What does 30 signify for n>30 for NLT?
For CLT to apply, you mentioned how the sample size needs to be 30.
Does this mean that the number of observations in each sample (since we're taking multiple in this sampling distribution) needs to be > 30 or the number of samples taken from the population needs to be > 30.
I believe it means you need at least 30 sample means. In the example shown of the prices, each price on the list is a sample mean; in other words, this list should have at least 30. To obtain each of these sample means, usually you need at least 100 prices/"observations".
Goes like this: Get at least 30 groups of car prices. Each group containing at least 100 prices. Calculate the mean of each group -> Now you have 30 sample means (e.g. the list of prices in the video example). Calculate the mean of these means -> this is your sampling mean that approximates the population mean.
[G1_1,...,G1_100], [G2_1,...G2_100], ..., [G30_1,...G30_100] -> 30 groups of 100 prices (e.g. sampled 100 car prices in/each 30 cities of New York )
[G1_mean, G2_mean, ..., G30_mean] -> (e.g. mean price in each city)
[SamplingMean] -> (e.g. approximates mean price in New York)
