General Stats Notes

Also wrong answer notes

Write down the topic to trigger the memory

Precision is increased as sample size increases. Inversely proportional to square root of n.

Bias would not change due to sample size if sample estimator is an unbiased estimator of population variable.

In non-probability sampling, the probability of sampling a unit is not equal for each unit so statistical methods can't be applied

Randomisation allows a causal relationship to be concluded from the data

USE THE FORMULA BOOK

Bias is systematic error

Proportional allocation is proportional to stratum size

Neyman allocation is proportional to stratum standard deviation size

If stratum standard deviation varies a lot, Neyman allocation is advantageous

S = population standard deviation. s = sample standard deviation.

For standard error of total, multiply N to the standard error.

In stratified sampling, the variation between strata is not counted for.

Just use the std deviation given, doesn't matter what, for Neyman allocation. No need to adjust.

Ratio estimator there is some error/bias in smaller samples.

X, Y and Z in randomisation, causation and experiments

Write the formula down for 1 mark

Write down the calculation even if putting it in the calculator

When they ask precision, they want variance

Get used to cluster sampling. Define what each of the notatin mean in each question.

As long as you clearly show the steps, the question will be given marks even if there is little error.

standard deviation measures variability of random variable

standard error is the standard deviation of the distribution of the sampling distribution of estimator

sampling error is due to random variability during the sampling process

non-sampling error is due to failure of the sampling scheme

If there is proportionality in the data, ratio estimators are preferred

Say what the symbols mean

Use tables when calculating to organise stuff

Cluster sampling and Ratio Regression and Estimation often overlap