www.lesswrong.com/posts/aKT6WCs3ASvBWfLw9/how-lesswrong-helped-me-make-usd25k-a-...

The sentence treats 22 weeks as the typical amount of time a job seeker spends unemployed before getting a job. But the Bureau of Labor Statistics says its unemployment-duration statistics do not mean that.

BLS's definition page explains that duration-of-unemployment measures reflect "still-in-progress spells of unemployment, not completed spells" and explicitly says they "should not be interpreted as the length of time it takes someone to find a job." In other words, using the 21.6/22-week BLS figure as the average time a job seeker spends searching is a misuse of the statistic.

BLS's own discussion of completed job-search spells gives a much shorter figure for successful searches: in 2019, "the median number of weeks that jobseekers were unemployed in the month prior to finding work was 5.6 weeks."

So the post's base-case claim is inaccurate because it presents an ongoing-spell unemployment measure as though it were the typical time needed to land a job.

This parenthetical repeats the same statistical mistake later in the post.

The BLS unemployment-duration figure around 21.6/22 weeks is a measure of how long people who are currently unemployed have been unemployed so far. BLS explicitly says these are still-in-progress spells and that they should not be interpreted as the length of time it takes someone to find a job.

When BLS discusses completed job searches instead, it reports that in 2019 the median number of weeks unemployed before finding work was 5.6 weeks.

So describing 22 weeks as "the average amount a job seeker looks for a job" is inaccurate.