How to do gender analysis in health systems research – Significance of gender as a variable

This blog post is part of a series looking at key questions related to gender analysis within health systems research. In this post we explore the significance of gender as a variable.

RinGs Steering Committee

20 October 2015

On 8 September 2015 Research in Gender and Ethics (RinGs): Building Stronger Health Systems held a cross-RPC webinar on “How to do gender analysis within health systems research”. The webinar involved 26 members from Future Health SystemsReBUILD, and RESYST.

Webinar participants asked some very interesting and relevant questions about how gender analysis can be incorporated into health systems research. In this blog series, we discuss some of the issues raised and we would be interested in your viewpoints as well. Please let us know in the comments section below!


Q: Within health systems research, should gender be mentioned when it is not significant significant as a variable? Should it still be reported? Female participant, China.


A: Gender and gendered power relations is relevant for all health systems research, including efforts to examine and change vulnerability to ill- health; household decision-making and health-seeking behaviour; access to and utilization of health services; the design and use of medical products and technology; the nature of the health labour force; what health programs or issues get financed; what health issues get prioritized and how they are framed.

Within quantitative research there are many things which may impact whether or not a variable is statistically significant, including: the overall sample size, the sex disaggregated sample sizes, the questions asked, and the type of analysis used. Therefore just because a variable was found to be statistically not significant, does not mean it doesn’t play a role, or indeed it is not important. As researchers we need to understand more about the context and explore why certain variables are not found statistically significant. It gives us an opportunity to revisit our conceptual frameworks or program theories and explore whether the variable is not relevant in the context we are examining, or did we not fully understand other dynamics going on that are masking social inequalities? For example, in India, parents tend to favor male over female children in terms of health care seeking. A finding of equal access to public sector doctors for male and female children may not indicate that gender bias does not exist, if the gender bias is predominately manifested in the private sector which makes up the bulk of care seeking for sick children.

It is important to show not just what variables were statistically significant, but also to revisit our conceptual frameworks and theories to enquire further about the variables that were found to be statistically not significant and examine why this is the case. Such insights can lead us to better understand nuances and unexplored research avenues.

Next week this blog series will explore the significance data collectors’ gender


To view a recording of the webinar or the webinar presentation slides, click here.

For more information about RinGs visit our website.