Fundamentals of Statistics contains material of various lectures and courses of H. Lohninger on statistics, data analysis and chemometrics......click here for more.


Exercise - Testing the Reaction Time of a Person

The reaction time of a person is of much importance in many critical situations. It is also well-known that the reaction time is strongly influenced by the consumption of alcohol or by the degree of fatigue. In order to decide whether the reaction time of a person is above some predefined threshold, a one-sample t-test should be applied. In order find significant differences for different experimental set-ups, a two-sample t-test should be used.

9 persons have repeated a reaction test on two different occasions between 10 and 40 times. The first test run was performed in early afternoon at the beginning of a two-hour course in statistics (column denoted by "t-early"). The second series of tests were conducted after the two-hour course ("t-late"). Go to the  DataLab    and try to answer the following questions:
 

  • Perform a t-test to find out which of the persons have shown a reaction greater than 500 msec. Use the identifier in the third column to discriminate between the individual test persons.
  • Client number 7 shows an average reaction time of 551 msec in the second run. The t-test however, does not indicate any significant deviation above the 500 msec threshold. Why?
  • Is there any significant difference in the reaction times of client 6 before and after the two-hour course?
  • Is the t-test appropriate for this data set?