English Deutsch Français Italiano Español Português 繁體中文 Bahasa Indonesia Tiếng Việt ภาษาไทย
All categories

Where I work we do a lot of surveys and we have software that tells us if a sample size is statistically significant. But nobody can explain the parameters that affect this (e.g. does the size of the population matter?). Would really appreciate a simple explanantion. Thanks!

2006-06-15 20:25:56 · 6 answers · asked by chris1089 2 in Science & Mathematics Mathematics

6 answers

This document will teach you how to calculate optimal sample size.

http://www.osra.org/itlpj/bartlettkotrlikhiggins.pdf

2006-06-15 20:47:20 · answer #1 · answered by uselessadvice 4 · 0 0

Statistically Significant Sample

2016-10-02 03:08:32 · answer #2 · answered by ? 4 · 0 0

Error=z*S/n^.5
S=Sample Standard Deviation
Error=IS the error from the average (Actual)
Z=Normal Distribution Value At (Alfa=.05)

so if ur accepeted error<.5....of something that has an average of 100
then n=Z^2*S^2/Err^2

where do u get S?
Have to do a pilot study...Compute S then do the sample size
Using High confidence makes the sample size need to increase...which is the Z Value (Can Get from Stats Table or excel)

if U do not want to do Pilot Study
Use this table for any thing if u impose he average on Surevy

average St devaiation
30 5.48
40 6.32
50 7.07
60 7.75
70 8.37
80 8.94
90 9.49
100 10.00
110 10.49
120 10.95
130 11.40
140 11.83
150 12.25
160 12.65
170 13.04
180 13.42
190 13.78
200 14.14
210 14.49
220 14.83
230 15.17
240 15.49
250 15.81
260 16.12
270 16.43
280 16.73
290 17.03
300 17.32

2006-06-15 21:19:15 · answer #3 · answered by Khalid 2 · 0 0

The sample size is statistically significant when it is large enough to accurately represent the population. For example, would you use a sample of 1 person to represent a population of 100 people? Statistically significant solely means that the sample size is big enough to reflect on the general poplace.

2006-06-15 20:30:30 · answer #4 · answered by King of Kings 2 · 0 2

I do believe it is 5% - 10% of the total population if the number is greater than 1000. Under that, I think it is 30% - 50%, as long as it is achievable.

2006-06-15 20:37:13 · answer #5 · answered by Anonymous · 0 1

It depends on how much error are you going to accept. The smalletr the error the bigger your sample. So on it depends on the size of the population. It's different to make an analysis on the consumer habbits of al US than making for your school.

2006-06-15 20:28:52 · answer #6 · answered by Rocko 3 · 0 1

fedest.com, questions and answers