Skip to main content
Search form
  • 00:00


  • 00:13

    INTERVIEWER: Libby Bishop, thank youvery much for talking to me.The first thing I wanted to ask youis what is secondary analysis of qualitative data?

  • 00:25

    LIBBY BISHOP: It can sometimes be a bit difficult to define,but I use an informal definition that secondary analysis isusing data for any purpose that isdifferent than the original purpose for which the data wascollected.And it can be done in a lot of different ways.It's a relatively new area, particularlyfor qualitative data.It's quite a bit more establishedfor quantitative data.

  • 00:48

    LIBBY BISHOP [continued]: But it can be done in ways such as taking an existing data setand asking a new question about it, which is somethingI did as a small example of working with an existing dataset where the original purpose was an investigation of healtheffects, generally a health study.And I used it to look at the examples of occurrencesof talks-- of conversation about food, which is a part of healththat the original researchers had notdealt in any detail with the food topic.

  • 01:19

    LIBBY BISHOP [continued]: Most people tend to reuse data for topics or themesthat are broadly similar to how the data were reused.Not always, but if the original studyhad to with collecting informationabout class or something, then perhaps a related topicmight come up.In other cases, it really, again,can be a completely different topic.

  • 01:41

    LIBBY BISHOP [continued]: Other sorts of uses of the data canbe very constructive to do comparative work,so you might have comparative workcross-sectionally geography.You might be collecting your own data in one geographyhere in the UK and want to do comparative workinternationally or comparison to another region in the UK,and increasingly, of course, with the idea of temporalityand time across cross-periods, you can alsodo longitudinal comparisons.

  • 02:09

    LIBBY BISHOP [continued]: So someone will have collected-- wehave huge quantities both in the USand in the UK of very valuable data collected in the '60sand '70s, and now people are revisiting that dataand trying to make it more contemporary.It's also an example that sometimes peoplecome revisit their own data, so it doesn'thave to be a new researcher.People go back and look at their own data from decadesago and sometimes find very, very new things in it.

  • 02:35

    LIBBY BISHOP [continued]: Increasingly as well as the research aspects of it,you can use readers' data to study methodology.It's a fantastic way to get below the surface of what'shappened in a study.So you might read a published account,and you'll get the official version, if you will.But sometimes reading the actual data,reading the flow of an interview,you'll get a sense of how the data collection actually wentand get a much better and deeper understandingof the methodology.

  • 03:02

    LIBBY BISHOP [continued]: I sometimes make a comparison with doing a literature review,but it's a much deeper kind of resultthan you can get with a literature review.And the other vitally important area to remembersis it's a very valuable resource for teaching,so being able sometimes new researchers,in particular, new researchers taking up teaching activities,only have a small set of data.

  • 03:23

    LIBBY BISHOP [continued]: They might only have their own area of datato share with students using archives like the UK DataArchive or new qualitative archives in Ireland,and in Europe, even, there's a wealthof resources that can be used for teachingboth substantive areas and methodology.

  • 03:40

    INTERVIEWER: Can I just ask at this stage,how is what you do in terms of secondary dataanalysis different to what, say, historiansare doing in revisiting data?

  • 03:51

    LIBBY BISHOP: I actually don't think that thereis a great deal of difference.It's not surprising that a coupleof the pioneers of doing secondary analysis, people--just, there are many names I could pick.But Paul Thompson's major Edwardian studywas founding one of the founding studiesfor qualitative analysis at Qualidata.Mike Savage, of course, is a key one who's redone data.

  • 04:12

    LIBBY BISHOP [continued]: These people come out of oral historyor historical traditions.So many-- I should be a bit unkind,but sociologists sometimes think they'rediscovering brand new things that, in fact, historiansare well-practiced in and already doing,and it's a highly enriching cross-fertilization, I think,to see the collaboration across the social scienceand the historical methodology.

  • 04:34

    INTERVIEWER: OK.So it's nothing really to do with the types of datathat you're working on.They're quite distinguished.

  • 04:40

    LIBBY BISHOP: The distinction wouldbe-- so much more data in social sciences is actively collected.So historians will-- much rarer for historians willdata have been produced from a historical record.Now you can argue when someone famous is writing a diary,is that really a document for their own personal use,or do they actually have in mind that itmight become a significant historical document,but most of the kinds of materials that historians usehave not been explicitly actively generatedby a research process like an interview elicitation process.

  • 05:16

    INTERVIEWER: So who does secondary data analysis?

  • 05:18

    LIBBY BISHOP: Well, increasingly, birds do it,bees do it, pretty much everyone is doing itin one form or another.I mentioned a couple of names, but some of our starsare doing secondary analysis.People have been doing it quite awhile now.There's Mike Savage, who's also here,and, in fact, just released his book hereat the Research Methods Festival thatis an example of using secondary analysis.

  • 05:41

    LIBBY BISHOP [continued]: So his work I found particularly innovative,because he drew on a very large [INAUDIBLE] setof multiple collections of data.So oftentimes, a reuse project mightfocus on only a single collection, whereas Mikedrew across a really wide range of periods and timesand different kinds of data actually drawingon even across the qualitative quantitative boundaries.

  • 06:06

    LIBBY BISHOP [continued]: So we used affluent worker data, which is housed at Qualidata.He also used the mass observation directives,which are a particular kind of datawhere people are asked to write on themes and produce data.He's investigated these issues of identityand particularly people's class identity over quitea long historical stretch now by beingable to draw these historical perspectives.

  • 06:28

    LIBBY BISHOP [continued]: Another good example, when I'm working with this teamquite closely now, also people here at the Research Festival--Graham Crow and Dawn Lyon, and theyare reusing data collected by Ray Paul in the 1960s and '70s,I believe.And Ray did a large number of projects, but one of themfocused essentially a community studydone on the Isle of Sheppey, which isan island southeast of the UK.

  • 06:54

    LIBBY BISHOP [continued]: So it investigated many aspects of community life in Sheppey,but one of the bits of data collection that Ray didwas draw on essays that young peoplewere writing about what they imaginedtheir future lives to be.And parts of Sheppey, anyway, wereregarded as somewhat impoverished, socially excludedareas, so part of this investigationhad to do with how people growing upin these circumstances imagine their futures.

  • 07:21

    LIBBY BISHOP [continued]: So Graham and Don are coming along and doingmany different kinds of work with this,but one of the exercises that they're doingis reintegrating themselves into the Sheppeycommunity, into the schools, and indeed,having current young people write an almost identical kindof essay, so they will now have a very nice comparison of howyoung people have imagined their futures across this time frame.

  • 07:43

    LIBBY BISHOP [continued]: The Timescapes Project, which is a qualitative longitudinalproject based at Leeds, one that I also work with,has got its projects that are studying the life force.And there are a series of projects,and there's a pair of projects that are studying parenthood,so one on motherhood, one on fatherhood.A couple other pairs of projects are focused on older generationand grandparenting issues.

  • 08:05

    LIBBY BISHOP [continued]: So another way it's a bit differentbut that secondary analysis is being doneis even within the auspices of this ongoing live project,if you will, there's cross-fertilizationwithin these teams.So you might call it primary, but on the other hand,these people are using the motherhood projectas drawing on data from the fatherhood projectand vice versa.

  • 08:25

    LIBBY BISHOP [continued]: So it's kind of a contemporaneous formof secondary analysis.So I think probably the most exciting thingabout people doing it is now increasingly,I'm opening up conference programsand looking at particular sessionsand finding things on secondary analysiswhere I don't even know any of the peopleor any of the topics, because it's proliferating quickly.

  • 08:47

    LIBBY BISHOP [continued]: It used to be that I sort of knew most of the players,and now it's quite exciting to seethat it's growing so quickly that it'sgetting a bit difficult to stay on top of it all.

  • 08:58

    INTERVIEWER: What might be the problemsof using secondary data?

  • 09:04

    LIBBY BISHOP: I think, actually, one of the problems,and I would say this in particular to younger studentsor postgrads thinking about exploring this as a method,I actually think it's an image problem as opposedto a real problem.So it's what I call the poor relation problem,and that is that somehow maybe because of the name,secondary analysis, it's also thought of as second class,second tier in some way.

  • 09:32

    LIBBY BISHOP [continued]: And there is a little bit, I suppose,still I think of a bias that would say that somehow primaryis privileged, that primary data is always better.I just don't think it's true.I think you certainly talk to any historian,as you mentioned before.How could any historian think that somehowversions of secondary data would not beequally valid or rich to use?

  • 09:58

    LIBBY BISHOP [continued]: I personally am trying to even get awayfrom the term secondary analysis, although it'svery hard to kill a term, unless you have a new termto replace it with.But I increasingly start trying to use languagelike reusing data, which is a bit clunky,revisiting data, maybe better.I tried recycling data.It sounds warm and friendly and climatechange friendly and whatnot, but typically reusing datais the term I will use.

  • 10:25

    LIBBY BISHOP [continued]: So it is a bit of an image problem of tryingto get away from that.The biggest challenge I think that probably comes upwith reusing data usually focuseson the issue of context.And by context, what is typically meantis the idea that there is something precious, invaluable,and unreproducible about the original contextand situation of the data collection moment,if you will, and there are hot debates about thisin terms of various sub0branches of methodology and so forth.

  • 11:02

    LIBBY BISHOP [continued]: But the idea is that even-- take our example.There are important contextual informationthat is happening in this particular interview.The room we're in, the temperature, body language,the fact that another person is watching this interview-- theseare all things that somebody would need to know,even analyzing this as an example of a formal interviewsituation, something like that.

  • 11:23

    LIBBY BISHOP [continued]: Now within a secondary analysis, evenif you have the transcript, quite frankly,even if you have the audio, even if youwere to have a video of this video session,it wouldn't be the same.It wouldn't represent-- things it might've missedwould be your briefing of the original setupfor the interview.It would lack context about why we'redoing this whole series of interviewsin the first place, any previous contacts that we might havehad, early images you might have of me, me of you,these sorts of things.

  • 11:54

    LIBBY BISHOP [continued]: So all these things, it is quite true,cannot be fully reproduced as context in doing secondaryanalysis.But then, the question is, do they need to be?Must that information be present for anybodyto make any good use of data?So I have two answers for that.One is we're getting better, actually,about collecting contextual information.

  • 12:18

    LIBBY BISHOP [continued]: So that's one answer is, OK, if the context is missing,try and fill it in.And all of the qualitative data archiveswork quite hard to do that, so theywould collect multiple genres of the data-- audio, image,photos, et cetera.They would collect background informationon how the project came into being, how it's funded,reports of the setting, any extensive kindof contextual information.

  • 12:45

    LIBBY BISHOP [continued]: So that is helpful, and I think that that'sa very useful way forward.I quite still, though, agree with the critics that says,it doesn't fully answer the question.It doesn't solve the problem of full reproducing the context.But I also think you don't have to fully reproducecontext for data to be fully reusable and useful in very,very different ways, because, in fact, any time a researchercomes to new data, that is part of their jobis to gather as much context as they effectively can,use good judgment to make interpretationsin recontextualizing that data for the current setting,and that's a term from Nieve Moore, who'swritten on exactly this issue.

  • 13:28

    LIBBY BISHOP [continued]: So in substance, you can't analyze datawithout context, because simply by in the processof doing the analysis, you are recreating the context.And that is part of the job of somebodydoing secondary analysis, reusing data ratherthan primary.And again, this is, of course a whole experience and kindof technique and tool set with which historians arequite expert and very familiar.

  • 13:51

    LIBBY BISHOP [continued]: And I think we can learn a great dealfrom that in terms of bringing our tools to data reuse.

  • 13:57

    INTERVIEWER: Now can I ask when secondary data mightbe preferred to original data.

  • 14:05

    LIBBY BISHOP: There are several situationswhen that might be the case.One is, think about any sorts of instanceswhere data already exist, and thereis anything particularly that burdensome about collectingthe data again.So think about vulnerable populations of various kinds,so elderly, ill, people who've been over researched,and I'm afraid, increasingly, I'mlistening to, again, some of the talks around the festivaltoday and yesterday.

  • 14:35

    LIBBY BISHOP [continued]: We know this is the case.You're seeing response rates dropon surveys, difficulty in accessing certain populationswho are deemed challenging.So this idea of burden is important in getting away from,and any time you can find data thatmight fit your requirements without burdening a new group,that is quite useful.

  • 14:56

    LIBBY BISHOP [continued]: I think sometimes people don't appreciate that another wholevalue of reusing data is although reusing datadoesn't-- it isn't always faster and easier.I'll come back to that point in a minute.But somehow it does allow your mind to shift and pay attentionto other aspects of research design and researchmethodology, so as opposed to havingto focus a huge amount of effort on the particularsof recruiting, for example, you still, of course,have to think deeply about samplingand why you're choosing some people and not others.

  • 15:26

    LIBBY BISHOP [continued]: But you can focus on design and research questions.You can focus on sampling design,and again, I'm just pointing out Mike Savage's work, I think,is brilliant on this, because he hadto think quite deeply about not only which archives to pickand which collections to pick amongst those archives,but there was a vastly greater quantity of materialthan he could use.

  • 15:48

    LIBBY BISHOP [continued]: And so within any archive or collection,he had to figure out effective ways of sampling,and I think that is an important thing for usto pay more attention to.We sometimes, I think, shirk that a little bitas qualitative researchers.And also, just getting down to the deeper layersof what I would call research analysis,which are things like thinking through implicationsfor theoretical development and thinking through thingslike causal mechanisms.

  • 16:14

    LIBBY BISHOP [continued]: I will personally go on record as sayingI don't think we have done enough of this,and I don't think we're doing it well enough.And again, other people at the conference, Julia Brannanmade these points yesterday that she thinks that there actuallywill be an increase, a growth in reuse of data,and then, although, there are some downsides of that.There's some key benefits, and one is exploring these areas.

  • 16:35

    LIBBY BISHOP [continued]: And that raises a point is that Ithink it is a bit tricky right now, because oneof the reasons, of course, there mightbe more pressure to reuse data in the near futureis that the money for collecting new data may be scarceand very hard to find.So this was the point that Julia made.So I think that's true, and I will sadly enjoy the benefitthat archives might get.

  • 16:60

    LIBBY BISHOP [continued]: If, in fact, people come to us and need to reuse our data,I will take advantage of that.But there are a couple of important points to make.One is that it's still a very importantto remember that reusing data is not always faster and cheaper.You still have to do your homework.You have to investigate that data,and you have to read data.You have to explore not only the context that you were given,but try to find the additional context.

  • 17:22

    LIBBY BISHOP [continued]: You still have to pay attention to ethical issues.Some of them will have been somewhat addressed,but you may well have to address them again or thinkthrough them also.So anybody who's thinking that somehow using existing datais the budget model should think again,because I'm not positive that that is necessarily the answer.I also, of course, do not want to see the availabilityof existing data used as a justificationfor cutting primary data collection.

  • 17:51

    LIBBY BISHOP [continued]: We need both, and that is not goingto be an easy argument, I know, to push in the near future.But nonetheless, they cannot be pitted against each other.They're very complementary, and theyneed to continue to be done that way.

Video Info

Publisher: SAGE Publications Ltd.

Publication Year: 2011

Video Type:Interview

Methods: Secondary data analysis, Archival research, Qualitative data collection, Historical research

Keywords: comparison; convenience; essays; fatherhood; finances; historical records; issues and controversies; motherhood; teaching; vulnerable populations ... Show More

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



Dr. Libby Bishop defines secondary data analysis as reusing an existing data set to pursue a different research question. She explains that it is very similar to the research methodology used in history, because few historical documents were created expressly for research purposes. Bishop also highlights the benefits of using secondary data analysis to study over-researched and vulnerable populations.

Looks like you do not have access to this content.

What is secondary analysis of qualitative data?

Dr. Libby Bishop defines secondary data analysis as reusing an existing data set to pursue a different research question. She explains that it is very similar to the research methodology used in history, because few historical documents were created expressly for research purposes. Bishop also highlights the benefits of using secondary data analysis to study over-researched and vulnerable populations.

Copy and paste the following HTML into your website