What impacts developers’ job preferences?
It’s not all about the money! — Stack Overflow Survey 2020 results
Nowadays, non-compensation factors arise as the key to hire a successful worker. Companies are able to offer enormous salaries, however, after breaking some level of the salary, it’s not the money that often is a key to accept this or that job.
If it’s not the money, then what? During the Stack Overflow Survey 2020 respondents got the following question:
Imagine that you are deciding between two job offers with the same compensation, benefits, and location. Of the following factors, which 3 are MOST important to you?
The results were as follows:
We can clearly see that for over 1/2 of respondents, technology is one of the most important factors when deciding on taking a job offer. It shouldn’t surprise since nobody would like to work in technology that is not interested in or is not familiar with it.
The second and third places are the Office environment or company culture and Flexible schedule. A similar number of respondents have chosen also Opportunities for professional development. These four emerged as dominant among others.
On the lower side of the list, we have just about 7% of respondents that picked the Diversity of the company or organization as the most important factor when taking a job offer.
Finding patterns
Recall, that the respondents could mark up to 3 factors, so it’s worth checking if some of them tend to be chosen together. For instance — among the respondents who have chosen the “Flex time or a flexible schedule”, 37% of them chose also the “Remote work options”.
Actually, we should expect this, because both questions are related to each other. Both are related to the time of working. It wouldn’t be too risky to claim that the group of developers choosing these options could consist of persons that require convenience and comfort regarding the work time organization.
Another interesting observation — developers who chose “Family friendliness” or “Diversity…” are less often care about technological stuff:
They are more tended to choose soft factors, strongly connected with relations with the people and respect, rather than technology.
And there are many more such patterns! Let’s identify them using the following heatmap — don’t be scared! I’ll explain below how to interpret this :):
For those who picked an option from the y-axis, it is more/less likely they would chose also a given option from the x-axis.
For instance: those who chose Office environment or company culture, would more likely choose also Diversity as an important factor. The stronger the color is, the greater is the deviation from the average (thanks to standardization the average is equal to 0 and the standard deviation 1).
We can easily confirm observations we noticed previously — about those who pick Remote work options and Flextime or those who picked Diversity are less likely to choose languages, frameworks, etc.
An even better way to visualize these relations is to use a simple multi directed graph:
The relations are limited to only those over the standard deviation. There is a couple of insights here:
- Individuals that pick “Family friendliness” tend to care less about career-related factors. They are less likely to pick at least one of the following: “Specific department”, “Tech stuff”, “Professional development”, and “Industry ”.
- As a contrary to above, those who picked “Work impact”, are more likely to pick also “Industry” or “Professional development”.
- Those who picked “Diversity ” tend to care more about company culture and office environment, while less about “Organization finance” and “Tech stuff”. Actually, that's what we expected since diversity and inclusion are the important values of many companies.
- Last but not least, we can observe some similarities between “Flex time” and “Remote work options”. The group of developers choosing these options could consist of persons that require convenience and comfort regarding the work time organization. Respondents choosing these two are overlapping much, that’s why “Family friendliness” occurs in both as the factor they tend to chose more likely, while they both less often care about their impact on the reality, which confirms the hypothesis ‘Convenience first’.
Last, but not least, I attempted to identify factors that impact individuals’ choices. Attempts to model respondents’ choices resulted in multiple quite poor models. Based on modeling results we can identify features that impact models output so they could be related to respondents’ choices. For instance, in the “Diversity“ case, the most important factors were: gender, sexuality, and compensation.
After further investigations, there was found a weak correlation between “Diversity” and being a hetero or a man. The coefficient doesn’t tell if it’s a positive correlation or negative.
Conclusions
- Languages, frameworks, and technologies are the most common non-salary related factor impact job preferences. Quite important are also office environment and company culture, flexible schedule, and opportunities for professional development.
- There are some patterns in respondent choices. Based on the respondents' choices it is possible to identify at least 4 groups of differing in their preferences.
- Features that impact respondents' choices are various. In the “Diversity“ case, for modeling the most important factors were: gender, sexuality, and compensation the respondent.
The project source codes: https://github.com/kmwolowiec/job-preferences
