). RR is usually expressed as a risk ratio comparing incidence or prevalence rates among two groups.

RD - Risk/Rate Difference

The risk/rate of disease among exposed persons that can be attributed to the exposure. (Difference in the risk/rate of exposed and nonexposed).

Linear

What relationship does a correlation coefficient reflect between to variables

When is a phi coefficient used?

When both are dichotomous

Covariation

a principle of attribution theory stating that for something to be the cause of a particular behavior, it must be present when the behavior occurs and absent if it doesn't occur; if your boyfriend is irritable only when you spend extended time with others (high covariation)

Third Variable Problem

a variable correlated with X and Y is suspected to be cause of both

Product-moment correlation

another name for Pearson r

Correlation Coefficient

calculated numerical value 1. indicates the direction of the relationship 2. indicates the strength/extent of the relation 3. ranges from +1.00 to -1.00

Prep

commuting __ will give you an estimate of the probability of replicating the same direction of effect as reported in each original study

Stronger

he further away from zero the correlation coefficient is, the __ it is a. Small = .1 b. Medium = .3 c. Large = .5

counternull statistic

measure of non-null magnitude of effect size

Zero Correlation

no relationship between the variable high score on one variable may be associated with a high OR low score on second variable

Power analysis

performed to determine sample size required to obtain a desired level of power and the probability that a Type II error was committed; formulas exist for calculating both power & sample size for a variety of statistical tests; calculations based on these formulas are presented in tables; use of tables requires determination of effect size; sample size & power calculations should be based on the primary outcome variable for the study

Temporal Precedence

principle that presumed "cause" must be shown to have occured before presumed "effect"

Relative Risk Reduction

reduction in relative risk expressed as percentage

Beta (b)

the probability of a type II error (if Ho is not rejected when it is incorrect).

alpha (a)

the probability of making a type I error

Pearson r

two continuous variables

Phi Coefficient

two dichotomous variables

Forms of correlations: Phi Coefficient ( (|) )

two discreet variables (e.g.,) correlation of gender w/ SAT performance

Spearman Rho (rs)

two ranked variables

Continuous variable

value can fall between any two adjacent scores (e.g.,) age

Dichotomous variable

variable divided into two distinct parts (e.g., gender)

Dummy coding

when numerical values such as 0 and 1 are used to indicate the two parts of a discreet variable __ mused in calculations of correlations

to predict occurrence of events

why use correlations

maximum correlation

± 1.00: a. perfect association b. the dots are more tightly clustered