Client Portal Login

Secure login

Quick sample size calculator below. To access additional statistical information and the Insights Academy, you will need to log in to the client portal.

Quick sample size calculator.

To use the calculator, enter the following...

Enter the desired margin of error (accuracy) for your study. The Auditor General generally recommends a 5% margin of error for state government research studies.

Select the desired confidence level (typically 95%) from the drop-down menu. A confidence level of 95% is acceptable for local government research studies.

Enter the population size. The population for the study is determined by the study. For example, the population for a community-wide study may be all residents aged 18 and over living in the municipality. Whereas, the population for the LGA's libraries would be all library users.

Enter the response distribution. This value is the expected distribution of the results. If you do not know, use 50% as this provides for a larger sample size to account for uncertainty.

Click on the Submit button.

Refer below for an explanation of each input.

Sample Size Calculator for Local Government

Common choice is 5%. Lower = larger sample.
Higher confidence = larger sample size.
Number of people in your target group (use 20,000 if unknown – sample size barely changes above this).
Use 50% unless you know the split will be very skewed (gives the most conservative/largest sample).

Some statistical information about the inputs

Understanding the "Margin of Error"

The Margin of Error is simply the "wiggle room" in your survey results. Since you can’t talk to everyone in the community, this number tells you how closely your survey’s results likely reflect the results if you had surveyed everyone in the community. Think of it like this:

  • The "Safety Gap": If your survey says 60% of people are satisfied with a service with a 5% margin of error, the real answer is likely somewhere between 55% and 65%.
  • The Landslide vs. The Toss-Up: If your results show a huge win (like 90% vs. 10%), you can afford a bit more error because the winner is obvious anyway. But if the results are a "toss-up" (like 51% vs. 49%), even a tiny error could change the outcome. In close races, you need your maths to be much more precise.
  • The Power of Numbers: To get a lower margin of error (a smaller "wiggle room"), you have to survey more people. The bigger your group, the more certain you can be.

Understanding the "Confidence Level"

The Confidence Level is basically the "certainty rating" of your survey. While the Margin of Error tells you how much "wiggle room" you have. The Confidence Level tells you how sure you can be that the real answer actually falls within that margin of error. Think of it like this:

  • The "Reliability Score": Most researchers use a 95% Confidence Level. This means if you ran the exact same survey 100 times with 100 different random groups, your results would be "correct" (within your wiggle room) 95 out of those 100 times.
  • Casting a Net: Imagine throwing a net to catch the "true" answer. A higher confidence level is like using a stronger, more reliable net. It gives you more peace of mind that you haven't missed the truth due to a random stroke of bad luck with who you surveyed.
  • The Trade-off: To be more certain (moving from 95% to 99% confidence), you either have to accept a much larger "wiggle room" or—more commonly—you have to survey significantly more people. The more "sure" you want to be, the more data you need to back it up. Typically, research for government bodies uses a 95% confidence level.


Understanding the "Population"

The Population is simply the total group of people you are trying to understand. In local government research, this isn't just "everyone" - it’s the specific group that your project or policy will actually impact. Think of it like this:

  • The "Target Zone": If you are planning a new playground, your "Population" might be all households with children under 12 years in the local area. If you’re changing parking rules, it’s the local business owners and commuters. You have to define the boundaries before you start.
  • The "Whole Pie": The population is the "whole pie," while your survey group (the sample) is just a "slice." You study the slice to make an educated guess about how the entire pie looks and tastes.
  • The Power of Accuracy: If you define your population incorrectly - like asking only homeowners about a project that affects renters - your results won't reflect the real-world community. Getting the population right ensures your data actually represents the people who matter most to the decision.


Understanding the "Response Distribution"

The Response Distribution is simply how "split" the opinions are in your community. It’s a measure of whether everyone agrees or if there is a massive divide on a local issue. Think of it like this:

  • The "Landslide" vs. "The Tug-of-War": If 90% of residents want a new library and 10% don't, that’s an uneven distribution (a landslide). But if 50% want a bike lane and 50% want more parking, that is a perfectly even distribution (a tug-of-war).
  • The "Wiggle Room" Factor: When a community is evenly split (50/50), your margin of error matters much more. A tiny mistake in your maths could mean you've crowned the wrong "winner". When the community is largely in agreement (90/10), you can tolerate a bit more error because the overall sentiment is crystal clear.
  • The Safety Play: Because local governments often don't know how split the public will be before they start a survey, local government researchers usually assume a 50/50 split (50%) (the "worst-case scenario"). This ensures you survey enough people to be safe, no matter how divided the final results turn out to be. So, if you're unsure, go with a response distribution of 50%!