Integration Testing Ownership Poll result analysis

Half month ago I did a poll about integration testing ownership. I have asked this question on three different services. Twiter, few slacks and Polish Testing group on Facebook.
Integration Testing mind mapI am still testing mind map zen sorry for a watermark.
The question was: „Who is responsible for integration test automation in your company/team?

Why this question?

But before We go into analysis, a little background why this question.
I have found article „Test devs aren’t testers” I had a lot of issues with author arguments. To the extent that I had prepared three blog post with my contr argument. But I wasn’t satisfied with the quality of my work, so the project was scrapped.  
This Survey was to collect data for one of the arguments:
„Many times test automation projects are started in QA teams, or departments by testers (either organically or managerial decree) and therefore are owned by testers.”
I had a problem with this generalisation cause unit tests (Which ARE TEST automation) are almost always owned by the Developers. But I am willing to accept that some people don’t count them as test automation.
Testers usually own UI and E2E. I haven’t done any poll on this subject, but the simple googling query was showing me enough results to treat it as a good theory.
Ok so in case if the author didn’t treat unit test as test automation I wanted to take a closer look at integration testing.
Unfortunately, Google search didn’t bring me any reliable information. So I decided to do this poll.

Integration Testing Poll result

There were 147 participants, and the substantial majority (53%) said that both Developers and testers are responsible.
[expand](Description of the poll was explaining that term Testers will also include SDET and similar roles)[/expand]
only testers and only developers were oscillating around 20%
So from the perspective of my argument. It means that in the vast majority of answers responsibility for automation was either shared all in the hands of developers.
Making author point Mute.

Error ins methodology:

I wish I would be able to say that. But there are few the problems:
According to this report, there are 22 million software developers worldwide.  Word tester is not mentioned at all in the report so it is safe to assume we are not included in that number. There is no similar report about testing. I think it is safe to assume that with testers it will be at least 30. 
So if I want to achieve Confidence Level 95% and Margin of Error 5% I would need around 375 participants. Unfortunately, I had only 147.
That number is good for CL 90% and MoE 10% which is from statistics perspective poor base for generalisations.
It still shows that issue is not clear as Author of that post would like to present it. But I can’t say for sure he is wrong.
I made also some mistakes preparing the poll.
I was mostly asking testers. I haven’t put my poll on any developer forums (I did try on two, but when I was asking for permission from admins I heard they have flat out ban on polls). That why I am afraid I may have fallen into Confirmation Bias.
I was asking for Responsibility, not Ownership. It is a minor difference, but in surveys, you have to be careful and specific, so all participants were on the same page.
The first poll I did was on the facebook group and well it has few more issues
It included 'other’ option – which didn’t bring any useful info. Since it was mostly about the responsibility, I have rolled it into „both.”
Another I forgot to blog adding new options, so few people added their own. I have rolled them into appropriate options (none and both)
Polls on the Straw poll (Sent on few slacks) and twitter went much smoother. The only problem is Twitter poll had only seven answers.


For me, it was a useful exercise
But Even though it did bring some info it was not enough to bring the full answer.
If I will make polls in future
I need to make more precise questions and answers.
Plus I need to tackle wider audience.
The straw poll is still open:
If you want you can still vote. I will check it and End of Aprill who knows maybe I will have enough votes to make the more solid statement?
If so I will update this article.
And what do you think? Do you find this poll useful or you would do it differently?
If you want to read more topics connected to science and testing you can check it here.

Dodaj komentarz