Hi Emma,
Formula for this is:
z = (xbar - mu)/SE; SE = standard error = sigma/sqrt(n)
Now, you want greater than for 20% of the time--which means you want the 80th percentile. Since fruit weights are normally distributed, you can check the interior of the z-table for 80%, or 0.8. There are also options to do this using software. Anyway, closest interior probability on my z-table is 0.7995; corresponding z-score is:
z0.8 = 0.84
Now, we want xbar, the sample mean, but we need standard error (SE) to get it:
SE= sigma/sqrt(n)
sigma= standard deviation = 14
n= sample size = 8
SE= 14/sqrt(8) = 4.95
Now, we can substitute into initial formula:
z= 0.84
xbar= xbar, weight we are looking for
mu= population mean = 766
SE = 4.95
Substituting:
0.84 = (xbar - 766) / 4.95
4.158 = xbar - 766
xbar = 770.158, to one decimal place;
xbar = 770.2
I hope this helps.