(Splitshire, n.d.)
This is a simulated study that using a range of fake wheat yield data to compare with a country-wide average to showcase the application of One-sample Student’s T-test. Let’s say the countrywide average is 2249 kg/ha, and I have a range of my experimental result:
my.yield <- c(2456, 2365, 2876, 2022, 2498, 2655, 1989, 1994, 2556, 2601)
summary(my.yield)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1989 2108 2477 2401 2590 2876
Choosing boxplot to quickly visualise these statistical metrics with the country-wide yield average:
Applying one-sided and two-sided method on One-sample Student’s T-tests to find out:
• statistically different, • statistically higher, or • statistically lower.
t.test(my.yield, mu = 2249)
##
## One Sample t-test
##
## data: my.yield
## t = 1.569, df = 9, p-value = 0.1511
## alternative hypothesis: true mean is not equal to 2249
## 95 percent confidence interval:
## 2181.755 2620.645
## sample estimates:
## mean of x
## 2401.2
t.test(my.yield, mu = 2249, alternative = "greater")
##
## One Sample t-test
##
## data: my.yield
## t = 1.569, df = 9, p-value = 0.07555
## alternative hypothesis: true mean is greater than 2249
## 95 percent confidence interval:
## 2223.375 Inf
## sample estimates:
## mean of x
## 2401.2
t.test(my.yield, mu = 2249, alternative = "less")
##
## One Sample t-test
##
## data: my.yield
## t = 1.569, df = 9, p-value = 0.9244
## alternative hypothesis: true mean is less than 2249
## 95 percent confidence interval:
## -Inf 2579.025
## sample estimates:
## mean of x
## 2401.2
There is no statistical different between experimental mean and the country-Wide mean (P-value>0.05).
There is not a strong evidence that the experimental mean is greater than the country-wide mean (P-value>0.05).
There is not a strong evidence that the experimental mean is lesser than the country-wide mean (P-value>0.05).
Splitshire, n.d., Field of growing wheat, https://freerangestock.com/photos/41001/field-of-growing-wheat.html