I am unfortunately distracted by the large number of significant digits. House 1 burned 2.785kg @ 0.038150685 kg/min, which implies a precision of micrograms per minute,
Using decimal math, that works out to exactly 72.99999986894075427479218263 minutes. Realistically, it's 73 minutes. There are only 2 significant digits so the reported rate should be 0.038 kg/minutes.
Also, there's is a factor of 3 variation in wood/meal unit, compared to 25% improvement in switching stoves. Why aren't all of the houses using the efficient cooking technique of house 3? For that matter, how reproducible are the numbers for each house?
Finally, I see that house 2 had no gains, using about the same wood/meal unit with all three stoves. Does this mean that about 1/3rd of the users of the stove will see no gain to switching? Or does it mean that there isn't enough data to draw a good statistical conclusion? What are the odds of seeing these results purely from random chance?
"Unfortunately, results from House 3 were often unreliable due to operational difficulties that caused large variations in meal size, type, and preparation."
I'd love to know the human story behind these "operational difficulties".
I am unable to find that text in the article. I am even unable to find the word 'operator'.
In any case, I pointed out that there's a 3x difference between two houses, so I am in complete agreement with that statement. My post asks a different question, which is, what is the error range on the observed results? What are the odds that the stove doesn't actually have an effect on wood use, but that some other factor ('operator and the conditions of use') resulted in the observed differences?
"The Hawthorne effect (also referred to as the observer effect) is a type of reactivity in which individuals modify or improve an aspect of their behavior in response to their awareness of being observed."
Using decimal math, that works out to exactly 72.99999986894075427479218263 minutes. Realistically, it's 73 minutes. There are only 2 significant digits so the reported rate should be 0.038 kg/minutes.
Also, there's is a factor of 3 variation in wood/meal unit, compared to 25% improvement in switching stoves. Why aren't all of the houses using the efficient cooking technique of house 3? For that matter, how reproducible are the numbers for each house?
Finally, I see that house 2 had no gains, using about the same wood/meal unit with all three stoves. Does this mean that about 1/3rd of the users of the stove will see no gain to switching? Or does it mean that there isn't enough data to draw a good statistical conclusion? What are the odds of seeing these results purely from random chance?