In: Statistics and Probability
5 A) Inferences based on t-statistics are said to be more robust than those using z-statistics. Explain what this means.
B) Confidence intervals using t-statistics are generally broader than those using z-statistics. Explain why this is.
C) Explain as completely as possible what a Student’s t distribution is and why William Gosset developed it.
Solution:-
(A)
Both Z score and t statistic have equal numerator, but they are different in terms of standard error calculation (denominator). T statistic are more variable than Z score. This is caused by extra variability of sample variance in t static with respect to Z. Epecially in small sample size its power is greater than for same significance level. So generally we can say that t test is more robust
B)
Due to critical t value which is higher than Z value.
t confidence inerval
t Z - confidence interval
(C)
student t distribution is a continuous probability distribution which is appled as below.
where
Z N(0, 1)
is a chi square random variable with D degree of freedom.
pdf:
He invented because he felt big sample to select best yielding varieties of barely (which he was worked) was telious. So he developed a way to extract a small samples, result in new discovery of t distribution.