Getting to grips with STATA (check). Getting access to Natsal-3 data (check). Analysing data (check, and check). Reviewing analysis with supervisors (check) resulting in discussing again and going round in circles about the best way to group contraception for my analysis (not quite, almost… and back to the drawing board).
Its not surprising because there are many different ways that contraception can be grouped. However it is important that I fully understand the different options, weigh-up the pros and cons, and am confident that my groupings are appropriate for my research enquiry.
Most contraception is reversible, meaning that once the intervention is stopped/removed, fertility levels return. However two methods are permanent; sterilisation for women and vasectomy for men.
My research is focused on young people aged 16-24. In particular my interest is about ‘good’ or ‘effective’ (exact terminology still work in progress) contraception use. In other words; once people have chosen and obtained their contraception its the application of the method which actually determines their risk of pregnancy. Therefore I’m not interested in permanent forms of contraception because a) very few young people use this and b) there is no variation in use of the method. This decision was straight forward.
This is where things become a little more complicated *Deep breath*
Different contraceptive methods have different levels of efficacy. For example; according to BPAS the contraceptive implant is 99.9% effective, contraceptive injection is 94% effective, combined contraceptive pill is 92% effective, and male condoms are 82% effective for preventing pregnancy. However there are two issues to consider:
In acknowledging that humans do not live in a social vacuum, its important to consider the ‘typical’ use of different methods. For some, particularly LARC (long-acting reversible contraception) perfect use is the same as typical use. However for non-LARC methods the difference between perfect and typical use is quite dramatic… the graph below shows rates of unintended pregnancy (%) following the use of different methods after a year:
I think its clear that my research is more concerned with typical use (i.e. what actually happens) rather than perfect use. However what I want to be able to do is compare ‘good contraceptive users/practice’ with ‘less good practice’, in order to find out what factors make it more like to be in one group than the other. There is no obvious way to group people, and it turns out there are many opinions regarding different grouping options. And I’m in the middle of this at present. What I need to consider is what is most appropriate for my research question, what can the Natsal data tell me (and what can’t it), and how can I best present and analyse the data.
Something I’ve realised recently is the way I’m writing up my work isn’t quite right. I’ve been writing things up as if it were going to be published (“this is what I did and this is what I found”) which leaves little room for outlining discussions and the decision making process along the way. Thanks to a discussion with a colleague yesterday, I’ve started to realise that capturing this discussion is important within a PhD thesis. I should outline my rationale for making groups, and write up the different options available, leading to the justification for the method I end up using.
And so following this mini brain-wave – time to whip up some words to explain my thinking to my supervisors.