How do inductions bias the due date statistics?

I wanted to make this simple data table after many visitors posed this question: "We know that many women are induced for to avoid going postdate, doesn't this bias our data?"

Fair question. Most of the graphs I have just throw out inductions and show data based on only spontaneous labor. But what if ALL those women who were induced were actually supposed to go late, how long would the real average pregnancy last?

Is the Mittendorf-Williams rule be more accurate?

I'll just say it: Many people also ask this question because they think the medical community has an agenda and wants to induce us "early" for convenience. I do not believe this; I like doctors. They are having babies too.

288 days is actually a huge difference and it would be VERY difficult to keep using 280 days if 288 was the truth. With the 9-10 day standard deviaitons we've found, an average of 288 would mean that on day 280, 80% of pregnant women would have to still be pregnant.

Is that even possible? On this page, I pretend that every single induction before 40 weeks would have been a spontaneous birth at 42 weeks... and I still can't get to a 288 day average. Sorry, Mittendorf-Williams rule. But so many women spontaneously go into labor in the days leading up to 40 weeks, there aren't enough inductions to make a difference.

Here are the median days for the spontaneous labors (12290 total), and the shifted median if every induction (3779 total) was actually supposed to be a very late birth:

Conclusion: if all the induced women who took my survey had been left to stay pregnant until the very end (past 42 weeks), then the median day for birth would shift by 2 days. Not an entire week.