Prof. Raewyn Connell discusses her work on masculinities and some of the issues raised by the subject.
This is my first attempt at an answer for this year's 52 Mark Question answer. It's not my best work as it was done very quickly between lessons, so a bit patchy - but hopefully gives some idea of some of the things that can be included in a question on sampling (which, in my opinion, is one of the toughest - and dullest - things you can be asked). Anyway, here it is. Enjoy:
Using the pre-release material and your wider sociological knowledge, explain and evaluate the use of sampling techniques in researching motherhood and social class 
The aim of Vincent et al’s study was to explore the ways in which mothers of different social class backgrounds experience life. This was a qualitative study, using semi-structured interviews. This suggests that Vincent et al are working from an interpretivist perspective, gaining rich, in-depth and valid data in order to fully understand the lives and experiences of the mothers in the study. The study took place over a five year period, though the pre-release material suggests that several different projects on several different groups was conducted in this time, so this may not have been strictly a longitudinal study (which would usually focus on a single group over a prolonged period).
In order to gain a sample for studying, Vincent et al used two sampling methods. The first was opportunity sampling. This involved Vincent et al visiting a range of different settings (e.g. nurseries, playgrounds, libraries) and finding mothers who were happy to take part in the research. Gathering a sample in this way meant that Vincent et al were able to gain the sample size they wanted; they could keep recruiting until they got the number they needed. Visiting different environments across two areas of London may help gain a more representative sample, but volunteer sampling overall is not the best method to ensure representativeness. Firstly, it is a non-random method, which means Vincent et al could have actively singled out or approached the mothers that they felt best matched their research (or what they wanted to find). Secondly, the mothers that agree to take part may not be typical of the research population, or may have their own agenda for taking part which is different to that of the researchers.
The second method was snowball sampling. This involves using one person or small group who fits the criteria for the sample and getting them to recruit others like them for the sample. This is a useful method for researchers, as it is quick and enables them to access groups and individuals they otherwise might struggle to reach. However, it is also non-random and is low in representativeness, because people are recruiting others that they know or who are ‘like them’.
When a sample is not representative, it is not usually possible to make generalisations from the research findings. This means it is not possible, for example, for Vincent et al to claim that what is true of the 55 working-class mothers in their study is true of all working-class mothers. To add to this, 55 is quite a small sample, considering the research population for this was working-class mothers in two big London areas, and a small sample has less chance of being representative.
The total sample was 126 mothers (55 working class and 71 middle class). Both of these are small samples considering the target population, although having smaller samples does have it advantages. The method of study Vincent et al opted for was semi-structured interviews. These can be time-consuming, as they are flexible by nature, so they are more suited to a smaller sample size. They were used to gain qualitative data, with takes more time to interpret and analyse than quantitative (numerical/statistical) data, so again a smaller sample size is beneficial here. Data gained through semi-structured interviews tends to be lower in reliability than that gained through structured interviews, as the flexible nature of the questions means interviews are less likely to be repeatable, but they are higher in validity, as the data is more detailed and the interviewer can built a rapport with the interviewee.
Overall, the use of sampling in Vincent et al’s survey could be argued to be problematic as the methods they used were low in representativeness and therefore mean that it would be difficult for Vincent et al to make generalisations from the results. Non-random sampling methods are open to bias and this, compared with an unrepresentative sample, can result in research that is low in validity, as it does not give an accurate insight in the real lives and experiences of the group being studied. Bias can be addressed through a range of methods, including a pilot study (pre-testing the main study to check for such issues) or respondent validation (getting participants to help review the results, to check the researcher hasn’t imposed their own bias or agenda) but it is not clear whether or not Vincent et al did either of these.
On the other hand, the use of non-random sampling techniques made sense on a practical level, as they ensured the sample was appropriate and they helped gain a sample quickly and efficiently, as no sampling frame would have been available for this group. Additionally, as Vincent et al are interpretivists, they may be less interested in gaining a large or representative sample and more interested in gaining rich, deep insights and understandings into a smaller group using qualitative data – and this is what they achieved.
For anyone sitting the G671 retake in May, here is the full 2007 Carol Vincent study on social class and childcare, a version of which provides the exam's pre-release material. Reading the full thing isn't essential for exam success, but is still useful in gaining a fuller understanding of the objectives and findings of the study and more detail on the methods used...
...Reading through the pre-release material, it seems clear that Social Class will be the theme of the exam. The study uses semi-structured interviews, but there is very little detail on these. It seems much more likely the 52 mark question will focus on the sampling techniques used.