Skip to main content
Search form
  • 00:09

    [Evaluating Project Servator Using Quantitative Methods]

  • 00:17

    PAUL BAINES: Hi, my name is Professor Paul Baines.[Paul Baines, PhD, Professor of Political Marketing,Associate Dean, External Relations, School of Business,University of Leicester]I am Professor of Political Marketingand Associate Dean, External Relationsat the University of Leicester School of Business.[How did you become interested in your research subject?]

  • 00:37

    PAUL BAINES [continued]: My primary research interest is around political marketing,marketing more generally, and propaganda.And specifically, some of the projects that I'm working onare around the use of guilt and fearas appeals in marketing communications, but alsoaround market segmentation and how to devise and organize

  • 01:02

    PAUL BAINES [continued]: audiences into specific types and groups of peoplein order to tailor marketing appeals, and products,and services to those particular groups.[What was the objective of the Project Servator researchstudy?]

  • 01:24

    PAUL BAINES [continued]: The project that I'd like to speak about todayis called Project Servator.And this is a national police initiative to effectively sweepup low-level crime, any kind of crime, really,but also act as a deterrent for--in terms of counter-terrorism.And it incorporates unpredictable police

  • 01:46

    PAUL BAINES [continued]: deployments--that means police units walking around particular areas,particularly around iconic sites like, say,for example, in London, St. Paul's Cathedral, or in York,York Minster, or the Bull Ring in Birmingham.And these iconic sites are effectivelypatrolled in an unpredictable way.

  • 02:07

    PAUL BAINES [continued]: They don't just look at those sites, but lots of othersas well.And this is incorporated into a wider communication campaignwhich aims to encourage the reportingof suspicious activity by the general publicand at the same time increase their vigilanceso that they are aware of their surroundings

  • 02:28

    PAUL BAINES [continued]: and potential risks, whether that's crimeor whether that's terrorism.

  • 02:35

    HELEN ISAACS: The impact of getting this campaign wrongwould have been really severe for the force,and for our community, for the confidencethat our community has in the City of London Police,so we had to get it right.We had to design a message that spoketo the different audiences in different ways, whichwas very challenging.But what we did was we designed messages

  • 02:57

    HELEN ISAACS [continued]: that conveyed to the public our need for them to be vigilant,our need for them to phone us if they saw anything they thoughtwas suspicious, and also the reassurance that we wantedthem to feel from the operations that we're doing on the ground.We used a range of different mediums.So we used PR articles.We used posters.

  • 03:17

    HELEN ISAACS [continued]: We used mediums in stations, in crowded places.And we looked at messages that would mean somethingdifferent to each audience.So for example, we used a messageabout a plainclothes police officer potentiallybeing in a crowd of people and asking the question,can you spot the plainclothes police officer?

  • 03:38

    HELEN ISAACS [continued]: And of course, it's very difficultto do that in a crowd of people.For the public, that would be a message hopefully reassuringto them to say that even though you can't see a uniform,there are police out working today.For the would-be terrorist, that wouldbe a message that would deter themfrom coming into the City of London,because they can see that, even though they might not

  • 03:59

    HELEN ISAACS [continued]: be able to see a uniform, it doesn't meanto say that there isn't a plainclothes operation takingplace in the area.We had another poster that we usedas well, which was one about the public being our extra 300,000pairs of ears and eyes, and the fact that we loverush hour for this very reason.Now for the public, that hopefully

  • 04:21

    HELEN ISAACS [continued]: makes them feel that they are part of the operationsthat we're running and impresses on themthe importance that we place on their role in the securityof the City of London.For the would-be terrorist, that would be a deterring messageto show them that it's not just the police that are involvedin this, it is the whole community of the City of London

  • 04:41

    HELEN ISAACS [continued]: who are watching out for anything that theymight feel is suspicious.

  • 04:44

    PAUL BAINES: So that's the project.It's quite an exciting project.It's an unusual, if you like, contextfor marketing It's social marketing,effectively, how the police use marketing in orderto keep us all safe.So I think that it's a very worthy project.And I'd like to tell you a little bit more about it.

  • 05:06

    PAUL BAINES [continued]: [What types of data were you able to collect?]The type of data that we gathered on this projectwas done through a street intercept technique, really.So it was a survey that was conducted in the street

  • 05:28

    PAUL BAINES [continued]: with members of the public who'd comeinto contact with the police.And what we wanted to ask them was several things.Were they reassured?Were they alarmed?Did they see particular assets that the police had placed?So some of those assets might be the communications,the posters, the airframes that they had,

  • 05:51

    PAUL BAINES [continued]: the handbills that were given out,the tunnel in train stations, had theyheard these kinds of elements of communication?Social media, had they heard about Servator on social media?So we asked about various communication devices,had they heard of Servator or had they seen this message?

  • 06:13

    PAUL BAINES [continued]: What we also asked around was, how preparedwere they to report suspicious activityand how prepared were they to report unattended items?And then we asked them, did seeing the police deploymentand the communications, but obviously separately,affect whether they were more or lesslikely to report suspicious activity

  • 06:34

    PAUL BAINES [continued]: and report unattended items?And we asked a whole bunch of demographic questionsaround who these people were.And we asked various questions to understand how likely theywere to report suspicious activity,and any barriers that they might feel to reportingsuspicious activity.And what this helped us do was build

  • 06:55

    PAUL BAINES [continued]: a picture up of the way in which people are prepared to reportsuspicious activity, report unattended itemsin the circumstances where a crime might be happeningor a terrorism incident might be taking place.[What ethical issues did you face?]

  • 07:24

    PAUL BAINES [continued]: There are always ethical considerationsin any piece of work that's undertaken.In this case, the research was undertaken in the streets.It was undertaken by a market researchagency who was aligned to the Market Research Society.And so the staff that had--were undertaking the work were part of a professional body.

  • 07:46

    PAUL BAINES [continued]: And that means that they had to adhere to the Market ResearchSociety's guidelines on data collection and other issues.So there are a variety of ethical issues.One is about whether you're reporting the right kindof data to the client.So you have a duty to report and behonest with the client about what it is

  • 08:08

    PAUL BAINES [continued]: that's coming out of the data.But the police were very good in that regard,because they genuinely-- this was something very newthat they had not done before originally.And what they wanted--so this was something they really wanted to know.It would genuinely help them improve their relationshipwith the public and help them police better,

  • 08:28

    PAUL BAINES [continued]: because they would receive information from the publicthat they might not otherwise have got if this project hadn'ttaken place.There are other ethical issues.There are ethical issues in relation to anonymity.We asked in one particular survey-- we've done a lot,so that's why I'm hesitating a little bit.

  • 08:49

    PAUL BAINES [continued]: But the average sample size was somewherebetween 600 and 1,000.And that's an appropriate number for this type of activity,because we wanted to look in detailat specific groups of people.And so you've got the issue of individual respondentswould not want to be identified with their answers.So you might have some, for example, somebody who says,

  • 09:11

    PAUL BAINES [continued]: you know, I have no interest in reporting suspicious activityto the police.And I'm not very confident in the police.And there is a small number of people,very small number of people who have that kind of attitude.But equally, you got lots of people who were veryenthusiastic-- and most are--about reporting activities that they see,criminal activities and terrorist activities,

  • 09:31

    PAUL BAINES [continued]: to the police.But they wouldn't necessarily want to be identified.So it's very important, I think in research, survey research,that you allow the respondents that privacy and anonymity.So that's another important issue.How you word certain questions is important.

  • 09:54

    PAUL BAINES [continued]: This is less ethics and more about good questionnairedesign.But you should ask questions that are very clear.And the clearer they are, especiallyabout an important topic like this, then the more likelyyou are to get that kind of honesty and real truthfrom the respondents themselves.So those are the key issues.

  • 10:18

    PAUL BAINES [continued]: And the issue about reporting backto the client is a similar issue,is that if a respondent asks who is commissioning the research,you do have to let them know.So we would have to say, well, this is something that we'redoing for the police.And so these are the key issues that we considered

  • 10:39

    PAUL BAINES [continued]: in the design of the research.[How did you analyze your data?]So the data were collected in the sensethat the 600 to 1,000-- there were different surveys,because we did a number of them.

  • 11:00

    PAUL BAINES [continued]: Once that data were collected from the street interceptinterviews, we then had to put that into, effectively,into a spreadsheet.And that data would effectively become a series of categories,often enough.So certain types of data would be

  • 11:23

    PAUL BAINES [continued]: made into numerical categories.Other types of data might be providedin its complete verbatim form.The verbatim form would be necessary in open questionsfor complete analysis, for qualitative analysis.The numerical questions would be importantfor statistical analysis.

  • 11:45

    PAUL BAINES [continued]: And as far as the--once the data in the spreadsheet was set up,that would then be imported into SPSS.And basic analyses would be undertakenin terms of frequency of responseto certain questions, standard deviations, means, averages,and so on.So on one of our questions, it was a five-point scale.

  • 12:07

    PAUL BAINES [continued]: And it was around to what extent areyou prepared to report suspicious activity, veryprepared and not very prepared kind of thing, five scales,midpoint.That kind of data would be stored in termsof one, two, three, four, five.And then the responses would be transformedinto that kind of data, which is known as ordinal data.

  • 12:32

    PAUL BAINES [continued]: So you've got nominal, ordinal, interval kind of data.And you would put it into that form.Once it's in that form, it's in SPSS,you can do lots of other things.SPSS is a very powerful engine.And so what we-- so aside from those frequency evaluations,we would also do something known as cross tabulation.And that's why you look at whether certain questions are

  • 12:55

    PAUL BAINES [continued]: related to each other.So what this could allow us to do is to look at say,are males and females saying different thingsabout the extent to which they are prepared to reportsuspicious activity?Or certain age groups, do they differ?Cross tabulations give us some indication,but they're not usually--they can be statistical.

  • 13:17

    PAUL BAINES [continued]: So you normally want to run statistical analysesto determine various things.Are there relationships?You might look at are the differences, in which casewe'd run t-tests.Or we might want to look at is one answer associatedwith another answer for nominal data?That's categorical data.

  • 13:38

    PAUL BAINES [continued]: And then we'd use chi-square tests and that sort of thing.So these things were done by the market research agency,actually.And I came in and did something extra.I looked at the extent to which weresome of the communications making the respondents moreor less likely to report suspicious activity?

  • 14:00

    PAUL BAINES [continued]: And I did that using a logistic regression.And logistic regression is an amazing technique,a very powerful technique.What it allows you to do is determinewhether some of the independent variablesare contributing, the extent to whichthey're contributing to your dependent variable.

  • 14:20

    PAUL BAINES [continued]: In our case, most of the time--I did various analyses.But the key things that we were interestedin as a dependent variable was the extentto which people were prepared to report suspicious activity.And logistic regression uses a binary variable,prepared or not prepared to report suspicious activity.

  • 14:45

    PAUL BAINES [continued]: You can use other techniques, but that was the oneI used most of the time.And then you put all these variables in.So we could see whether some of the communicationswere actually improving people's abilityor making them more likely to report suspicious activity.We could see whether the posters were working,whether the dogs were working, whether the armed officers

  • 15:05

    PAUL BAINES [continued]: with tasers, with firearms were working.That allowed us effectively to evaluate the effectivenessof the different communication techniques and the policedeployment.We also did something called a cluster analysis.So once I'd done that we could find outwhich of these communication techniqueswere enhancing people's likelihood of reporting

  • 15:26

    PAUL BAINES [continued]: suspicious activity, what we could also dowas group some of those people thatwere more or less prepared to report suspicious activity.And we could see whether some peoplewere more or less prepared.And in one particular analysis, what we foundis that roughly 50% were completely and utterlyprepared to report suspicious activity.

  • 15:46

    PAUL BAINES [continued]: This group of people were highly vigilant, watching out,very compliant the message, really interestedin doing what they police messages were saying.Of the other 50%, half of them, so 25% of the total sample,were totally clueless.They didn't see any of this communication.They didn't really know what was going on.

  • 16:08

    PAUL BAINES [continued]: They could be persuaded to report suspicious activitymore, but they really didn't take much noticeof the communications.The other 25% were willfully not preparedto report suspicious activity.We don't quite know why.So these were people who usually sawthe message, so we know they were aware of it,

  • 16:32

    PAUL BAINES [continued]: but they weren't prepared to respond to it,probably thought that it was--that this kind of communication was unnecessary, and so on.And they tended to be a bit younger, quite a lot younger,on average, than the other groups.So we think that it might be at least partlyto do with the idea of a kind of authority message

  • 16:53

    PAUL BAINES [continued]: that younger people are less prepared to listen to.So the cluster analysis is of powerful technique.In this case it uncovered three groups.Other times it's uncovered four, sometimes two.But closer analysis allows us to determinethe extent to which different groups arerelated to different variables.In this case, it allowed us to determine

  • 17:14

    PAUL BAINES [continued]: whether different groups were responding in different waysto reporting suspicious activity.[What tools & resources would you recommend for similarresearch?]For the logistic regression, I used different techniques.

  • 17:37

    PAUL BAINES [continued]: Originally, initially, I used R. And R is a very powerful,it's a very quick way of doing analysis.But you've got to be able to program.It's effectively statistical computing.And so it's not for the faint hearted.Learning it is quite hard.You would need to go on a course to understand it.

  • 17:59

    PAUL BAINES [continued]: I mean, you would do for SPSS, but the course for Rwould take you much longer, I think.So SPSS is an easier tool for somethinglike straightforward logistic regression.You could also do-- you could do anything in R.It's such a powerful technique.And it's free.And it's open source.And that's its attraction, really.Can do almost anything in R. And it's

  • 18:24

    PAUL BAINES [continued]: updated regularly, and so on.So it's a very useful way of doing it.But actually, I also used SPSS.And SPSS is a bit more user friendly.And it spits out certain things that you'dhave to program how to do, because it doesn't spit themout with an actual--necessarily with the particular commands.So SPSS is probably much more friendly for student projects.

  • 18:50

    PAUL BAINES [continued]: The cluster analysis, I could have done that in R. Icould have done k-means analysis,cluster analysis in SPSS, but I chose to use Latent GOLD,because Latent GOLD does something called latent classcluster analysis.And that's a more robust technique,much stronger technique to identify clusters.

  • 19:14

    PAUL BAINES [continued]: You can determine the number of clustersmuch more scientifically, if you like, through latent classcluster analysis than you can with k-means.And it's much easier to identify the numberof clusters that come out.So latent class cluster analysis is also available in R,

  • 19:34

    PAUL BAINES [continued]: but I chose to use Latent GOLD, again,because the graphical interface for itis much easier to interpret.[What were your research findings and how did you decideto present them?]So once you've got the data, it'svery important to determine how to use that insight that

  • 19:58

    PAUL BAINES [continued]: derives from the data inside the organization that'scommissioned the research.I produced a report for the police.The report outlined in detail the various clustersof groups and how they were respondingand also the different factors that made communicationabout reporting suspicious activity and unattendeditems more likely.

  • 20:19

    PAUL BAINES [continued]: That was disseminated to the national team and ProjectServator in the City of London.They would then have used that datato talk in training to their different various police forcesthat are using Project Servator and this technique.

  • 20:40

    PAUL BAINES [continued]: And so it would help.So there will be a cascade methodof briefing for the different commandersof the different Servator forces and different policeauthorities.But what we also did was use that data to communicatein various ways about the project at various fora

  • 21:00

    PAUL BAINES [continued]: that incorporated police, and military,and so on, in order to raise the public'sand associated parties, people interested in this technique--so we talked at the Countering Violent Extremism conferenceand gave an outline of some of the datathere, for example, at Cranfield University, whichruns every year.

  • 21:22

    PAUL BAINES [continued]: And we've in the past also presentedto other partners within the home office about this data,because this technique in Servatoris also being considered by the border force and by--so in ports and elsewhere.So it was communicated not just to the police,but also to interested partners, and to some extent, members

  • 21:46

    PAUL BAINES [continued]: of the general public that are concerned with someof these issues, and other academics whoare considered or are interested in these issues.[What unexpected challenges did you face?]There were no specific methodological challenges

  • 22:07

    PAUL BAINES [continued]: that we came across in terms of-- or unexpected eventsin collecting the data, but one important issueis, of course, the fact that these dataare associated with the police.And whilst in this case, we didn't necessarilyhave to apply for home office permission for this typeof data, other types of data looking at offenses

  • 22:34

    PAUL BAINES [continued]: that people undertake and so on usually are--usually do require a license from the home office.And in one particular case when we looked at some data from--some similar type of data from the ports,from deployments that were being done in ports, that was

  • 22:55

    PAUL BAINES [continued]: overseen by the home office.And there were certain permissionsrequired in looking at that data in the home office,and obviously certain permissions in the wayin which that data are used.And of course, when I mentioned earlier in our interview,I was talking about how there is a duty to the client

  • 23:15

    PAUL BAINES [continued]: to report to them properly.And obviously, one thing I didn't mentionis the duty not to make that data available to anybody elseif they don't want me to.But in this case, there is nothingspecifically unexpected, because everybodyagreed with all these things.But there can be an expected challenge

  • 23:36

    PAUL BAINES [continued]: is more generally in relation to sample size, for example.Sometimes your interviewers can't find as many peopleas they need, in which case you justhave to add on an extra day and do more interviewing,but that would come at a higher cost.It might be that when you do the analysis,you don't find some of the things

  • 23:56

    PAUL BAINES [continued]: that you thought you might find.And if that happens, so be it.As long as your hypotheses were robust,and based on literature, and had some theoretical structure,if what you find is counter to that,that may be because the phenomenonis genuinely different from what it says in the literature.

  • 24:20

    PAUL BAINES [continued]: Equally, it could be just that your sample is notnecessarily robust.And so there might be a problem with your data.The only way you could solve that problemis to run this survey again and see whether youget the same kind of findings.And we've run this survey so many timesnow in different police forces that we

  • 24:41

    PAUL BAINES [continued]: know that some of the findings that we haveare being replicated in these other contexts,and not always in exactly the same way, but broadly very,very similar.So we know that our findings are robust.[What tools and resources would you recommend?]

  • 25:06

    PAUL BAINES [continued]: Well, I think-- I mean, there are a number of great booksout there.Some of the thicker tomes on market research I thinkare absolutely fundamental for the marketing student.Examples would be Malhotra's book on market research, whichI think is--it's so detailed that there is nothing you couldn'tfind, more or less, in there.The Herrera et al. text on multivariate methods--

  • 25:28

    PAUL BAINES [continued]: brilliant, brilliant book for the more advancedstatistical analysis.In terms of computing packages, thereis a great book by Julie Pallant calledSurviving SPSS, which I think--sorry, SPSS Survival Guide, which I thinkis a really, really useful book on undertaking analysis

  • 25:51

    PAUL BAINES [continued]: using SPSS.And there are better--well, not better, but there are other booksthat talk about the kind of survey methods and techniques.But I think what students should do is look at some of the booksthat I've mentioned, but also look at journal articles, whichnot many students do when they're

  • 26:12

    PAUL BAINES [continued]: looking at the methodology.They might look at journal articleswhen looking at literature review,but they should also be looking at journal articlesfor their methodology.And there are plenty of articles talking about cluster analysisas a technique.There are plenty of--whether that's latent class or k-means.There are plenty of articles talkingabout survey design, and design of certain scaled questions,

  • 26:33

    PAUL BAINES [continued]: and so on.So that would be my advice to the average student.And for all the students out there doingresearch projects, good luck.

Video Info

Publisher: SAGE Publications Ltd.

Publication Year: 2020

Video Type:Video Case

Methods: Evaluation, Latent class analysis, Cluster analysis, Survey research, Marketing research

Keywords: counterterrorism and community policing; public safety and collaborative prevention; Social marketing

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



Professor Paul Baines and Superintendent Helen Isaacs discuss the design of messaging for Project Servator, a national police initiative to improve crime rates and deter counter-terrorism using unpredictable law enforcement deployment.

Looks like you do not have access to this content.

Evaluating Project Servator Using Quantitative Methods

Professor Paul Baines and Superintendent Helen Isaacs discuss the design of messaging for Project Servator, a national police initiative to improve crime rates and deter counter-terrorism using unpredictable law enforcement deployment.

Copy and paste the following HTML into your website