The Survey Octopus is a friendly creature who will help you to think about all the crucial issues in crafting a survey.
Presentation by Caroline Jarrett @cjforms for the 2014 Content Strategy Summit #CSSummit
How to get better results from a survey: Meet the Survey Octopus
1. How to get better
results from a survey
Caroline Jarrett
@cjforms
Content Strategy Summit 2015 #CSSummit
2. Caroline Jarrett @cjforms (CC) BY SA-4.0
I’m a forms specialist
2
Image credit: Flickr, taxrebate.org.uk
3. Caroline Jarrett @cjforms (CC) BY SA-4.0
Why do people answer questions?
3Caroline Jarrett @cjforms (CC) BY SA-4.0
4. Caroline Jarrett @cjforms (CC) BY SA-4.0
A dollar bill with a mail survey works
better than $10 (guaranteed) later
4
Caroline Jarrett @cjforms (CC) BY SA-4.0
5. Caroline Jarrett @cjforms (CC) BY SA-4.0
Response relies on
effort, reward, and trust
People will only respond if they trust
you. After that, it's a balance between
the perceived reward from filling in the
survey compared to the perceived
effort that's required. Strangely
enough, if a reward seems 'too good to
be true' that can also reduce the
response.
Diagram from Jarrett, C, and Gaffney, G (2008)
“Forms that work: Designing web forms for usability”
inspired by Dillman, D.A. (2000)
“Internet, Mail and Mixed Mode Surveys: The Tailored Design Method”
5
7. Caroline Jarrett @cjforms (CC) BY SA-4.0
7
The aim of a survey is to get a number
that helps you to make a decision
8. Caroline Jarrett @cjforms (CC) BY SA-4.0
8
To get better results from your survey,
think about the Survey Octopus
9. Caroline Jarrett @cjforms (CC) BY SA-4.0
People ask me about surveys
• “How many people do I need to ask?”
• “How many questions can I have?”
• “Please tell me whether this is a good question”
• “I think it’s best to have 5 points in my rating scale, but
my boss wants to have 7. Who is right?”
9
10. Agenda “How many people do I need to ask?”
“How many questions can I have?”
“What makes a good question?”
“How many points in a rating scale?”
10
Caroline Jarrett @cjforms (CC) BY SA-4.0
11. Caroline Jarrett @cjforms (CC) BY SA-4.0
To find out how many people to ask,
start at how many we need to answer
Caroline Jarrett @cjforms (CC) BY SA-4.0
Fieldwork:
Who answers?
11
12. Caroline Jarrett @cjforms (CC) BY SA-4.0
Whether they’ll answer depends
on effort
Questions:
What are you
asking about?
How many
questions?
12
Caroline Jarrett @cjforms (CC) BY SA-4.0
13. Caroline Jarrett @cjforms (CC) BY SA-4.0
And on the reward you’re offering
Goals:
Why are you asking?
Is helping you a reward in itself?
Are you offering any other
incentive?
13
Caroline Jarrett @cjforms (CC) BY SA-4.0
15. Caroline Jarrett @cjforms (CC) BY SA-4.0
We don’t just want answers,
we want answers from the right people
Response
Caroline Jarrett @cjforms (CC) BY SA-4.0
15
16. Caroline Jarrett @cjforms (CC) BY SA-4.0
So it matters where
we get our sample from
Sample:
the list you
sample from
Caroline Jarrett @cjforms (CC) BY SA-4.0
16
17. Caroline Jarrett @cjforms (CC) BY SA-4.0
17
And now it’s easy to work out
how many to ask
Caroline Jarrett @cjforms (CC) BY SA-4.0
Sample:
the number of
people to ask
18. Caroline Jarrett @cjforms (CC) BY SA-4.0
We thought about a lot of topics
to work that out
Goals
Sample
Questions
Fieldwork
Response
18
Caroline Jarrett @cjforms (CC) BY SA-4.0
19. Caroline Jarrett @cjforms (CC) BY SA-4.0
To get really good results, we want
useful answers from the right people
Response
Response
19
Caroline Jarrett @cjforms (CC) BY SA-4.0
20. Caroline Jarrett @cjforms (CC) BY SA-4.0
“You're using this site from outside the
UK. Where are you answering from?
We’d like answers like these
– Anchorage Alaska U.S.A.
– Cameroon
– Dubai, United Arab Emirates
Even better, extra-interesting answers like these:
– Cote d'Ivoire / France / UK.
– Canada, originally from the UK
– Philippines, would like to work in UK with my husband if possible
But definitely not like these:
– ä¸å›½
– its not important
– nnnnnn
– none of your business 20
21. Caroline Jarrett @cjforms (CC) BY SA-4.0
Insights are the numbers
that you use for decisions
Insights 21
Caroline Jarrett @cjforms (CC) BY SA-4.0
22. Caroline Jarrett @cjforms (CC) BY SA-4.0
We’ve thought about a lot of issues
Goals
Questions
Sample
Fieldwork
ResponseResponse
Insights 22
Caroline Jarrett @cjforms (CC) BY SA-4.0
23. Caroline Jarrett @cjforms (CC) BY SA-4.0
To get better results from your survey,
think about the Survey Octopus
23
24. Agenda “How many people do I need to ask?”
“How many questions can I have?”
“What sorts of questions are best?”
“How many points in a rating scale?”
24
Caroline Jarrett @cjforms (CC) BY SA-4.0
25. Caroline Jarrett @cjforms (CC) BY SA-4.0
“How many questions can I have?”
Let’s work backwards from insights
25
Which answers do you need to make the decisions?
26. Caroline Jarrett @cjforms (CC) BY SA-4.0
You need accurate answers
to the questions
26
Will your respondents give you
good answers to all of them?
27. 27
In your last five days at work, what
percentage of your work time do you
estimate that you spent using publicly-
available online services (not including
email, instant messaging, and search) to
do your work using a work computer or
other device?
%
Caroline Jarrett @cjforms (CC) BY SA-4.0
29. 29
"Phone photography" by Petar Milošević -
Own work. Licensed under CC BY-SA 3.0 via
Commons -
https://commons.wikimedia.org/wiki/File:Pho
ne_photography.jpg#/media/File:Phone_phot
ography.jpg
Modified by Caroline Jarrett
30. Caroline Jarrett @cjforms (CC) BY SA-4.0
You need questions that your
respondents can answer accurately
Which
questions
will get you
the answers
that you
can use?
30
31. Caroline Jarrett @cjforms (CC) BY SA-4.0
Now we get to fieldwork
31
How many
of those
questions
are people
willing to
answer?
32. Caroline Jarrett @cjforms (CC) BY SA-4.0
Fieldwork used to be expensive
so a survey was a rare event.
32
Image credit: http://www.census.gov/history/www/genealogy/decennial_census_records/
33. Caroline Jarrett @cjforms (CC) BY SA-4.0
1950s mindset: “Ask Everything”
33
Survey =
Big Honkin’ Survey
34. Caroline Jarrett @cjforms (CC) BY SA-4.0
2015 mindset: the Light Touch survey
• Choose ONE question
• Find ONE person
• Ask the question, face-to-face
• See if you can make ONE decision
• Improve, iterate, increase
34
36. Time for new question
Caroline Jarrett @cjforms (CC) BY SA-4.0
36
One way to
iterate,
improve,
increase
37. Big Honkin’ Survey
• Negotiate to get survey
down to 20 questions
• Ask 10,000 people
• Get 1,000 responses
• Take a week (or more) to
analyze the responses
• Have a big presentation
a month later
Light touch survey
• ‘Question of the day’
(one question)
• Ask 100 people
• Get 50 responses
• Analyze them same day
• Present same day
• Repeat 4 x 5 = 20 times
37
Caroline Jarrett @cjforms (CC) BY SA-4.0
38. Caroline Jarrett @cjforms (CC) BY SA-4.0
Do both!
Big Honkin’ Survey
• Big numbers are
impressive
• Can compare answers
from different segments
• Easier ‘sell’ to
stakeholders
Light touch survey
• Quick, useful results
• Rapidly get better
at doing surveys
• Wonderful way to test the
questions for the Big
Honkin’ Survey
38
39. Agenda “How many people do I need to ask?”
“How many questions can I have?”
“What makes a good question?”
“How many points in a rating scale?”
39
Caroline Jarrett @cjforms (CC) BY SA-4.0
40. Caroline Jarrett @cjforms
40
"Phone photography" by Petar Milošević -
Own work. Licensed under CC BY-SA 3.0 via
Commons -
https://commons.wikimedia.org/wiki/File:Pho
ne_photography.jpg#/media/File:Phone_phot
ography.jpg
Modified by Caroline Jarrett
41. Tip
Always allow for ‘other’
Design by @RickyBuchanan; t-shirt from nopitycity.com or zazzle.co.uk
41
42. 42
In your last five days at work, what
percentage of your work time do you
estimate that you spent using publicly-
available online services (not including
email, instant messaging, and search) to
do your work using a work computer or
other device?
%
Caroline Jarrett @cjforms (CC) BY SA-4.0
43. Caroline Jarrett @cjforms (CC) BY SA-4.0
Response relies on
effort, reward, and trust
People will only respond if they trust
you. After that, it's a balance between
the perceived reward from filling in the
survey compared to the perceived
effort that's required. Strangely
enough, if a reward seems 'too good to
be true' that can also reduce the
response.
Diagram from Jarrett, C, and Gaffney, G (2008)
“Forms that work: Designing web forms for usability”
inspired by Dillman, D.A. (2000)
“Internet, Mail and Mixed Mode Surveys: The Tailored Design Method”
43
44. Caroline Jarrett @cjforms (CC) BY SA-4.0
44
A good question works in three ways
Appropriate
Obvious Interesting
45. Caroline Jarrett @cjforms (CC) BY SA-4.0
45
Why did you visit our website today?
Appropriate
Obvious Interesting
46. Caroline Jarrett @cjforms (CC) BY SA-4.0
Would you recommend us to
a friend or family member?
46
In a shop,
buying a
baby carriage
In a hospital,
having a
miscarriage
Obvious Yes
Interesting Yes
Appropriate Yes Cruelly
inappropriate
47. Tip
Test your questions by
interviewing in context
Caroline Jarrett @cjforms (CC) BY SA-4.0
47
48. Agenda “How many people do I need to ask?”
“How many questions can I have?”
“What makes a good question?”
“How many points in a rating scale?”
48
Caroline Jarrett @cjforms (CC) BY SA-4.0
49. Caroline Jarrett @cjforms (CC) BY SA-4.0
Likert had several different types
of question in his response formats
Likert, Rensis. (1932). A Technique for the Measurement of Attitudes.
Archives of Psychology, 140, 1–55.
49
50. Caroline Jarrett @cjforms (CC) BY SA-4.0
You can find an academic paper to
support almost any number of points
• Krosnick and Presser refer to over 80 papers
50
Krosnick, J. A. and S. Presser (2009). Question and Questionnaire Design.
Handbook of Survey Research (2nd Edition) J. D. Wright and P. V. Marsden, Elsevier.
http://comm.stanford.edu/faculty/krosnick/docs/2010/2010 Handbook of Survey Research.pdf
51. Tip
Don’t stress too much about
the number of points in your
rating scale
51
Picture credit: Flickr - Bill Soderman (BillsoPHOTO)
52. Caroline Jarrett @cjforms (CC) BY SA-4.0
Well, OK, stress a little bit.
52
This scale is
downright
peculiar. Avoid.
54. Caroline Jarrett @cjforms (CC) BY SA-4.0
54
The aim is to get the best number you
can, within the resources you have
55. Caroline Jarrett @cjforms (CC) BY SA-4.0
The aim is to get the best number you
can, within the resources you have
What you want to ask about
The resources you have
The questions you ask
The answers you get
The answers you use
The number
Who you want to ask
The list that you sample from
The sample you ask
The ones who answer
The ones whose answers
you can use
55
56. Caroline Jarrett @cjforms (CC) BY SA-4.0
The aim is to get the best number you
can, within the resources you have
What you want
to ask about
The resources
you have
The questions
you ask
The answers
you get
The answers
you use
Who you want
to ask
The list you use
to sample from
The ones you
ask
The ones who
answer
The ones whose
answers you can use
The number
56
57. Caroline Jarrett @cjforms (CC) BY SA-4.0
57
Survey statistic
Post-survey
adjustments
Respondents
Sample
Sampling frame
Representation
Edited response
Response
Measurement
Construct
The aim is to get the best number you
can, within the resources you have
Resources
What you want
to ask about
The resources
you have
The questions
you ask
The answers
you get
The answers
you use
Who you want
to ask
The list you use
to sample from
The ones you
ask
The ones who
answer
The ones whose
answers you can use
58. 58
Total Survey Error diagram as presented in
Groves, R. M., F. J. Fowler, M. P. Couper, J. M.
Lepkowski, E. Singer and R. Tourangeau (2009).
Survey methodology. Hoboken, N.J., Wiley.
People will only respond if they trust you. After that, it's a balance between the perceived reward from filling in the survey compared to the perceived effort that's required. Strangely enough, if a reward seems 'too good to be true' that can also reduce the response.
This is a genuine invitation from local government, but the layout and images in the invitation make it look as if it's an approach from some sort of spammer or scammer.
The survey sits between 'what you want to ask', 'who you want to ask' and 'the number'
The survey octopus has 8 tentacles. We'll visit each one in the next few slides. We’ll get our survey to the people who will answer in what the survey methodologists call ‘fieldwork’ – that might be a pop-up a website, a mail survey, or face-to-face interviews.
The octopus again. This time we're looking at 'the questions we ask'.
The resources you have will help you to decide on the reward you’re offering
Prank leaves Justin Bieber facing tour of North Korea
By Daniel Emery Technology reporter, BBC News
5 July 2010
Image caption It is highly unlikely Bieber would be given permission to enter North Korea Canadian singer Justin Bieber's has become the target of a viral campaign to send him to North Korea.
A website polled users as to which country he should tour next, with no restrictions on the nations that could be voted on.
There are now almost half a million votes to send the singer to the secretive communist nation.
The contest, which ends at 0600 on 7 July, saw North Korea move from 24th to 1st place in less than two days.
Many of the votes are thought to originate from imageboard website 4chan, which has built a reputation for triggering online viral campaigns.
The octopus, with focus on "the ones whose answers you use"
The octopus, with focus on 'The list you sample from'
The octopus again; we've looked at 6 of the 8 tentacles.
The octopus, looking at two tentacles that are about response: 'The ones whose answers you can use’, and ‘the answers that you get’
We're looking at the last tentacle of the octopus, this time 'The answers that you use'
The 8 tentacles of the Survey Octopus are:
Left side:
Goals: the resources you have
Questions: the questions you ask
Response: the answers you get
Insights: the answers you use
Right side:
The list you sample from
The sample you ask
The ones who answer
The ones whose answers you use
Two questions from a survey:
'24: Do you use a Windows or Mac computer'?
'25. What is your gender'?
Photo of a Samsung (android) mobile with the same questions as previous slide. If you only have an Android mobile, how do you answer ‘do you use a Windows or Mac computer’ when the answer options are ‘Windows’, ‘Mac’ and ‘Both’?
A process starting with one person face to face, continues through 10 people by phone, gets to 100 people by email or pop-up.
It’s best to check that your question works with one person before you hassle 10 people with it. Then check it works with 10 people before you send it to 100. Once you’ve tried it on 100 people, you might be more interested in a new question than getting more answers on this question
Photo of a Samsung (android) mobile with the same questions as previous slide. There aren’t any options for ‘other’, but some people don’t use a Windows or a Mac. Most people identify as ‘Male’ or ‘Female’ but some people don’t.
A model wears a t-shirt with Gender: 'Male' (crossed out), 'Female' (crossed out) and 'Other' (added and ticked'.
People will only respond if they trust you. After that, it's a balance between the perceived reward from filling in the survey compared to the perceived effort that's required. Strangely enough, if a reward seems 'too good to be true' that can also reduce the response.
Obvious questions require no effort to answer. Interesting questions are rewarding to answer. Appropriate questions are ones that inspire trust in the respondent.
People come to the web with their own questions, so they’re likely to know why they are on your website. It’s a relatively interesting question to answer, and it’s appropriate to ask visitors why they’re visiting.
This is a more conventional way of looking at the octopus tentacles
If we just look at the issues (no tentacles) we get this slide
This slide translates the issues into the technical terms used by survey methodologists