Biostatistics Review

Define this:

Names or labels categories with no specific order

Cannot calculate mean or average data

Examples: Color, Sex, Types of Chocolate

Nominal Data

Define this:

Categories into a meaningful and ranked order (a logical sequence)

Examples: Small, Medium, and Large

Ordinal Data

Define this:

Most precise level of measurement

Zero is arbitrary

Meaningful intervals

Examples: Temperature, Size

Interval

Define this:

Ranked in specific order that includes a constant difference in magnitude of change

Zero is NOT arbitrary

Examples: Blood Pressure, Enzyme activity, Weight, Age

Ratio

Define this:

A statistical test that makes specific assumptions about a population’s probability distribution and population parameters like the mean and variance

Parametric test

Define this:

The odds that a person affected by a condition was exposed to a risk factor versus the odds that a person unaffected by the same condition was exposed to the same risk factor.

Example: Odds 1 = 60/40; Odds 2 = 30/70

(60×70)/(30×40)

Odds Ratio

Define this biostatistical test:

2 or 3 categorical variables and determining if the two variable might be related or correlated

Chi-Square Test

Define this biostatistical test:

Used to compare two categorical variables, to determine correlation but with a very small (30 or less than participants)

If one of the cells is less than five then we also use this test

Fisher’s Exact Test

Define this biostatistical test:

Compares the average mean of one number value between two variables (one categorical & one continuous)

Example: Average blood pressure in a drug group vs. a placebo group

t-test

Define this:

Compares the averages between three or more groups

Uses one continuous dependent and three or more categorical

Analysis of Variance (ANOVA)

Define this:

a variable that serves in place of an unobservable or immeasurable variable

Proxy Variables

Define this:

How well a patient follows an agreed-upon medical treatment plan

Adherence

Define this:

A nominal variable that contains only two categories or levels

Seen typically in percentage or proportions

Ex: Male/Female

Dichotomous

Define this:

An input, which may be varied or simply observed by the researcher. Sometimes called an experimental or predictor variable

Independent variable

Define this:

The output, outcome, or effect of interest

Dependent variable

Define this:

Variable that have a finite number of possible value; typically nominal or ordinal variable

Discrete (Categorical)

When is ordinal data treated like continuous?

Likert agreement scale (Strongly agree, or disagree)

Define this:

Error associated with how participants are selected for studies

Selection Bias

Define this:

When the interviewer influences the participants response duirng an interview

Interviewer Bias

Define this bias:

The error associated with remembering

Recall Bias

Define this:

Error associated with not selectively submitting and/or publishing research

Publication Bias

Define this:

A mathematical value describing the numerator which is also included in the denominator

Proportion

Define this:

Frequency distributions of multiples variables represented in a single table

Contingency Table

Define this:

The odds of exposure among diseased versus non-diseased

Uses non-events rather than total events

Odds ratio

Define this:

Frequency distributions of multiple variables represented in a single table

Contingency table

Define this:

The likelihood of developing the disease/outcome given that one is exposed divided by the likelihood of developing the disease/outcome given that one is non-exposed in total out of an event.

Relative risk

Define this:

Do the questions look like they measure what they say they are measuring

Face Validity

Define this: Is it generalizable beyond the sample

External Validity

Define this:

How well did it measure a specific construct

Construct Validity

Define this:

The extent of agreement between two or more raters

Interrater reliability

Define this:

The degree to which rest or instruments scores are consistent from one point in time to the next

Test Re-test Reliability

What is the alpha value?

0.05

Define this statistical test:

Use if n > 30 and the continuous variable is NOT approximately normal

Must be stated in paper

Parametric test

Define this:

Three or more groups, with one continuous and two categorical variables.

Must use a post-hoc comparisons

ANOVA

Which of the following variables could be analyzed via student t-test? (age, weight, insurance (insured versus non insured), or # of medications)

Insurance (insured versus non-insured)

What is the non-parametric equivalent of paired t-test?

Wilcoxon

Data that are organized according to no categories are also called:

Nominal

What type of dependent variable is used in logistic regression?

Categorical

Define this;

Non-parametric test

Used to test two variables correlations

Spearman’s Rho

The probability that an individual given Aspirin will present an MI relative to the probability that an individual given a placebo will present that same outcome is called»

Odds

What test are most associated with ordinal data?

Non-parametric

Define this:

Non-parametric test

Compare two independent groups (t-test)

Mann Whitney U

Define this:

Non-parametric

Ranked

similar to ANOVA

Kruskal Wallis

Define this statistical test:

Statistical model able to predict a continuous outcome (using multiple predictors)

Linear regression

Define this:

The change in our outcome for every one unit change in our predictor variable

Unstandardized B

What correlation rule of thumbs is 0.7 - 1.00?

Very strong

Correlation does not always mean??

Truth (coincidence can happen)

What are the two things that describe a p-value?

Strength and Correlation

Define this:

Holding all other variables in the model at their mean; assuming all other things are equal

Controlling for

Define this:

Standardized to allow for comparison across variables

Largest magnitude of effect

Beta

What variables can regression models use?

Categorical (dichotomous) and Continuous

Define this:

Astatistical method used for data grouped or nested in multiple levels, such as students inside schools or patients inside hospitals. It handles data dependencies by adding random effects, which lets group baselines or trends change safely without breaking basic testing rules

Multi-level regression

Define this:

A set of statistical methods used to study the time until an event happens (such as Death, disease, reoccurrence,

Survival analysis

Define this:

A dichotomous outcome (Can be either continuous or categorical)

It calculates the probability that a given data point belongs to a specific category (sigmoidal)

Presented as odds ratio

Better fit closer to the 0 and 1 (binary or dichotomous outcome)

Logistic regression

What is the non-parametric equivalent to survival anaylsis?

Kaplan Meier method

What is the survival analysis of three or more groups?

Cox Proportional Hazards

What test would we use to measure difference between the three groups for continuous variable? If low sample?

ANOVA and Kruskal Wallis (Wallis)

What test would we use to measure differences between the three groups for categorical data? If less than 5?

Chi-Square and Fisher Exact Test