Early last year, we began marking SteelyKid’s height off on a door frame in the library. She occasionally demands a re-measurement, and Saturday was one of those days. Which made me notice that we now have a substantial number of heights recorded, and you know what that means: it’s time for a graph.

The “featured image” above shows the results, and I’ve stuck in a zero-day point using her length at birth. The solid line in the graph is a linear fit to the data, because nothing could possibly be wrong with that. According to the fit, we can project that she’ll reach a height of 3.0 meters a few months before her twentieth birthday, and since the R2 value is 0.99, you know that’s solid science.

Her current height, for the record, I rounded up to 44.5 inches. And if you want to see just how big she’s gotten, here’s a photo from an after-dinner trip to the playground at what will be her school in just a couple of weeks, when she starts kindergarten. Yikes.

SteelyKid at the top of the rope climber on her school's playground.

SteelyKid at the top of the rope climber on her school’s playground.

I got The Pip to stand by the door frame on Saturday as well, and his official height was 33 inches. As it’s a single data point, though, it’s not worth graphing– you’ll have to wait until we accumulate more data before we can compare the two in a truly scientific manner…


  1. #1 Wes Morgan
    August 25, 2013

    Now do an analysis of your grocery bills. *chuckle*

  2. #2 Bee
    August 25, 2013

    Interestingly the curve is indeed pretty much linear for quite some time before it stagnates, but I don’t think the length at birth should be on the linear trend. They grow faster in the first months, see eg http://happybabyusa.files.wordpress.com/2011/05/growth-chart.jpg

  3. #3 Michael Kelsey
    SLAC National Accelerator Laboratory
    August 25, 2013

    Scientific parenting, indeed :-) If you were doing the regular post-natal checkups, you should have a whole bunch of data points up to around day 400 or so, and you’ll see that the actual curve down there is roughly logarithmic.

    When my daughter was born, I started recording height, weight, and head size in Excel. I also found the raw data from both the CDC and WHO which are used to generate the height-weight curves.

  4. #4 RM
    August 25, 2013

    Bee, I don’t think this is inconsistent with that burst phase effect. It looks like if you remove the t=0 point and refit the SteelyKid data, you’ll get an intercept of 65-ish rather than 50. From the linked chart, burst phase accounts for ca. 10 cm. of height difference.

    What we can conclude from this, though, is that Chad is a terrible parent. ;-) Either for malnourishing his kid during a crucial growth phase, or by being negligent enough to skip recording crucial height data during the early period of this experiment. – He really needs to make it up to SteelyKid. Buy her a pony or something. It’s safer not to have a 2+ meter tall teen upset at you.

  5. #5 Chad Orzel
    August 25, 2013

    I do have higher-frequency data covering the gap in this graph, from her regular doctor appointments during that span. Those are recorded via a different technique in another data repository, though, and as such I’m saving them for a separate publication.