The Labor Probability Graph was designed for anyone who enjoyed the Labor Probability Calculator and wanted a little more detail about the underlying distribution. The Graph uses a left skewed normal distribution to modal the odds of spontaneous labor.
Important Tool Caveat
This tool predicts the odds of spontaneous labor without consideration of possible negative outcomes. It should not be used when making medical decisions.
About the Model
Prior research has shown that the distribution of spontaneous labor approximates a normal distribution with a standard deviation of 9 [1] or 13 days [2]. These studies provide the initial building blocks for our model, but fail to paint a complete picture. Neither distribution does a good job of estimating the number of preterm deliveries. The CDC has found that 9.6% of babies are born prematurely, before 37 weeks. The normal distribution based on [1] and [2], however, would predict less than 3% of babies are born prematurely (.03% and 2.7% respectively). In fact, [1] specifically excluded preterm deliveries from their analysis which is why their model predicts so few preterm babies, which explains why it's model predicts so few of them.
The normal distribution is favored in these types of applications for it's simplicity and tendency to fit the data. In this case, however, the normal distribution is likely an over simplification. A key drawback is that the normal distribution is a symmetric distribution. That means two points equally far from the mean will have the same probability. A symmetric distribution that predicts 10% of babies will be born before 37 weeks will also predict 10% of babies will be born after 43 weeks. For all the births prior to 30 weeks, there would be a similar number of births after 50 weeks. Additionally, in a symmetric distribution the median (point which half of all pregnant individuals would have gone into labor), is equal to the mode (most common day to go into labor.) In prior studies, however, the mode date is typically after the median [2]. Although not scientific, conventional wisdom is that the most common day to go into labor is around 41 weeks.
Taken together, these data points suggests that a skewed normal distribution might be more appropriate. The skewed normal distribution is a family of distributions that includes the normal distribution, however the skewed normal distribution need not be symmetric. Where the normal distribution is defined by two parameters (mean and standard deviation), the skewed normal distribution is defined by three (location, shape and scale). Using mean squared error we identified a skewed normal distribution that closely approximates the normal distribution identified with prior research (MSE of 0.002), accounted for 10% of spontaneous labors occurring prematurely, and predicted roughly half of all pregnant individuals would go into labor before their due date and half after.
Differences between Datayze's model and those found on other websites
The key difference between our model and every other model out there (that we know of) is our model better incorporates the possibility of preterm spontaneous labor. Like [1], most online models predict the probability of spontaneous labor before 37 weeks is approximately zero. As you can see from above we estimate the probability of spontaneous labor before 37 weeks as 9.6%, matching the number from the CDC. If you're past 37 weeks, this distinction will be less important. Before 37 weeks and our model will serve you much better.
Differences between Labor Probability Graph and the Labor Probability Calculator
You may notice the probabilities differ between the Labor Probability Graph and the Labor Probability Calculator. That is because the two apps are modeling two similar sounding, yet different events. This Labor Probability Graph shows the probability of spontaneous labor for a pregnant person without considering how far along she is in their pregnancy. The Labor Probability Calculator shows the probability of spontaneous based on how far along she is by renormalizing the distribution to include only the possible remaining days in a person's pregnancy. After all, for a person who hasn't gone into labor by today the probability of spontaneous labor starting yesterday is, by definition, 0%. Statistically speaking it's the difference between the probability of labor at 40 weeks 0 days in general, p_labor(40w0d), and the probability of labor at 40 weeks 0 days for a person who is already 39 weeks along, p_labor(40w0d|39WeeksAlong).
This website is intended for general information & entertainment purposes only. This website is not intended to be considered medical advice. I am happy to answer questions about the apps, and the mathematics behind them. I cannot answer medical questions.
Pregnant? You may enjoy our other pregnancy apps like the personalized week by week calendar.
Into Probability Distributions? You may also be interested in our Daily Miscarriage Probability Chart, which gives both the probability of miscarriage and probability of carrying until birth.
Wanting to become pregnant? Our Time to Conception Estimator can estimate how long it may take.
Have baby names on the brain?
Try our Baby Name Apps, including Baby Name Uniqueness Analyzer or Baby Name Explorer. Still can't find a name you like? Try the Unique Baby Name Generator or the Name Blender which generates new names based on the latest trends.