Calculating probability of being out of work
June 26, 2019 11:34 AM   Subscribe

I'm trying to work out the probability that someone is either out of work or in a less-than-ideal working situation (i.e. working part time, when they want to work full time, for example) in the future. I need help figuring out if the categories I have are mutually exclusive.

In order to calculate this, I have 6 different categories: 1) Long term disability, 2) Unemployment (using the unemployment rate), 3) Underemployment, 4) Laid-off from job, 5) Working part time, but not able to full time work, 6) Non-participation in the workforce (for example, choosing not to work to be in school or to stay home and raise children). I'm trying to figure out if there is overlap between the categories. For example, I'm not sure if the chance that someone has a long term disability would be captured in the unemployment category. So, my question is - I need help identifying where there is overlap in the above categories or if there's anything I might be missing. Thanks.
posted by NoneOfTheAbove to Education (5 answers total)
It depends on the dataset.

For example, in the U.S., a common source of employment data is the Current Population Survey. This page has definitions of the different concepts. For example, there are some people who are not in the labor force and give "ill health or disability" as a reason. But there are people with a disability who are employed.

The main categories are in the labor force, of which employed are a subset. Of the employed, a subset are "At work part time for economic reasons, also referred to as involuntary part-time workers".
posted by Mr.Know-it-some at 11:46 AM on June 26, 2019

You can do exactly what you want to do with CPS microdata. There are several sources for this, look for CPS PUMS.

Realistically, you will need an actual no-shit statistical package to work with it.
posted by GCU Sweet and Full of Grace at 12:36 PM on June 26, 2019 [2 favorites]

A person with a long-term disability and outside of the labour force because of it would not be counted in the unemployment category. Unemployed is typically defined for data-purposes as not having a job your'e working at, not having a job you're not working at, and actively looking for work. The disabled person not looking for work is not unemployed.

That said, you can't calcualte the probability that you want using "the unemployment rate" (i.e. you can't say the unemployment rate is 4% and this plus that etc. and combine that with the other stats.) You need actual person-by-person microdata that includes precise data on the employment situation of each person in the data. The good thing abuot this is that then you'll be able to define the categories *yourself*. Do you think a person who is laid off but likely to be called back and looking for work is unemployed? What is underemployment (note that this is a whole field of research)? Your call.
posted by If only I had a penguin... at 1:30 PM on June 26, 2019 [1 favorite]

"Labor Force Participation Rate and Why It Hasn't Improved Much--Five Reasons Why Workers Dropped Out and Won't Come Back"--a good explanation of all that.
posted by Carol Anne at 1:33 PM on June 26, 2019

#4 - laid off is a way of becoming unemployed but those people would be counted in the unemployed or other categories.

What about discouraged workers - not actively looking but would re-enter the labor force if they thought they could be successful at finding appropriate employment. The are subset of non participation but lack of opportunity (could be issues of lack of local employment opportunities, declining industries or discrimination) is what is keeping them out, not personal preference.
posted by metahawk at 2:05 PM on June 26, 2019

« Older Best archival software to create online archive?   |   Where are my-my-my-my-my (comfy, sturdy) boogie... Newer »
This thread is closed to new comments.