General Stats and Odds Ratios Flashcards Preview

SBM II December > General Stats and Odds Ratios > Flashcards

Flashcards in General Stats and Odds Ratios Deck (35):
1

What are some common types of analysis?

Univariate statistics
Bivariate statistics
Multivariate statistics

2

What are univariate statistics?

-Frequency
-Mean
-Median
-Mode
-Standard deviation, etc.

3

What are bivariate statistics?

Relationship between 2 variables. Chi-square, Pearson's correlation (r), t-test.

4

What are multivariate statistics?

Relationships among more than 2 variables

5

What are the four main Scales of Measurement?

NOIR
Nominal
Ordinal
Interval
Ratio

6

What are nominal measurements?

Measures mere differences (ex. hometowns, gender)

7

What are ordinal measurements?

Measures differences + rank order (ex. place finish in a race; SES [working class, middle class, etc.], religiosity)

8

What are interval measurements?

Measures differences + rank order + equal distances between values of the variable (ex. IQ - few examples of these non-meaningful zero values)

9

What are ratio measurements?

Measures differences + rank order + equal distances between values of the variable + true, meaningful zero point (ex. income)

10

What two Scales of Measurement are categorical/dichotamous?

Nominal, Ordinal

11

What two Scales of Measurement are continuous?

Interval, Ratio

12

What is the definition and examples of Nominal?

2 or more categories/groups; measures differences
(ex: eye color, city of birth)

13

What is the definition and examples of Ordinal?

2 or more categories/groups with order/rank; measures differences + order
(ex: place finished in a race, SES status)

14

What is the definition and examples of Interval?

Numerical value measured on a continuum; measures differences + order + equal units of measurement
(ex: temperature, IQ)

15

What is the definition and examples of Ratio?

Numerical value measured on a continuum; measures differences + order + equal units of measurement + true/absolute zero point
(ex: height, weight)

16

What things should you keep in mind with Scales of Measurement?

-How we measure phenomena is a choice (ex: race, gender)
-Technically, the statistical test you run depends on: how your variables are measured, your research question
-It is common for researchers to break test & measurement rules

17

What statistical test is used for nominal vs. nominal measurements?

Chi-squared

18

What statistical test is used for Nominal (2 groups only) vs. Interval or Ratio?

t-test

19

What statistical test is used for Nominal (2 or more groups, 1 IV only) vs. Interval or Ratio?

One way analysis of variance (F-test) - ANOVA

20

What statistical test is used for Nominal (2 or more groups, 2 or more IV) vs. Interval or Ratio?

Factorial design analysis of variance (F-test)

21

What statistical test is used for Ordinal/Interval/Ratio vs. Ordinal/Interval/Ratio?

Pearson Correlation Coefficient

22

What is the most common stat you'll see with categorical outcomes?

Chi-Square test and accompanying p-value
-Categorical = dichotomous outcomes = yes/no values (often as %'s)

23

What are quantitative outcomes also called?

continuous variables

24

What is an example of when you'd do a Chi-Square test?

Ex. comparison of the % of men vs % of women who are diagnosed with major depression
Ex. Comparison of freshman, sophomore, junior and senior UMD undergrads earning at least 1 "F" letter grade.

25

When is a Chi-Square test significant?

When P-value significant data

26

What does a t-test do?

Another way of comparing values.
-Checking to see if means are significantly different
-T-test designed to compare MEANS - mean values are generally generated for continuous values

27

What is the Odds Ratio (OR)?

Measure of association between two variables (ex. an IV and a DV)

28

What is another way of defining OR?

"Odds" that some outcome would occur in the presence of a predictor (IV) vs. absence of the predictor

29

When are Odds Ratios often discussed within medicine?

Case-control studies
-Though they are approbate for all kinds of categorical data

30

In many studies we don't actually know incidence data. What do we use to estimate relative risk in these cases?

Odds Ratio

31

What does OR = 1 mean?

Exposure does not affect odds of outcome.

32

What does OR > 1 mean?

Exposure associated with higher odds of outcome.

33

What does OR

Exposure associated with lower odds of outcome.

34

What is the equation for OR?

AD/BC

35

What is A, B, C, and D?

A = Disease +, RF +
B = Disease -. RF +
C = Disease +, RF-
D = Disease -, RF-