The industrial company Acme is looking at buying new computer based technology to replace similar but older technology. The task the workers have to perform is relatively complex, even with the assistance of the technology, so the company has run a small user trial to compare two alternate pieces of equipment (A & B) available on the market.
Purchase issues such as capital and running costs etc were likely to be very similar so the company was only really concerned with the likelihood of workers making errors with the new technology because errors impact on productivity. The company wanted to see if there was any real difference between A & B in terms of the number of errors experienced by the workers.
Twenty workers were randomly selected from the workforce and were randomly allocated to two groups of 10. Group A trialed the Technology A and Group B trialed technology B. The number & type of errors each worker made was recorded. However because of an unforeseen breakdown in the system, only 6 of Group B workers were able to fully complete their trial – the data are given in Table 1 below and for the purposes of any statistical calculations; it can be assumed that the error data form a normal distribution.
The data sheet is attached in the MS Word document.
So the questions that I have are:
For Null Alternative Hypothesis:
I came up that the mean errors for group A will equal group B
Alternative: Mean errors for group A will not equal group B
What type of graph would best summarize the data? I'm confused as to what graph would be appropriate to represent the data.
Also, how would i calculate the standard deviation for this. All of the data has a value of 1 - therefore, I'm getting a standard deviation of 0.
In order to calculate the 95% confidence interval, should I use the 2 sample t confidence interval - if so, would I use the pooled one or the unpooled one?
Finally, to carry out the statistical test, I chose to use the 2 sample t test.
Was I correct?
Any help will be greatly appreciated.