How do you define a Missing Person?

[Daryl]

In helping to develop the SAR Management application that I have introduced I have asked for an option to include the ability to cliassify Missing People so that we can start to create Missing Person Profiles.

Below is a list to start you off, but what classifications do you use? Maybe you would like to add your thoughts to the Blog also?

Despondent
Dementia Alzheimer Type
Runaway
Lost
Mental health
Natural disaster
Man-made disaster
Injured
Health problem
Unspecified
Abduction
Criminal investigation

[Robert]

The first major missing persons profile data was collected by William Syrotuck, his work has laid the foundation for this type of search analysis and has certainly helped the BMRC with there data. I think that one problem with too many categories of missing persons are the human input element, were not so great at making choices when filling out forms. When we started search profiling we looked at the available statistics and written procedures and found most are limited by man made funnels, roads, footpaths, pylons etc. and where we are in the south of England you are mostly lost due to being unable to move, rather than the old fashioned meaning of lost.
Syrotucks categories were.
Small Children (1-6 years)
Children (6 - 12 years)
Hunters
Hikers
Miscellaneous outdoor persons
Elderly Persons (over 65)

Special categories
Mentally Challenged
Despondents

The main focus on this type of profiling was distance from PLS and Distance walked. Ground is catergorised uphill down hill and flat.
This may not encompass everything but it is a great work from a great man well worth a read, and for this type database simple to use and easy to interpret the results.

Robert Greenwood

[Pete]

Although it is possible to have too many categories, research would require as much information as possible.

The origional categories did not give sufficient information for collation of data regarding Misper Stats (Just take a look at the Grampian Data, it has many categories).

With the world of SAR changing so quickly wording such as 'misc' and 'other' does not do justice to who we are asked to try and locate.

I would have to agree with the origional categories placed by DT.

Pete.

[Daryl]

I have to say that in my experience the 'Old' style of MISPER stats that has categories such as 'Hunter' and 'Hiker' (US Wilderness) which presumeably came from the states was not particularly useful for Lowland (Urban & Rural) searches in the UK.

Grampian Search Data seems to be the current best standard used by the Police in the UK (in my area anyway).

[Robert Bradley]

There are inevitably problems with any classification system; however, as Daryl stated the Grampian Missing Person data (despite its own drawbacks) is the one published by NPIA and used as the standard by many police forces in the UK.

Obviously whilst Syrotuck's work did lay the foundation for this type of statistical analysis, the many flaws in his work has meant that it has been superceded in many ways. The data collected by Bob Koester et al. for the International Search & Rescue Incident Database (ISRID) being the international standard these days and useful for comparison with our UK data.

Issues with Daryl's first list;

Dementia Alzheimers Type - as we all know there are several other causes of dementia and whilst many refer to all dementia mispers as Alzheimers, better to put it as just dementia.

Mental Health - there are too many types of mental health issues, with so many symptons that this category would not provide any useful data - because of the large differences between the different groups.

Abduction - the data for this category is already collected in great detail by NPIA at Bramshill in the CATCHEM (hope I've got the acronym correct) database.

It would also be beneficial for all the data to be able to link into the ISRID data in some format - otherwise everyone is replicating each others' work, pooled together we would have a much greater resource.

One final thing - the NPIA/ACPO guidance on missing persons gives a reasonable guide to the differences between lost and missing persons, although dementia mispers breaks their definition laid out in table 2, page 23.

(Looks a good piece of software though, and great it compares actual incidents to exercises and training - something I have been preaching for years!)

Robert

PS T'other Robert - I'd be very interested to learn more about how you use search profiling.

[Daryl]

I agree with Robs points.

Rob - what are the categories used in the ISRID Databse?

[Robert]

Nice debate, I may have made you think the wrong thing. My main concern with collecting data in this way would be the problems of our type of area, the UK doesn't have the large areas where the Syrotuck type system would work. However the merits I was trying to highlight are the simplicity in data collection, and general use. The Descicions for heroes site is good because of its simplicity, if you overload with options you can run the risk of it not being used. For the purpose of the software the intergration of stats for profiling needs to be well thought out to meet ease of use and purpose of data.
We have used profiling within the Coastguard (of which the origin of I'm unsure) and have collected information which when tested against another team on the same incident was un comparable due to complicated forms. The challenge is to make a system sufficient, easy to use (UI) and purposeful data, which needs to account for rural, urban and complicated searching.
At the end of the day we still have to look and local knowledge has ended many a search in a timely manner.

Robert

[robinb]

I've been in touch with these guys http://www.searchresearch.org.uk/www/ukmpbs/

They seem to have lots of info & definitions.

-Robin.

[Robert Bradley]

The categories are being added to all the time; at last count I think there were over forty.

Many, of course, are out of the remit of the type of search we tend to be used for but with data being collected from all around the world that is understandable.

Anyone really interested should contact Bob Koester - who should be a lot freer now he has finished his latest book on the subject!

With regard to simplicity etc.
The more simplified a model becomes, generally the less relevant it is and less useful for predicting behaviour.
This doesn't mean, of course, that the interface needs to be complicated - but that the underlying data must, by its very nature be rather more complex to be useful.
If I was designing such a system, as it started collecting data, these "most popular" types of incident would populate the first drop-down box or whatever, with a link to a section with all the rest of the categories. This way you are not degrading the usefulness of the data, in order to ensure an ease of use.

Robert - if you do have sets of data on this subject I'd love to see them, even if they seem that they are not comparable; You can still learn quite a lot from them.
You are right that the UK geography makes it different to say the US, where Syrotuck did his studies, which makes it doubly important that the UK and teams collect statistics of their own. However, the UK stats we have at the moment show that search could be improved with greater knowledge of these. For example, the distances generally travelled by despondents is very low once they are on foot, even in quite rural areas. Search managers, having searched the near-by areas, tend to expand their search areas too quickly, rather than re-search these high probability areas, which would be suggested if they had seen the probability density maps these statistics produce. (Of course, trying to remember Syrotuck off the top of my head, this is one of the issues with his work in that he suggested working at different points - mean or median? something like that, rather than calculating the probability density which would be far more useful - although with a maritime search background this will be all to familiar to you.)

Robin - if I remember rightly the UK stats collected by the centre for search research are forwarded for use in the ISRID - however, the definitions given online whilst useful are not the ones used by ISRID. Bob Koester would be able to give you a lot more information and is very approachable.
As an aside, Charles Twardy, who commented on your blog, produces some great wire diagrams of probability density - something that would be very useful for managing searches.

[Daryl]

The definitions used by MR (on the centre for search researchwebsite) can be found here http://www.searchresearch.org.uk/downloads/ukmpbs/Definitions.pdf