Sampling strategy
Part 2 sampling strategy (women)
Who made up the sample of pregnant women?
Women were included if they attended classes run by the three YfP teachers from part 1. This was in order to control key aspects of their experience while seeking as much diversity as possible by including women who attended three teachers’ classes rather than one teacher’s, and keeping within practical time constraints (Gerson & Horowitz, 2002). To ensure the women had enough experience of YfP for it to have the potential to effect change, a minimum attendance at six classes was set for continued inclusion in the study at the second and third interview stages.
How were the women chosen?
The aim was to offer participation to every pregnant woman in each of the three YfP teacher’s classes from September 2014 until the sample size was reached. The YfP teacher participants were asked to send a part 2 participant information sheet (PIS2, appendix 2) via email to every woman who requested to join their YfP classes. An example of one of the teacher’s email to participants is copied below:
The PIS2 (appendix 2) outlined that participation in the study would mean 20-30 minute interviews at three time points and informed potential participants that:
The study focused on general themes around why women attend yoga and their experience of the class, not specific people or situations
Interviews would be audio recorded, transcribed and all identifying features (names, hospitals, towns, midwife/GP names) would be removed
The study had received ethical approval from the University of Worcester and NCT
Data storage would be secure Hi there
Further to my previous email with details about my Monday night yoga classes A friend and colleague of mine at NCT is doing a PhD study into yoga in pregnancy and has asked if I could ask the women planning to come to yoga to talk to her about why they are choosing to come to yoga.
She is keen to talk to anyone - whether they are a first-time mum or have had a baby before, and whatever type of birth experience they are hoping for - whether it is a hospital, home or caesarean birth.
There is more information attached and she would be very happy to provide more details if necessary, her details are [email protected] or 0208 752 XXXX.
She would like to talk to you before you start your first class if possible.
Would you be happy for me to pass on your email to her?
With many thanks on her behalf
(Standard email to women booking T2’s classes)
If they agreed to participate, they had the right to withdraw from the study at any time with no pressure to continue and no need to provide an explanation for their withdrawal.
The women were asked to contact the researcher if they wanted more information or to see the full study proposal. If they were willing to take part, they were asked to sign the consent form and email it to the researcher.
The perceived benefits of working with the YfP teachers as gatekeepers were facilitating access to women attending YfP classes and reassuring the women of the researcher’s credibility and trustworthiness. The risks of using gatekeepers (King &
Horrocks, 2010) can include bias, where participants are consciously chosen by the gatekeepers in order to present a certain perspective. There is also an ethical danger that gatekeepers in a position of power may deny participants a genuine choice of taking part. The YfP teachers were not only gatekeepers, but also ‘insider assistants’, helping with recruitment. King and Horrocks (2010:32) produced guidelines for this research scenario, including considering the insider assistants’ trustworthiness, keeping them briefed, being in regular contact and ensuring that, after the initial contact, participants respond to the researcher not the assistant. These guidelines were adhered to during the present study.
How many pregnant women were chosen?
For part 2, in order to achieve theoretical sufficiency, enough data needed to be gathered to show development over time as the women progressed from joining the YfP classes to having their babies. The researcher’s experience led her to believe that only 30%-50% of pregnant women who booked the YfP classes would complete at least
six sessions; therefore, the preliminary design was to recruit twelve women to allow for at least six to have interviews at all three time points. The estimate was that this would yield eighteen interviews for analysis, approximately nine hours of transcript.
This, combined with the part 1 interviews and observations/videos, would mean twenty-seven separate data sources in total for the study. In practice, forty-five interviews, totalling approximately fifteen hours were analysed, for reasons described in the ‘part 2 participants’ section below. Therefore fifty-one data sources were analysed for the study in total, which is higher than most recommendations.
Participants
Part 1 participants (teachers)
Six teachers volunteered of whom three were chosen (T1, T2, T3) using convenience criteria of the location of their classes. They had been teaching YfP for 1 (T1), 10 (T2) and 7 (T3) years and had been teaching NCT couples courses for 2 (T1), 9 (T2) and 18 (T3) years. All the classes were facilitated by the teachers on a self-employed basis with remuneration partly based on class attendance numbers. Six classes were videoed and three teacher interviews held between May and July 2013.
Part 2 participants (women)
Recruitment was piloted with one YfP teacher (T2) in August 2014. Following T2 sending PIS2s to women booking her classes, six women contacted the researcher in August and September 2014. Two of these women did not respond to email requests for an interview date. With the aim of avoiding coercion and in consultation with the DoS, the decision was taken to email each woman who had offered to participate once
and then follow with a second email (or telephone call if the woman had given a number), before stopping contact if she did not reply.
The first four interviews with pregnant women who were about to start T2’s classes were held in August and September 2014. As the pilot access method appeared to be successful, on 29th September, T1 and T3 were also asked if they would send out the PIS2 to the women booking their classes which elicited the following responses from the teachers:
‘Yes of course. Hopefully will get a few participants’ (email from T1, 30th September 2014)
‘Yes happy to do this and get newbies all the time so can send some out soon’
(email from T3, 29th September 2014)
However, no participants from T1’s and T3’s classes contacted the researcher during October or November 2014. During that time, three more women from T2’s classes volunteered to participate, two of whom responded to interview requests. Data collection challenges are not uncommon for grounded theory research; Birks (2008) talks about how data collection can take twice as long as expected, and Miller and Bell (2002) describe how researchers’ principles at the outset of a study may be challenged by the realities of gaining a viable number of participants within deadlines. Personal reflective memos from this time show the researcher’s dilemma around how to encourage recruitment without using her power within the NCT organisation or negatively affecting her working relationship with the teachers.
A reminder was sent to all the teachers in November 2014 which resulted in two women from T3’s classes offering to participate. Another reminder to the teachers was sent in January 2015 followed by a phone call to ask if there was anything about the study that the teacher would like to discuss. T1 said she had forgotten to send out any invitations. Personal reflective memos from this time show that the researcher wondered if the YfP teachers were worried that the PIS2 might discourage women from joining their classes, leading to a consequent drop in teacher income, although this was not mentioned by the teachers. The researcher has no knowledge of how many invitations were sent out by any of the teachers.
The January 2015 reminder to the teachers resulted in thirteen more women offering to participate, ten from T2’s class, three from T3’s, and one from T1. In addition, a woman from T1’s class heard about the study from another woman and requested to take part. As recruitment had been challenging and over a longer time-period than anticipated, the decision was taken to interview all the women who volunteered, although the original intention had only been to interview twelve.
In total, twenty-five women contacted the researcher but three did not respond to requests for an interview date. Consequently first interviews were held with twenty-two women between August 2014 and January 2015.