# Analyzing Responses from Post Event Surveys

While it’s great to see that your delegates enjoyed your event, you now need to translate the information that was gathered into something with substance. The data gathered from your surveys is very valuable, and this blog will help you understand how to analyse the responses from your post event survey.

## Formatting the Data

There are many ways to record attendee feedback. Whether you use paper forms, online surveys or event apps you’ll most likely be left with a large pile of raw data.

In my experience, organizations record or export the data into Excel as it’s a piece of software that most people are familiar with. While you can pay for custom software or for a survey website to analyse it for you, sometimes it’s better to be able to get into the data yourself and look at it many different ways.

## Converting Conference Survey Responses

Looking at a large batch of raw data can be overwhelming, but like learning any new language you need to start simple.

The first thing you will want to look at is how the survey converts the responses. Some online surveys convert responses to numbers, so what was a yes or no question becomes ones and zeros. This means it’s very important to keep consistent during the survey creation so that you don’t have ten questions where one equals yes and another question where one equals no.

## Analysing the Conference Survey Data Using Means, Medians and Modes

Finding the “average” is a quick and simple method of comparing responses but I’m sure many of you remember from math at school that there is more than one way to find the average and some data types are best suited to different kinds of averages.

- The
**mean**is the average you are probably most familiar with; take all the values and add them together, then divide them by the number of responses.

The mean is great because all of your data is used to find the answer. However the disadvantage of the mean is that outliers, such as very large or very small answers, can distort the final answer. - The
**median**is when you arrange all of your data in order and find the value in the middle of the data.

The advantage of using the median is that outliers won’t affect the answer anywhere near as strongly as it would using a mean, but arranging the data and finding the median can be a pain if your data isn’t already recorded in Excel or similar software. - The
**mode**is the response that appears most frequently and can be a great way to measure non-numerical data such as “why did you attend the conference?”

Unfortunately the mode doesn’t always work out. You could have a set of responses where nobody gave the same answer as another person, so no mode would exist or alternatively you could end up with 3 people saying “for networking” and 3 people saying “for the education” and so both responses would be the mode.

## Determining the Most Accurate Average to Use

So it’s obvious you need to use the right average for the right situation. Below are some guidelines for when to use each method.

For data where there are two or more categories that can’t be organised into an order e.g. what province do you live in? Use the mode.

For data where you can organise the categories into an order e.g. please rate on a scale of 1-5 the speaker’s presentation. Use the median.

For data where there is no distinct categories and a large range of possible answers e.g. what is your age? It depends on what statisticians call “skewness” or a skewing of the data.

For example, if your conference has 50 attendees all in the age 30-50 range but one attendee who was 80, the 80 year old would be a skew in the data. If your data is skewed you’d use the median, if it’s not skewed then use the mean. Please note this is a basic example of the *very* complicated topic of statistical skewness.

## Comparisons and Comments

Now that you have your data sorted, organised and averaged out, what are you going to do with that data? It’s all fine and good seeing that speaker Bob got an average of 5 for his talk but what does that mean? Is that good or bad?

In order for your data to mean anything you need to be able to compare it to something. Whether you are comparing speaker sessions against each other or this year’s conference to last year’s being able to compare and contrast your results is important.

So if Bob got an average of 5 for his talk and Dan got an average of 7 for his talk, does that mean that Dan is the better presenter or speaker? Not necessarily.

This is where creating a good post event survey is really important, you MUST give your delegates places to be able to comment as to why they scored the way that they did. It could be that people scored Bob low because of AV issues, maybe his session description didn’t match the delegate expectations or his slides were too busy and hard to read.

So while it’s great to have the data to find out where the issues are, it’s really the attendee comments that help you to improve the event for next year.

## Utilizing Graphs and Graphics in Your Final Conference Report

A well done graph is a thing of beauty; clear, concise comparisons in an easy to understand visual format. Oftentimes it’s easier to show a client survey responses in a graph than in large tables of numbers. Below are my recommended graph types for different comparison types.

- Are you trying to compare values?
- Column, Bar, Line, Scatter or Bullet

- Are you trying to show how a whole is broken down into its individual parts?
- Stacked Bar, Stacked Column or Pie

- Are you trying to show data distribution?
- Scatter Plot, Line, Bar or Column

- Are you trying to analyse trends in your data set?
- Line or Column

So whether you’re finding out where your delegates are from or how well the food went down, it’s important to use the correct method of analysis for the data, to look at feedback comments, and to present all the information in a clear and concise manner.