A first year undergraduate module was developed with an assignment that could be in two different modes of presentation:
Version 1 - conventional paper based
Version 2 - purely electronic
A study was carried out to see if the two versions produced different results. Two independent random samples of students were selected and each sample of students were given one of the assignment versions to complete. Version 1 produced a mean mark of 75.5 per cent and Version 2 a mean of 80.8 per cent. The 95 per cent confidence interval of the difference in the means was 5.3% ± 4.3%:
(i) Would this confidence interval suggest that the module developers could be 95 per cent confident that there was a difference in the marks? If so, why?
(ii) What reservations would you have regarding the design of this study?