|The When and How Of Learning To Code|
|Written by Janet Swift|
|Wednesday, 10 April 2019|
The results of the 2019 Stack Overflow Developer Survey are out and, as usual, provide lots of useful insights into the global developer landscape. Here we look at when devs stated coding and their, ongoing, education.
This is the ninth edition of the Stack Overflow Developer Survey and nearly 90,000 took part by completing a 20-minute online questionnaire. This is around 10 percent fewer than last year which had over 100,000 respondents
As usual there's a lot of information to digest, but one finding, which we were alerted to in the blog post link:
"all of the developers are above average"
tends to call the survey methodology into question - or does it?
This reference is to a question in which respondents were asked to evaluate their own competence, for the specific work they do and years of experience they have. Almost 70% of respondents say they are above average while less than 10% think they are below average.
I have to concur with Stack Overflow's comment:
This is statistically unlikely with a sample of over 70,000 developers who answered this question, to put it mildly.
After all if this was a representative sample of developers, surely we should see a normal distribution of competence with average being the most common response. Instead the results look like this:
(click in chart to enlarge)
It perhaps make sense that only above average developers are attracted to Stack Overflow. And perhaps that developers who consider themselves below average aren't interested in completing surveys. But it does mean that we should consider that all survey results are skewed towards developers who consider themselves of above average competence.
Stack Overflow looked further into competence, discovering that men are much more likely to say they are above average than "gender minorities in tech". i.e women, accounting for less than 8% of respondents, or in a third category, "non-binary, genderqueer, or gender non-conforming" also referred to as "transgender", and used by just over 1% of respondents.
An analysis looks at the effect of experience and gender together:
The report notes:
New developers are much less likely to evaluate themselves as above average for their experience, and this effect does not flatten out until about 10 years of experience. We see evidence here among the most junior developers for impostor syndrome, pervasive patterns of self-doubt, insecurity, and fear of being exposed as a fraud. Among our respondents, men grew more confident much more quickly than gender minorities.
This time around Stack Overflow's questionnaire included a question about the age at which respondents wrote their first line of code, something that was covered in last year's survey of developers from HackerRank, see Never Too Early To Code According to HackerRank. The results confirmed what was found by HackerRank - that today's programmers started early:
Overall, over half of respondents had written code by the time they were sixteen, but Stack Overflow found differences by gender, with women writing their first code at an average age of 16.9 years, later than men at 15.3 years. On the other hand non-binary respondents started coding earlier at 14.3 years. There are also marked geographical differences. In the United States, which accounted for the largest proportion of respondents, the average age was 15.2 years, but India, second largest, had the oldest average 17.0 years. The youngest average, 14.2 years was recorded from Australia, followed by United Kingdom 14.3 years. The report notes:
If we control for developer age today, we see an even more dramatic disparity between India (where developers who are older today started coding later in life) and the UK (where developers who are older today started coding earlier in life).
In terms of educational attainment, and among professional developers who accounted for 83% of all respondents,three quarters of respondents have the equivalent of a bachelor's degree or higher:
Of those with a university education almost two thirds had majored in Computer Science.
Another finding was that:
Developers are lifelong learners; almost 90% of all developers say they have taught themselves a new language, framework, or tool outside of their formal education. Among professional developers, about 60% say they took an online course like a MOOC (up significantly from last year), and about a quarter have participated in a hackathon.
Enthusiasm for learning new skills is also reflected in the high proportion, 80%, of respondents who claim they code as a hobby. So while programming is what developers do to earn a living, it is also what they do for fun and for self-improvement.
or email your comment to: email@example.com
|Last Updated ( Thursday, 11 April 2019 )|