5: Sampling, Random Error & Intervals ✅ Flashcards
Sample
Selected subset of a source population
Source population
Group of all individuals that we are interested in assessing
Purpose of a sample
To study something that cannot be studied as a whole due to practical restrictions
Source population can be
Can be general population of a sub-population
Source population in descriptive research
It is important the sample reflects the source population
Source population in Analytic research
In analytic research it can be more general depending on the research question
Source population: when investigating the biological effect of a disease
Source population can be more general
Source population: investigating social/ cultural effects
Source population has to be more restricted from where the population is derived
What should the sample be representative of?
The source population
Sampling frame
List of all the individuals in a source population
Sampling units
Individuals to be potentially selected
Most often individual people but can be sometimes larger eg families, streets, hospitals
Statistical interference
When sample estimate is used to draw conclusions on the population
It involves using stats to determine the degree of uncertainty
What are we measuring when using samples?
We are measuring estimates which carry sampling error
Parameter
Measurement of a quantity in a population
Estimate
Measurement of a quantity in a sample which aims to represent the true quantity
What does the sample estimate aim to do?
To quantify the population parameter
Sampling variation
Difference between different sample estimates
Sampling error
Difference in magnitude between sample estimates and the actual population parameter
-> caused by measuring a quantity in a sample rather than in a sourced population
-this is due to chance so: random error
-sample size plays an important role
Standard error describes..
The uncertainty of how well the sample estimate represents the population
Standard error estimates..
The SD of the sampling distribution
“average error that can occur whenever we take a sample from a certain size”
When does standard deviation exist?
For all statistical quantities
How many samples does it take to estimate standard error?
1
Standard error equation
For the mean:
SE = S / (square root)n
S= sample standard deviation
n= sample size
Confidence interval
They indicate a range within we are confident the true population lies
95% confidence interval
We are 95% confident the population parameter is contained within the interval sample estimate +/- 1.96 standard error
Lower confidence interval:
Sample estimate - 1.96*standard error
Upper confidence interval:
Sample estimate + 1.96*standard error
What is the sample estimate in the 95% confidence interval?
The sample mean
If the difference between the lower and upper confidence interval is small..
Precise
Aka low uncertainty regarding true population mean
If larger difference between lower and upper confidence interval..
Not very precise
Aka HIGH uncertainty regarding true population mean
