Skip to main content
Search form
  • 00:11

    My name is Phillip KAPker.I'm a post doctoral research associateat University College London School of Slavonicand East European Studies.I'm a political scientist and my research dealsmostly with presidents and the wayin which elected heads of state in European democraciesget involved into political decision making.

  • 00:34

    But I'm also interested in new parties and party systemchange.In this case study, we're going to lookat presidential activism, the useof formal constitutional powers by presidents,and the way in which we can study it using mixed methods.Specifically we're going to look at how

  • 00:57

    I study presidential activism with a combinationof quantitative and qualitative methodsand the specific challenges that I faced in doing that.To give you some context, the democratic transformationof Central Eastern Europe after 1989

  • 01:17

    produced quite a number of political systemsin which presidents possessed more than justa ceremonial role.they were given the power to veto legislation--so to send it back to Parliament and refuse their signature--or to play a key role in the appointment of governmentsafter and between elections.There's a great amount of scholarship

  • 01:39

    on how we can measure these different presidential powers.However, there's surprisingly little on how presidentsactually use them in practice.In my research, I wanted to find outwhy and when presidents used their powersand what factors can explain that presidents

  • 02:01

    decide to become active.I did not only want to give a general explanationof presidential activism, but I alsowanted to shed more light on the different and intricatemechanisms that underlie presidential decisionmaking and the use of presidential powers.

  • 02:25

    To achieve this goal, I decided to use nested analysis.Nested analysis is a mixed methods approachdeveloped by Evan S. Lieberman, which allows researchersto combine quantitative and qualitative methods into onecoherent research design in order

  • 02:46

    to achieve both generalizable results as wellas qualitative, in-depth insights into the subjectthat they're studying.In this case study, we're going to lookat how I use nested analysis to study presidential activismand discuss some of the problems and challengesthat I encountered.

  • 03:07

    The first step in nested analysisis usually a large scale statistical analysiswhich allows researchers to test general hypothesesabout relationships between variables.In my case, two of the hypothesesI was mostly interested in was whether the presidents who

  • 03:28

    are elected directly by popular votewould be more active than presidentselected by Parliament.Another hypothesis was that presidentswould be more active if they're in cohabitationor ideological opposition to the government thenwhen their own party participated in the cabinet.To test my hypotheses with this data,

  • 03:52

    I used both event count regression models and eventhistory analysis.This allowed me to analyze the general occurrence of vetoesas well as the time in between vetoesby individual presidents.

  • 04:13

    My result showed that popular presidential electionsand the presence of cohabitation were in fact positivelyand significantly associated with a more frequent useof presidential vetoes.However, there was still a few problems.On the one hand, the fit of some of my models was not perfect

  • 04:34

    and there were a few variables which I could not includein my statistical models.Also, how could I know that my results were not merelycorrelation, but also an indicationof causal mechanisms?The second step of nested analysisis specifically designed to deal with some of these problems

  • 04:57

    that I've just described.However, it was here that I faced even morecomplicated challenges.Based on the predictions of my statistical models,I chose 12 specific president cabinetpairings from four countries-- Estonia, Poland, Hungary,and Slovakia-- to serve as case studies for in depth analysis.

  • 05:22

    These case studies which allows me notonly to validate the results of my quantitative analysis,but also to find additional or alternative explanatoryfactors.But in order to conduct these case studies,I needed more and better data.Therefore, I decided to conduct elite interviews.

  • 05:43

    In particular, I wanted to interviewpeople who were directly involvedin presidential decision making and the useof presidential powers or people who could tell memore about the relationship and powerrelations between president and government and presidentand parliament.There were three groups of elitesI was most interested in.

  • 06:05

    First, these were presidential advisersand other high ranking members of the presidentialadministration.I also want to talk to former members of governmentor former members of Parliament whowere in power during the specific time periodthat I selected for my case studies.Last, I wanted to talk to national experts

  • 06:26

    and journalists who would help me to send the responses that Igot from my elite interviewees into a wider contextand validate them.There were a number of challengesthat I faced in arranging and conductingmy elites interviews.The most general problem was that not all people I contacted

  • 06:50

    were immediately available for interviewsor wanted to talk to me.While this is normal interview based research, in my case,it made things slightly more difficult.I selected very specific time periodsfor case studies and number of peoplewho could give me the information that Ineeded for my research was relatively limited.

  • 07:10

    To counter this problem, I among othersengaged in snowball sampling.I asked the elites who had agreedto be interviewed by me to recommend other respondentsor people who might have similar knowledge.I also tried to interview people who were active immediatelybefore or immediately after the time periods

  • 07:31

    I was interested in.Despite these challenges, I was stillable to complete 65 elite interviews overall.Another problem was language.Political elites are generally more likelythan other respondent groups to speak a foreign language.However, in my case, not all of the people I wanted to talk towere actually fluent in a foreign language.

  • 07:53

    This was overall less of a problem in Estonia and Hungarywhere most elites were fluent in English or Germanand I could interview them in these languages.In Poland and Slovakia, however, elitestended to lack foreign language skills.While I speak Polish and could thus interview Polish elitesin their language, in Slovakia, it

  • 08:15

    meant that I could simply not interview as many peopleas I wanted to.And being a PhD student at the time,I also lacked the funds to hire interpreters.To deal with this problem, I eventuallydecided to interview more expertsand to use more archival materialto at least approximate the information I could have

  • 08:36

    gained from elite interviews.A third challenge was the recording of interviewsand subsequent attribution of responses.Some of my respondents were quitewary to have their interviews tape recorded.On the one hand, it was due to the factthat we talked about their personal political viewsor informal formal mechanisms of political bargaining, which

  • 08:59

    they did not want to be made public.On the other hand, there had beena number of political scandals in my case study countrieswhich involved, albeit secret, wiretapping of politicians.Eventually, I decided to take most of my notesfrom interviews by hand in order to increase rapport

  • 09:21

    and establish a relationship of trust with my respondents.Furthermore, I assured confidentialityfor my respondents, which meant that I was not coaching thempersonally or attributing anythingthey had said directly to them.Unfortunately, this meant that I could not use all information

  • 09:42

    that I'd gained in my study, as includingit would have meant it violates this agreementof confidentiality.After finishing my interviews and writing my case studies,I still have to bring together the quantitative andqualitative results.The case studies also showed that the causal mechanismsthat I assumed to be behind these results actually

  • 10:05

    existed in practice.However, I also found that there wereresults from the statistical analysis whichwere caused by different factors that I had initially thought.While in this cases the qualitative evidencethat I gathered was not enough for further generalizations,the in-depth study of the mechanisms involved now

  • 10:27

    provide the starting point for future studies on the subject.Furthermore, the synthesis of my resultsalso showed variables that were notincluded in my theoretical or statistical models.For example, I could show that presidents use their powersmore often when there were deep divisionswithin or between government parties.

  • 10:49

    And they used their veto more often irrespective ofwhether they were in a friend relationshipwith the government or in cohabitation.In summary, my research could, among others,show that directly elected presidents use their powersmore frequently than presidents elected by Parliament.

  • 11:12

    And it could also test a number of other hypothesesadequately for the first time.Using nested analysis, I could therebynot only provide a general test of relationshipsbetween variables, but also provide in-depth insightsinto presidential politics and the decisionmaking mechanisms involved.However, in each step of the research process,

  • 11:35

    I faced certain challenges that could not be completely offsetby the combination of methods.Nevertheless, mixing methods can provide a solid foundationto reach valid and reliable conclusions.For this, the relation between the different methodologiesand the role of each method must be clearly defined

  • 11:57

    from the start.Researchers must also embrace the respective ontologicaland epistemological assumptions thatare associated with each methodological strand.Only when both approaches are recognizedas equally contributing to the results canthe full power of mixed methods approaches unfold.


Dr. Philipp Koeker describes the benefits and challenges of using mixed-methods research. His study investigated how and why Eastern European presidents use their veto powers, mixing statistical analysis and elite interviews.

Looks like you do not have access to this content.

Researching Presidential Activism Using Mixed Methods

Dr. Philipp Koeker describes the benefits and challenges of using mixed-methods research. His study investigated how and why Eastern European presidents use their veto powers, mixing statistical analysis and elite interviews.