## ​​Survey Analysis – The One Thing You Need to Know!

These days pretty much everybody has to ​run a survey occasionally.

Whether you work for a business that collects customer data, a supplier that wants to know which way the prevailing winds are blowing or a healthcare provider with patient data, surveys are everywhere.

But if you've never run a survey before, how do you know what to do?

It's just a matter of coming up with a few questions isn't it?

Well, it's a little bit more involved than that.

For a start, there are different types of survey, all designed for different purposes.

And then when you get your data, how do you analyse them?

Gahhhhh!!!!

It's starting to get complicated.

Not to worry, though, because I've got just the thing for you - I'm going to show you all the different types of survey that you'll encounter, and I'm going to ​teach you the one thing that they all have in common!

### Choices, Choices...

And now I'm going to give you a choice...

Ooh, I like choices!

It's up to you...

And just so you know, this video lesson is actually one of the lessons in The Hive from our exclusive video course How to Analyse Categorical Survey Data in Excel and in R, which is Open Access to start with (you don't need to register).

If you want to continue learning, you can start from the beginning here:

### ​What is a Survey?

It might seem like a silly question to ask, but until we define what a survey is, we really can't go any further.

So here it is:

​What is a Survey?

​​A survey is a research method used for collecting data to gain information and insights on a particular topic of interest

In other words, we:

• observe the real world
• ​to gather information,
• ​convert the information into data,
• analyse the data
• and interpret the results
• to ​understand what the information means

And once we understand what the information means we can use it to predict the future to make gains in some way.

Here, the word 'gain' means:

• ​Financial profit
• ​Scientific discovery
• ​Better treatment of an illness

#### Types of Survey

​It might surprise you to know that there are only 2 types of survey!

These 2 survey types are:

• ​Cross-sectional study
• ​Longitudinal study
##### Cross-Sectional Surveys

We use cross-sectional surveys (aka cross-sectional studies) to study a topic of interest at a specific point in time, and they are used to get a quick and cheap summary of our data.

For example, we can ask the question 'Do you have lung cancer?'.

We can then analyse the data gathered and get an idea of what's happening out in the world, like this (note - not real data):

#### Pin it for later

Need to save this for later?

Pin it to your favourite board  and you can get back to it when you're ready.

##### Longitudinal Surveys

On the other hand, a longitudinal survey (aka longitudinal study) is used to study a topic of interest over a period of time, and we use these studies to examine cause and effect.

For example, as well as asking the question 'Do you have lung cancer?', we can also ask 'For how long have you smoked?'.

This has the effect of adding the passage of time into our data, which is what's needed to examine cause and effect, like this:

​From these data we can discern trends, and from these trends make predictions about what might happen in the future, like this graph that shows that the longer you have smoked, the higher risk you have of getting lung cancer (note - not real data):

#### ​​Types of Longitudinal Survey

Just to complicate things a little, there are also 2 types of longitudinal survey:

• ​Case control ​study
• ​Cohort study
##### Case Control Surveys

In a case control survey (aka case control study) we compare the data where the outcome is present (the cases) with the data where the outcome is not present (the controls), where the outcome that we seek might be disease (healthcare), sales (retail) or opinion (politics), like this:

We use case-control studies to try to figure out the cause when we already know the effect (the outcome).

Basically, we have a number of cases where the outcome is present and some other cases where the outcome is not present. We follow each of these cases over time, gathering data where the cause is present (at least, what we think the cause might be - this is our hypothesis) and where the cause is not present.

We can then compare the data like this:

##### Cohort Surveys

​On the other hand, in a cohort survey (aka ​cohort study) we compare the data where the ​cause is present (the cases) with the data where the ​cause is not present (the controls), where the ​cause that we seek might be ​smoking (healthcare), ​product placement (retail) or advertisement (​marketing), like this:

We use ​cohort studies to try to figure out the ​effect (outcome) when we already know the ​cause.

Basically, we have a number of cases where the ​cause is present and ​other cases where the ​​cause is not present. We follow up on these cases over time, gathering data where the ​outcome is present ​and where the ​​outcome is not present.

We can then compare the data like this:

#### ​​Types of ​Cohort ​Survey

​And as if it wasn't bad enough that there are 2 types of londitudinal surveys, there are also 2 types of cohort surveys:

• ​Retrospective study
• ​Prospective study
##### Retrospective Surveys

​A retrospective survey (aka retrospective study) is where you check back into existing records to investigate cause and effect - in other words, you're looking backwards into the past to try to get answers.

##### ​Prospective Surveys

A prospective survey (aka prospective study) is where you investigate cause and effect over time - you're looking forwards into the future to get answers.

How to Analyse a Survey - ​The One Thing You Need to Know! (video) #​survey #​surveyanalysis @chi2innovations @eelrekab

#### ​​Survey Type Summary

At this point you might be wondering how all this fits together. Well, maybe this graphic will help:

​There are 2 types of survey - cross-sectional (for a quick summary) and longitudinal (for ​more in-depth cause & effect studies).

Of the longitudinal surveys, there are case control surveys (to investigate ​causes) and cohort surveys (to investigate ​effects).

And of the cohort surveys, there are retrospective surveys (looking back over existing data) and prospective surveys (examining future data).

#### ​​The One Thing All Analysts Need to Know

There may be different types of survey, but the one thing that you need to know as a data analyst is that whichever type of survey you're conducting...

...the way you analyse them all is exactly the same:

• ​Same types of data
• Same descriptive statistics
• Same data analysis techniques
• Same statistical tests

​I'm not going to go through exactly what those analyses are ​here - that's what the course is for - but it's useful to know that if you're running a study, whatever type of study it is, you only need to know one methodology for analysing the data, not several.

That's a HUGE relief!

### ​How to Analyse a Survey - A Summary

When it all comes down to it, surveys are much simpler ​you might think.

Sure, they're a little more complicated than just putting together a random set of questions, but if you know the different survey types you'll know which type is appropriate for your study.

And when you know that, it gives you a much better understanding of which questions to ask, and in which order.

And finally, when you've collected your data you know that there are no 'special' ways to analyse each type - here, it's one size fits all!

I appreciate that surveys aren't the sexiest beasts on the planet, so if you've got this far without running away, well done!

Give yourself a boost by clicking our 'Awesome Button'!

### ​How to Analyse Categorical Survey Data in Excel and in R

I hope you enjoyed this video lesson.

It's actually one of the lessons in The Hive from our exclusive video course How to Analyse Categorical Survey Data in Excel and in R ​, where I'll teach you everything you need to know about analysing categorical survey data.

This course is Open Access to start with (you don't need to register), and after you've gone so far through you'll need a Free Plan to continue.

As part of the course you'll get all the data and resources you need to practice with.

If you want to continue learning, you can start from the beginning here:

### ​How to Analyse Categorical Survey Data in Excel and in R

I look forward to seeing you on the inside!