STATS FLASHCARDS

Linear interpolation for grouped data

lower class boundary + ((position - cumulative frequency before the class) / frequency in the class) x class width.

Use n/2 for median, n/4 for lower quartile, 3n/4 for upper quartile, and 0.9n for the 90th percentile. First use cumulative frequency to find which class the position is in.

Normal approximation to binomial

If X follows a binomial distribution and n is large, you can sometimes approximate it using a normal distribution.

Use:

mean = np
variance = np(1 - p)

Then use the normal distribution with a continuity correction.

Continuity correction

Use continuity correction when approximating a discrete binomial distribution with a continuous normal distribution.

Examples:

P(X <= 20) becomes P(Y < 20.5)
P(X < 20) becomes P(Y < 19.5)
P(X >= 20) becomes P(Y > 19.5)
P(X > 20) becomes P(Y > 20.5)

PMCC hypothesis test

For correlation tests, use:

H0: rho = 0

Then H1 depends on the question:

positive correlation means rho > 0
negative correlation means rho < 0
correlation/different from zero means rho not equal to 0

Then compare the calculated r value to the critical value.

Sampling frame vs sampling unit

A sampling frame is the list of all members of the population that the sample is chosen from.

A sampling unit is one individual member/item selected from that frame.

Example: if a school samples students from a register, the register is the sampling frame and one student is the sampling unit