In short...
None of the stuff claimed is true and nobody at Google Science Fair apparently read their project report.
They won for being cute little girls. Possibly for having a puppy in the presentation for extra cuteness.
I initially wanted to correct myself on numbers above, cause it's just the germination that was up to 50% and Google Science Fair summary DOES state that the results showed "crop germination by up to 50%, and increased barley yields by 74%".
And then I checked the video and their results.
Which are both loaded with weasel words, omissions and plain old padding the numbers.
From the project documentation:
https://www.googlesciencefair....
The optimum concentration of r.japonicum for the germination of barley seeds was found to be 2x107CFU/ml (13% reduction; ANOVA p<0.0264).
R.leguminosarum had a positive effect on the germination of Barley and reduced germination time by approximately 40% at 25oc (ANOVA p<0.0001).
For Oats, an optimum concentration of 4x106 CFU/ml of r.japonicum was observed to be most efficient and resulted in a reduction in germination by 22 hours (28% Reduction; ANOVA p<0.0001).
Lower concentrations of r.leguminosarium were most effective on oat germination. A concentration of 16x104 CFU/ml reduced germination times from 86 to 66 hours (23% reduction; Dunnett test p<0.0001).
13%, 40%, 28% and 23% reduction in germination time for various crops. Reported as 50%.
Small Scale Agricultural Tests
R.japonicum was seen to have a positive effect on the length and dry mass of barley crops. (+10.4% length increase:+13% dry mass; p<0.0328), the effect was more notable at higher concentrations.
It was observed that Oats treated with a higher concentration of r.japonicum (4x106 CFU/ml) produced a greater dry mass (p=0.0248) and longer length (p=0.0043) than water treated seeds.
10.4% increase in length for barley.
13% increase in dry mass for barley and "a greater dry mass" for oats in small scale test.
Only problem is... length increase was noted for n=300 plants.
Dry mass increase for only n=24. Cherry picking? P-hacking?
You won't find those numbers in the text though. Only in the tiny low resolution graphs.
Large Scale Agricultural Tests
Lower concentrations of r.japonicum (3x109 CFU/ml) with peat as a carrier were the most successful treatments (ANOVA p<.0001) and resulted in an average increase in plant dry mass of 0.284g/5 seedlings (74%).
Spraying the seeds with aqueous culture post planting increased dry mass by a mean of 44% (Dunn p<0.0001)
74% increase (and 44% increase for an alternative method) in dry mass is there BUT...
It's dry mass of the entire plant. Roots and all. And this time, without the numbers on the length of the plants.
And no information on if there is correlation between the length of the plant and its weight.
I.e. Is it barley grain or barley grass?
Cause, as we are not talking about acres but of mass, crop yield of barley is just a fraction of the mass of the plant.
So "an average increase in plant dry mass" IS NOT "increased crop yield by an average of 30% with some results exceeding 70%", as stated in the conclusion.
This is just Google throwing money at anything that will make them look good.
No proof of results necessary. Just make it LOOK good.
Which gives me a very icky feeling of exploitation. Of children, minorities, certain genders...
2011 - three girls, from USA, two of them racial/ethnic minorities.
2012 - a "Caucasian" girl from USA, three boys from Spain (i.e. Latinos AND foreigners so it's a little more diverse and not all USA) and another "Caucasian" boy, from USA.
2013 - ethnic/racial minority boy from Australia, a "half Filipino" Canadian girl and an Asian boy from USA.
2014 - 3 "Caucasian" girls from Ireland, one from Canada, one Asian boy from USA and an ethnic/racial minority boy from India.
In Google America, only girls and Asians can science?
Or is it that "we are the world, we are the children" photos just look better?