Discussion questions

Q1- A researcher wants to investigate the safety of a procedure that is carried out on women whose babies get ‘stuck’ and who are therefore in distress during labor in childbirth. In particular he
wants to compare the outcomes for both mother and baby from this procedure with those for another, more common, procedure that is used in similar situations. He has listed the outcomes of interest.
He intends to collect data over the next 12 months from mothers who attend a particular large maternity hospital and who experience one of the two procedures during routine care.

(a) Is this an experimental or observational study? Justify your answer.

(b) Describe any variables, other than outcome variables, that you think would be important in this study, explaining why you believe them to be important.

(c) Describe any problems you would anticipate in collecting the data.

(d) Briefly state the steps that would be necessary in processing the data ready for analysis.

Q2- A team of university researchers is investigating the experiences and opinions of nurses working in nursing homes for elderly people. They specifically want to compare responses across three
areas of the United Kingdom: England and Wales, Scotland and Northern Ireland. This is because there are variations in the regulations relevant to their research questions between these three
areas. The research is to be led by three universities, one in each geographical area. There are many more nursing homes for elderly people in England and Wales than in the other two areas.
The researchers have decided to select 30 nursing homes in each of the three areas. These are to be near to the participating universities and known to the researchers. It is believed that this
will increase the response rate of the selected participants in the research. Within each selected nursing home, three nurses will be chosen at random from the nursing staff at the home.

(i) Explain why the proposed survey uses stratified sampling with the three areas of the United Kingdom as strata.

(ii) Within each stratum, what sampling method has been used to select the homes? Comment on and criticize the method used.

(iii) Comment on the claim that the method of selecting nursing homes is likely to increase the response rate.

(iv) The researchers claim that they will get a random sample of nurses working in nursing homes in the United Kingdom. Comment on this claim.

(v) Comment on the appropriateness of aiming for equal numbers of nurses from each of the three areas.

Q3- Explain the meaning of the above 95% confidence interval as if to a non-statistician.

A researcher wishes to conduct an experiment to see whether or not consuming alcohol affects the mean reaction time of drivers.
i. What would be a suitable null hypothesis, and an appropriate alternative hypothesis?
ii. What would be the result of making a Type I error and what would be the result of making a Type II error?
Q4- A survey is to be carried out on the alcohol consumption of full-time students attending a particular university. Many of the questions are sensitive in nature. All such students have an email
account and ready access to a computer which they are required to use in their courses. Each student’s term-time address is recorded centrally by the university. It has been decided that it would
not be appropriate to carry out face-to-face interviews or phone interviews for this population.

(i) Comment on the decision not to use face-to-face interviews or phone interviews.

(ii) The team is considering the use of either questionnaires sent by post or an internet survey. What would be your advice and why?