I have a problem using a 2 sample t test. We need to test the assumptions of the
ID: 3055133 • Letter: I
Question
I have a problem using a 2 sample t test. We need to test the assumptions of the statistic, explain the results and draw conclusions. If I run a normality test from the data using Anderson-Darling test, one sample has a p-value of 0.013 and the other has a p value of 0.750. This means that one sample is normally distributed and the other isn’t. Also when testing variance my p values are high for Levene’s and Bonett’s test. My questions are these. I think The high p value in both variance tests means that the ratio of the standard deviations is 1, what does that mean exactly? My 95% CI is listed as 0.75 for both samples. Lastly, if one sample is not normally distributed, what does that mean for my t test overall ? I’m not sure what implications it has on the data. The t test has a p of 0.000. I would reject the null hypothesis but I am not sure given the other assumptions are not clear.I have a problem using a 2 sample t test. We need to test the assumptions of the statistic, explain the results and draw conclusions. If I run a normality test from the data using Anderson-Darling test, one sample has a p-value of 0.013 and the other has a p value of 0.750. This means that one sample is normally distributed and the other isn’t. Also when testing variance my p values are high for Levene’s and Bonett’s test. My questions are these. I think The high p value in both variance tests means that the ratio of the standard deviations is 1, what does that mean exactly? My 95% CI is listed as 0.75 for both samples. Lastly, if one sample is not normally distributed, what does that mean for my t test overall ? I’m not sure what implications it has on the data. The t test has a p of 0.000. I would reject the null hypothesis but I am not sure given the other assumptions are not clear.
My questions are these. I think The high p value in both variance tests means that the ratio of the standard deviations is 1, what does that mean exactly? My 95% CI is listed as 0.75 for both samples. Lastly, if one sample is not normally distributed, what does that mean for my t test overall ? I’m not sure what implications it has on the data. The t test has a p of 0.000. I would reject the null hypothesis but I am not sure given the other assumptions are not clear.
Explanation / Answer
Result:
I have a problem using a 2 sample t test. We need to test the assumptions of the statistic, explain the results and draw conclusions. If I run a normality test from the data using Anderson-Darling test, one sample has a p-value of 0.013 and the other has a p value of 0.750. This means that one sample is normally distributed and the other isn’t. Also when testing variance my p values are high for Levene’s and Bonett’s test.
My questions are these.
I think The high p value in both variance tests means that the ratio of the standard deviations is 1, what does that mean exactly? My 95% CI is listed as 0.75 for both samples.
Lastly, if one sample is not normally distributed, what does that mean for my t test overall ? I’m not sure what implications it has on the data. The t test has a p of 0.000. I would reject the null hypothesis but I am not sure given the other assumptions are not clear.
High p value in both variance tests shows that the two groups have equal variance. This means the equality of variance assumption is not violated.
You not mentioned about the sample sizes. If the sample size is large( about 30) you need not worry about the non normality of one group. You can infer from the p value of 0.000 that the two group means are significant.
Further T test is "robust" to the assumption of normality. Note that the real question is not exactly normal, but rather normal enough for the given problem. For moderate deviation from normal assumption may not affect the results.
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.