CHS 753 Data Skills

Population of units

a group of entities having some quantifiable characteristics in common

Population of observations

A group consisting of the numerical vallues of a quantifiable characteristics determined for each member of a population of units

Sample

A subset of a population

Sampling strategy

population —> sample —> sample statistic —> estimate population parameters

Statistical inference

A conclusion concerning a population of observations made on the basis of a sample of observation 

Population vs. Sample

Sample: Find reasonable estimators of the parameter (population)

Sample Probability samples

simple random sample 

stratified sample 

cluster sample

Simple Random Sample (SRS)

size n is taken when every possible subset of n units in the population has the same chance of being the sample. The simplest form of probability sample. Easy design and analyze

Stratified Sample

the population is divided into subgroups called strata. Then an SRS is selected from each stratum, and the SRS in the strata are selected independently. Often used for subgroups of interest

Cluster Sample

Observation units in the population are aggregated into larger sampling units, called cluster. Cheap, convenient to manage but less precise and extensive statistical analysis

Forms of Bias 

Selection, measurement, confounding factors 

Selection bias

convenience sample, purposefully selecting “representative” sample, mis-specifying the target population, failing to include all the target population in the sampling frame

Measurement bias

do not tell the truth, do not understand the questions, forget, different answers to different interviewers, particular interviewer, questionnaire design

Confounding factors

Situation in which a measure of the effect of an exposure on risk is distorted because of the association of exposure with other factors that influence the outcome under study 

Statistics

the science of collecting, organizing, summarizing, analyzing, and makign inferences from data

Descriptive statistics

includes collecting, organizing, summarizing, and presenting data. Used to summarize and present data related to public health issues. Focus on describing what is currently known about a population’s health without making inferences.

Inferential statistics

Includes making inferences, hypothesis testing, and determining relationships. Helps public health professionals determine rates after an intervention, is statistically significant and likely to apply to the entire population

Qualitative data 

Categorical-anything that is not a # (Sex, etc)

Quantitative data (numeric)

Anything that is measured (Age, weight, etc)

Measure of Effect

Calculate point estimate and confidence interval of the ‘risk’ associated with an exposure 

Rate ratio

If =1 there is no relationship between the exposure and the outcome

Interpreting measures of effects

OR= 1, no association

OR>1, risk factor

OR<1, protective factor

What is a p-value? 

measure the likelihood that the observed estimate is due to random sampling error 

What affects statistical power?

Beta error (type II)- lower statistical power

total sample size and group within sample size

alpha error (type I)- probability of rejecting a null hypothesis when it is actually true

T-Test

A statistical hypothesis test used to determine if there is a significant difference between the means of two groups or samples A

ANOVA 

analysis of variance- is a statistical test used to analyze the variation between multiple groups or conditions to determine whether there are statistically significant differences