Need suggestions for young students analyzing T/F data

I teach an intro course in robotics to middles school students. I need help on how they can conclude from data that is T/F.

I encourage them to create a mini-experiment each week like how far will their car roll after going down incline plane of different angles. Then they must do a bit of mathematical analysis like a graph and y=mx+b.

But experiments are frequently a change in independent variable until a T/F result such as how high can LEGO bricks be stacked before falling. They end up with data like (F=stack not fall, T-fell): 1 brick:F, 5 bricks:F, 10 bricks:F, 15 bricks:T.

Any suggestions on how they can summarize this data into a more sophisticated conclusion? Obviously they can state that less than 15 bricks is stable. A graph doesn't really make sense. I've had them repeat to find the most accurate value, like above test 11 bricks, 14 bricks, 12 bricks, 13 bricks.

Maybe there is no better treatment than the English statement. But if there is better I'd appreciate the advice.
Last edited:


Less is more. Stay pure. Stay poor.
This reminds me of a survival curve, but instead of time to event (e.g., say relapse, death, heart attack), you have number of blocks until fall. So on the X axis is plotted number of blocks, on the y axis is probability, given as you stated you repeated the test under attempted identical circumstances. You may get away with doing a simple Kaplain Meier Curve.

P.S., You won't have any censoring (righthand I believe), so your line will eventually intersect the x-axis. In medical research, some individuals neveer have an event or are not followed long enough, so the line eventually plateaus.

You could repeat the test with blocks of different shapes and plot different curves (which look like steps actually) on the same graph to see how they compare.


TS Contributor
That would work as should binary logistic regression. The hard part is keeping it within the abilities of middle school students. Though they could create this graph less the BLR curve.