

Compute the number of comparison performed in the evaluation of the experiment and choose such an adjusted significance threshold using Bonferroni correction so that the family-wise error rate 0.05 is maintained. Use Student's t test to compare each condition against each other separately for both variables (var1| cond1 vs var1 | cond2, var1| cond1 vs var1 | cond3, var1| cond1 vs var1 | cond4, …, var1| cond2 vs var1 | cond3, var1| cond2 vs var1 | cond4, …, similar to the boxplots bellow). In an evaluation of the experiment, we want to know whether the variables are different for certain experimental condition.
MATLAB STUDENT T TABLE SKIN
The data could originate from an experiment where a test subject was viewing images with certain contents (1 - neutral, 2 - adventure, …) while the size of his/her pupil was measured by eye tracking device and his/her skin conductance by polygraph (the variables). The data contain three columns, first column indicates experimental condition (integers 1 to 5), the second and third columns contain certain measurement. Use data multidata.mat to examine the problem of multiple comparison with Student's t test.

Submit the implementation as an m-file alongside the report from the lab. This variant is appropriate when the two samples don't have the same sample size and their standard deviations are similar. The test statistic is $$ t = \frac$$ and the t statistic follows t-distribution with $n_1 + n_2 - 2$ degrees of freedom. One sample Student's t test can be used to test whether the sample mean is different from 0 (or a specific value).
