MIXED METHODS IN HEALTH SERVICES & POLICY RESEARCH Shoshanna Sofaer, Dr.P.H. School of Public Affairs Baruch College OVERVIEW Why use mixed methods? Key issues Timing/sequencing Multiple methods for collecting data for a single variable v. multiple variables How one type of data collection can inform the other(s) Conundrums of analyzing mixed data Reporting challenges and opportunities with mixed data June 27, 2005 Academy Health ARM 2 WHY USE MIXED METHODS When you want to ask a question that has rarely been asked or has been asked with questionable results When you want the strength of multiple methods triangulation When some, and only some, of your variables are easily quantifiable at this stage of inquiry To “illuminate the black box” June 27, 2005 Academy Health ARM 3 TIMING/SEQUENCING Model One: Qualitative first, use results to refine study questions, think through sampling and develop measures Model Two: Qualitative and quantitative in tandem Model Three, Quantitative first, qualitative to explore results in depth and/or “in situ” June 27, 2005 Academy Health ARM 4 DATA COLLECTION Using qualitative approaches to confirm or further explore existing quantitative data (secondary) Using both qualitative and quantitative methods to explore the same variables (primary) June 27, 2005 Academy Health ARM 5 DATA COLLECTION Using qualitative methods to explore some variables (e.g. independent variables) and quantitative methods to explore others (e.g. dependent variables) Using qualitative methods to identify key variables for further study June 27, 2005 Academy Health ARM 6 USING ONE KIND OF DATA TO INFORM ANOTHER Even in a “purely” qualitative project, it is often wise to gather as much existing data about your site or respondent before you arrive In a more mixed project, qualitative data can help you identify and test alternative interpretations of quantitative findings, and vice versa June 27, 2005 Academy Health ARM 7 CHALLENGES OF ANALYSIS Mixed methods projects lend themselves to analysis along the way rather than just at the end Issue: do you analyze each kind of data separately and then look at how each sheds light on the other? Or do you analyze the data “together?” June 27, 2005 Academy Health ARM 8 CHALLENGES OF ANALYSIS The elephant in the room: what if your quantitative data imply one set of findings which are open to question when qualitative findings are examined? This has to lead to a thorough re-examination of your methods and experiences This is one reason for starting analysis early so you can continue to do empirical exploration June 27, 2005 Academy Health ARM 9 REPORTING Challenges: How much do you integrate? How do you integrate? Do you present all kinds of data to all audiences? June 27, 2005 Academy Health ARM 10 REPORTING Opportunities Using qualitative findings to illustrate quantitative results Even though some audiences claim they only want hard, numerical evidence, most people really connect with the stories available through qualitative studies June 27, 2005 Academy Health ARM 11