Estimating Sampling Errors
Increasing the sample size will reduce the sampling error and give increased
precision in survey estimates. So, when planning your survey, think about how
‘accurate’ you want the results to be.
- If you want all estimates (results) to be
within about ± 2%, then you will need a sample
size of 2,000.
- If you can afford a survey with 5,000 respondents, then your maximum standard
error is likely to be ± 1.4%.
The following is a useful table for estimating standard errors for various
percentages found in your survey. For example, if you get a survey value of 30%
with a sample of 2,000, then the error range will be 2.0% - and the 95%
confidence interval will be 30 ± 2% = 28% to 32%.For simple random sampling:
Range of Error * (±) for 95% Confidence and
Interval
| 100 |
4.4 |
6.0 |
8.0 |
9.2 |
10.0 |
| 200 |
3.1 |
4.2 |
5.7 |
6.5 |
7.1 |
| 500 |
1.9 |
2.7 |
3.6 |
4.1 |
4.5 |
| 1000 |
1.4 |
1.9 |
2.5 |
2.9 |
3.2 |
| 2000 |
1.0 |
1.3 |
1.8 |
2.0 |
2.2 |
| 5000 |
0.6 |
0.8 |
1.1 |
1.3 |
1.4 |
| 10000 |
0.4 |
0.6 |
0.8 |
0.9 |
1.0 |
|
|
* Note – the range of
error here is twice the standard error (1.96 x SE). |
