In: Statistics and Probability
Think of (and describe) a problem in your work that can be addressed using a t test for two INDEPENDENT or two DEPENDENT/CORRELATED samples.
Clearly identify the independent and dependent variables. Note that the dependent variable must be continuous (interval or ratio). The independent variable must be dichotomous.
Two sample independent t-test
Suppose we want to study the prevalence of smokers and non-smokers
in males and females. We want to test whether the prevalence of
smokers is significantly higher in males as compared to females. We
can use two sample independent t-test to study the above problem.
Our dichotomous variables are Gender and whether they smoke or not.
We have to calculate the proportion of smokers in male and female.
Let P1 be the proportion of smokers in male and P2 be the
proportion of smokers in Female.
Here, our hypothesis to be tested is
H0 : P1 is equal to P2
H1 : P1 is not equal to P2
Then by calculating the test statistic and comparing it with the
table value we can decide if we H0 is to be accepted or
rejected.
Two sample dependent t-test
Here we want to study if the student perform well in exam after
giving them extra homework to be solved. We want to test whether
giving extra homework helps students to perform better in
examination. The variable to be used are the exam score of the
students before giving them extra homework to be solved and the
exam score of the students and after giving them extra homework to
be solved. We have to calculate the average score in exam by the
students before giving them extra homework to be solved and average
score in exam by the students after giving them extra homework to
be solved. Let M1 be the average score in exam by the students
before giving them extra homework to be solved and M2 be the
average score in exam by the students after giving them extra
homework to be solved.
Here, our hypothesis to be tested is
H0 : M1 is equal to M2
H1 : M1 is not equal to M2
Then by calculating the test statistic and comparing it with the
table value we can decide if we H0 is to be accepted or
rejected.