Statistical Test Flashcards
Number of standard deviations s that a data point on a curve lies from the curves mean
Z score
What type of statistics is a z score used for?
Descriptive
interval which is expected to typically contain the parameter being estimated
Confidence interval
What does the confidence interval depend on?
- Sample size
- Size of the standard deviation
- Degree of confidence that you want
What does the p value depend on
- Difference between the means that are being compared
- The standard deviation of the sample
- The size of the sample
number, calculated from a statistical test, that describes how likely you are to have found a particular set of observations if the null hypothesis were true
P value
What is the advantage of using a confidence interval?
Gives you a meaningful range of means to look at which augments the information of the p-value
What is the benefit of using a p-value
Gives you a single probability but can be very specific about the size of the probability
Regardless of the shape of a population curve you will get a normal curve if you take numerous repeated random samples of a set size from the population and plot the means of those samples.
central limit theorem
What does a very small standard error of the mean suggest?
That the sample mean is close to the population mean
The ____ this sample size (n), the ____ the standard error of the mean and narrower the curve
Larger… Smaller
Standard error of the mean versus standard deviation
Standard deviation measures the spread of the data
Standard error of the mean measures how well you know the population mean
Interpretation of a 95% confidence interval
You can be 95% sure that the true population mean lies within the 95% confidence interval
Used to compare one data point with a known population whose mean and standard deviation are known
Z score
Used to compare a sample mean that the known population is mean and standard deviation are known
Z test
Test used to compare means when the population standard deviation is unknown (which is typically the case)
T test
Compares a sample mean to a known population mean (like comparing the stressful sample mean with the mean of the US population)
Single sample t test
This test compares two distinct samples (like comparing Chesterfield sample with the sample from neighboring worry town )
Independent samples t-test or unpaired t-test
This type of test compares two matched samples ( like comparing the stressful sample with the same stressful people the previous year)
Dependent samples t-test or paired t-test
Give at least one example of when you would use a paired t-test
1 comparing the before and after of a treatment on a single patient
2. Comparing treatment on one side of the body with the other in a single patient
3. Comparing one twin with the other
4. Close matching of two different groups for many variables such as age, gender, cultural and economic background, medical history, education, and area of residence
Is the smaller or larger sem desired? Why?
Smaller, the smaller the SEM, the smaller the confidence interval and more certain the population mean
Incidence of a disease progression in people taking the placebo minus the incidence of disease progression in people taking the treatment
Absolute risk reduction
The incidence of disease attributed to the risk factor minus the incidence of disease in persons not exposed to the risk factor
Attributable risk
1 - the relative risk
Relative risk reduction
What causes differences between absolute risk reduction and relative risk reduction?
Since relative risk reduction is a proportion, you will see bigger results for rare diseases
Equation for sensitivity
Number of people who have the disease and test positive divided by the number of people who have the disease
Equation for specificity
Number of people who do not have the disease and test negative divided by the number of people who do not have the disease
Equation for positive predictive value
Number to test positive who have the disease divided by the number who test positive
For negative predictive value
Number of people who test negative and don’t have the disease divided by the number of people who test negative
Compares One data point or sample mean with the mean of a population of known standard deviation
z test
Compares sample mean to a known population mean
Single sample t test
Compares means of two distinct samples
Unpaired t-test
Compares means of two matched samples
Paired t-test
Calculate the effect size
Cohen’s d or glass delta
Compares means of more than two groups; one independent variable factor and one dependent variable
One Way anova
Compare means of more than two groups; two independent variable factors, when dependent variable
Two-way anova
Compare means of more than two groups; two or more dependent variables
Manova
Control of influence of a covariate in anova test
ANCOVA
Correlation between one variable and another; interval or ratio data
Pearson correlation
Correlation between one variable and another; ordinal data
Spearman rank order correlation
Fit best straight line through a scattergram of linear data
Simple linear regression
Fit the straight line through a scattergram of nonlinear data
Simple nonlinear regression
Test association between categorical variables by comparing proportions against proportions predicted by theory
Chi-square goodness of fit test
Analyze contingency tables of categorical variables for statistical significance
Chi-square test of Independence
Analyze 2x2 contingency tables when cell sample size is less than five
Fisher’s exact test
Analyze contingency table with one category with two levels
Binomial test
Analyze contingency table with closely matched samples
McNemar test
Compare two unpaired groups, when their data do not have a gaussian distribution
Mann Whitney u test
Compare two paired groups when their data do not have a gaussian distribution
Wilcoxon test
Anova test that uses ranking
Kruskal Wallace test