# HLTB15H3 Lecture Notes - Lecture 7: Sampling Frame, Population Study, Sample Size Determination

Cue Column:

Note-Taking Area:

Adv: saves $ & HR & time

Loses accuracy

□

Based on assumption

□

Disadv: only find about sample, and estimate for population

Small group from population

○

Keep going while you keep finding

□

Basically continue sampling as long as you make new

findings

□

Based on subjective judgments

When no new findings, it's called saturation point

□

Analysis done on saturation point

In qualitative research

○

Concept of sampling

•

Group of individuals/units etc.

Shares same characteristics

Population / study sample (N)

○

Subset of population

Sample

○

# of individuals used to draw info

Representatives of population

Sample size (n)

○

Random, non-random, mixed designs

Sampling design/strategy

○

Unit is an individual

Unit of analysis

Sampling unit/sampling element

○

Every indi/unit in study sample

Where can sample be drawn

ID individuals

Sampling frame

○

Statistics ex mean, etc. from sample

Sample stats

○

# inferred from stats

Estimate coming from sample stats

Predict certain population mean

Population parameters/ population mean

○

Terminology

•

*true population mean - the true mean of entire

population

□

The difference is in relation to sample units

□

There is going to be diff between sample stats and population

mean

As sample size increases, more accuracy between sample

stats and true pop mean

Ex. Really young and really old in class, it's going to be

so off

□

The greater diff in variable, greater diff between sample stats

and pop mean

Principles of sampling

○

Need to be precise

It still can occur though

□

Sampling is non-random

Doesn't cover accurately

◊

Sampling frame

Can occur if:

□

Minimize bias

Aims

○

Guiding principles

•

Undertaking research - sampling methods

HLTB15 Lecture 7

Thursday, March 14, 2013

12:16 PM

HLTB15 Page 1

## Document Summary

Disadv: only find about sample, and estimate for population. Basically continue sampling as long as you make new findings. When no new findings, it"s called saturation point. There is going to be diff between sample stats and population mean. *true population mean - the true mean of entire population. The difference is in relation to sample units. As sample size increases, more accuracy between sample stats and true pop mean. The greater diff in variable, greater diff between sample stats and pop mean. Really young and really old in class, it"s going to be so off. Have equal and independent chance of being chosen. Randomly select certain column and row or something. Sampling error can be reduced if heterogeneity can be reduced. Proportionate: # from each stratum is proportion to total pop. More heterogeneity, more respondents needed a lot of diff, need large sample size.