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;
32 Cards in this Set
- Front
- Back
|
Analytic method that should be selected depends on:
|
the nature of the data to be analyzed.
|
|
|
Characteristics of the sample and variables
|
Number of independent and dependent variables
Data scale the variables are measured Nominal, ordinal, interval or ratio |
|
|
Distribution of the variables
|
Parametric vs. non-parametric test
|
|
|
Overview of statistical tests
|
see slide 5
|
|
|
parametric tests
|
The variable is normally distributed or an assumption is made based on a large enough sample to consider the variable normally distributed (central limits theorem)
The variable is continuous or, if it is discrete, it at least approximates a normal distribution. The variable is measured on an interval or ratio scale. |
|
|
non-parametric test
|
Considered to be distribution-free methods
Useful in analyzing nominal and ordinal scale data ** when in doubt, use the non-parametric test ** |
|
|
T-tests for Independent samples
|
Statistical method to test for differences between the means of two independent groups.
--Dependent variable: interval or ratio scale --Independent variable: nominal (binary) Null hypothesis assumes that the means of the two populations are equal |
|
|
T-tests for Independent Samples
assumptions: |
The two samples are random samples drawn from two independent populations of interest
The measured variable is approximately normally distributed and continuous. The variable is measured on an interval or ratio scale The variance of the two groups are similar |
|
|
example of T-test
on slide 9 thur 13 |
true
|
|
|
Mann-Whitney U Test
Also called the Wilcoxon rank-sum test what type? para or non para? |
Non-parametric test
|
|
|
Mann-Whitney U Test
when is it used? |
when data are measured
Ordinal scale Non-normally distributed (compare with 2 independent samples) |
|
|
Mann-Whitney U Test
test of the equality of: |
medians rather than means
|
|
|
Mann-Whitney U Test
A more powerful test than student T when the assumptions of the T test are violated!! |
true
|
|
|
Mann-Whitney U Test example
|
Suppose your are interested in examining risk factor for death in patients with S. aureus infection. In particular, you are interested in examining the relationship between death and length of hospitalization prior to infection. You hypothesize that patient who die have a longer length of hospitalization prior to infection than patients that do not die.
What would you do? see slide 16 thru 20 |
|
|
Matched or Paired Data
When two samples are NOT drawn from two independent populations: |
One group is measured twice
--Pre and post treatment/ intervention --Purpose is to decide whether the intervention makes a difference Subjects are matched on one or more relevant characteristics --Analyzed as though they were the same subject |
|
|
Pairing data:
|
the same subject is used to collect data for both groups
--Cross-over design --Pre and post test |
|
|
Matched data:
|
match subjects for certain characteristics
--Lodise et al. CID 2002. --Comparison of outcomes between vancomycin resistant enterococcus (VRE) bacteremia and vancomycin susceptible enterococcus (VSE) bacteremia |
|
|
Normally distributed continuous data
is: |
paired T-test
|
|
|
Non-normally distributed continuous data or ordinal scale
|
Wilcoxon signed-ranks test
|
|
|
Paired design allows the researcher to detect the change more readily
|
true
|
|
|
Chi-Square for Two Independent Samples
used for: |
≥1 groups and compares the actually number within a group to the expected number for that same group
Expected number is based on theory, previous experience or comparison groups |
|
|
Chi-Square for Two Independent Samples
address research questions related to: |
rates, proportions, or frequencies
|
|
|
Chi-Square for Two Independent Samples
which data ? |
nominal and ordinal
|
|
|
Chi-Square for Two Independent Samples
assumptions: |
--Frequency data
--The measures are independent of one another --Categorization of the variables “Yes” or “No” “Presence” or “Absence” “Treatment” or “placebo” “Exposure” or “No Exposure” |
|
|
The chi-square is the most commonly used method for comparing frequencies or proportions.
|
true
|
|
|
Chi-Square for Two Independent Samples
Categorical data often arranged in a table consisting of: |
columns and rows.
Rows→ categories of one variable Columns→ categories of the other variable |
|
|
“2” x “2” table
Chi-square is a comparison of the expected frequencies in each cell compared to the actual or observed frequencies in those same cells Can be used to measure differences and/or associations |
true
|
|
|
Chi-Square for Two Independent Samples
example |
Suppose you are interested in examining the relationship between prior antibiotic exposure and MRSA. You hypothesize that patients with MRSA are more likely to have a prior antibiotic exposure than patients without MRSA.
How would you address your hypothesis? see slide 29 to 32 |
|
|
Other Methods of Inference for Categorical Data
Fisher's Exact Test |
--Cell within a matrix→ expected frequency <5
--Small sample size |
|
|
Other Methods of Inference for Categorical Data
McNemar's Test |
(Paired or Matched Data)
|
|
|
Overview of Statistical Tests
see slides 34 and 35 |
trued
|
|
|
Comparing more than two groups
Analysis of Variance: |
--The null hypothesis assumes that the means of the various groups being compared in the study are not different
--Assumptions Each of the groups is a random sample from the population of interest The measured variable is continuous The variable is measured on a ratio or interval scale The error variances are equal The variable is approximately normally distributed --If mean difference detected→ Post Hoc Comparisons |
