2. Evaluating ABC
• Why evaluate
• What indicators can we measure
• Methods for measuring indicators
– Dog population survey
• Novel tools for surveys
(developed with funding from HSI)
– Combining clinic database with
survey data collected on a smart
phone
3. • Are you involved
in running an
ABC programme?
• Do you evaluate
your programme?
• Does this
evaluation go
beyond reporting
number of dogs
through ABC?
4. What is evaluation?
• Evaluation
“the assessing and judging the value of a piece
of work, an organisation or a service. Its main
purpose is to help an organisation reflect on
what it is trying to achieve, assessing how far it
is succeeding, and identify required
improvements”
Evaluation Trust
5. What is evaluation?
• Evaluation
“the assessing and judging the value of a piece
of work, an organisation or a service. Its main
purpose is to help an organisation reflect on
what it is trying to achieve, assessing how far it
is succeeding, and identify required
improvements”
Evaluation Trust
6. Why evaluate?
• Donors – reward their generosity and build
commitment
• Government and animal welfare
organisations – inspire them to replicate
• Your organisation – pride and motivation
• Your end beneficiary – the animals that
matter to us
Learn and improve
7. Indicators
• Evaluation involves measuring a range of
indicators to expose important changes
• When selecting indicators consider the aim of
ABC programme and opinions of stakeholders
– Government including health department
– Local community both dog owning/caring and not
– Donors
– Your own organisation (encompassed in the aim)
What do they want to see change?
8. Indicators
• Animal welfare:
– Body condition score
– Skin condition
– Lactating females
(more visible and
consistent to score
than litters of puppies)
– Disease incidence
(dog rabies cases)
– Euthanasia rates
– Dog population size
• Community
satisfaction
– Nuisance complaints
– Attitudes towards dogs
• Human health
– Dog bites
– Human rabies cases
9. Methods of measurement
• Official records
• Community questionnaires/focus groups
• Dog population surveys
10. Methods of measurement
• Official records
• Municipality records of nuisance complaints
• Health centre records of dog bites and human rabies cases
– Samples are sufficient, find a friendly source
– Note locations (e.g. postcodes) if this is possible, can
you find data from within the ABC area and outside
for comparison over time (incl pre-ABC) and between
treatment and control
– Note human population increase – e.g. don’t quote
just dog bite data, quote bites per 100,000 people
– Right to Information Act (RTIA)
11. Colombo dog rabies cases
Dog rabies cases in CMC against elimination, vaccination and
sterilisation from 1990 to end 2010
0
2000
4000
6000
8000
10000
12000
14000
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Year
Numberofdogseliminated,
vaccinatedorsterilised
0
10
20
30
40
50
60
70
Numberofdograbiescases
Elimination
Vaccination CMC
and BPT
Number of
sterilisations
Dog Rabies
Start of project
12. Methods of measurement
• Community questionnaires/focus groups
– WHO (2004) Guidelines for conducting
community surveys on injuries and violence
– Questionnaires can reveal changes in
quantitative data such as dogs bites (e.g. has
anyone in your household been bitten in the
last 12 months)
– Focus groups can reveal attitudes and
practices (e.g. do you think stray dogs are
dangerous or safe? What do you do if you are
bitten by a dog?)
14. Methods of measurement
• Dog population surveys
– Index of abundance (aka ‘indicator survey’)
• Quicker/cheaper
• But only indicates change in density and not total
population size
– Estimate of total population size
• Requires more time investment ahead of ABC, so
only invest in this if necessary
• Can be relatively simple once ABC had begun
16. Example – surveying for a new
ABC project
• Baseline
1. Dog population estimate based on exhaustive
search in small number of sample areas
extrapolated to whole area (assuming you need a
total dog population estimate)
• Refine after approximately 1 year of ABC
2. Mark-resight of dogs notched through ABC (novel
tool)
17. 1) Baseline population estimate
• Establish smallest authority area for which information
on covariates is available (e.g. human population, road
length)
• Select small sample of these areas (as many as can be
realistically managed, suggest minimum 5%)
• Complete count of all dogs visible on public areas at
peak roaming time (usually early morning)
• Extrapolate to entire city:
– Using sampling fraction (multiply number of dogs counted by
total number of areas/number of samples areas)
or
– Find reliable covariate (e.g. dog:human ratio) and extrapolate to
entire area
21. 1) Baseline population estimate
• Extrapolate to entire city:
– Using sampling fraction (multiply number of dogs
counted by total number of areas/number of samples
areas)
• E.g. counted 462 dogs in 7 wards out of a total of 47
Estimated total roaming dog population = 463 X (47/7) = 3,102
or
– Find reliable covariate (e.g. dog:human ratio) and
extrapolate to entire area
• E.g. counted 462 dogs in 7 wards in which live 16,000 people
= 1 dog:35 people. 100,000 people live in the city hence
estimated total roaming dog population = 100,000/35 = 2,857
22. 1) Baseline population estimate
• This is a ‘rough’ minimum roaming dog
population estimate:
– Small sample size
– Dogs roaming and visible in the early morning – not
total roaming dog population as some will be missed
– But can be a basis for lobbying and budgeting with
these clear limitations
– Be clear this will be refined and is not the final figure
23. 2) Mark-resight of notched dogs
• Dogs are notched during the ABC programme
providing a sample of ‘marked’ dogs in the city.
• Taking into account a survival estimate (some
ABC dogs will have died since the operation) the
number of notched dogs in the city can be
calculated
• Combining that with proportion of notched dogs
can be used to measure the population size
25. 2) Mark-resight of notched dogs
• ABC example
– ABC programme sterilises 3,500 dogs in first year, with an
annual survival estimate of 0.7 the number of dogs notched and
alive in the city is estimated to be 2,800
– The percentage of notched dogs on the street is estimated to be
46%
– Total dog population estimate is 2,800 / 0.46 = 6,087
Hiby et al (2011) A mark-resight survey method to estimate the
roaming dog population in three cities in Rajasthan, India
• Novel tool combining database for recording
clinic data with a smart phone app
26. 2) Mark-resight of notched dogs
• Mark-resight estimates the total number of
roaming dogs rather than the number roaming at
the time of the initial baseline survey
• Relies on a survival estimate
– Can use the estimate in Reece et al (2008) Fecundity
and longevity of roaming dogs in Jaipur, India; annual
survival of 0.7
– Can estimate a survival for your own area if using
individual tattoos
27. 3) Index of abundance routes
• Several standard routes of known length
• Representative areas
• Confounding factors
– Time, person, weather, protocol…
• Quick so can repeat
– Seasonal breeding, politically important time…
• Evaluate ‘like with like’
28. What to record
• Information for each dog:
– Age
• Adult
• Puppy – agree maximum age (e.g. 4 months)
– Sex
• Male
• Female
• Unknown
– Reproductive status
• Lactating
• Sterilised (notched)
– Welfare state
• Body condition score
• Skin condition
29.
30.
31.
32.
33. Welfare measures
• Body condition score (5 point BCS)
– Reduction % of dogs in score 1 BCS over time and
between notched and unsterilised dogs
• Skin condition (Presence/absence)
– Reduction in % of dogs with skin conditions visible
• TVTs?
• Behaviour?
1 – runs away immediately
2 – keeps a distance but remains in area
3 – approaches
4 – in your pocket!
34. Using an Android phone to record
data
• An offline database plus some
free smartphone apps can help
you to monitor your intervention
73. ADMIN:
download and
analyze city data
FUNDING AGENCY:
view by NGO and time period
NGO 1: upload
by month, view
NGO 2: upload
by month, view
NGO 3: upload
by month, view
74. Discussion point
• Which indicator is best for dog abundance?
• Total population for the city?
– But cities are expanding in size with increasing human
population in expanding areas = increasing carrying capacity =
increasing total dog population
– Average city dweller has no concept of total dog population size,
but they do perceive the density and behaviour of dogs
– Takes a lot of time/resources to establish a really reliable dog
population estimate ahead of ABC – although easy once ABC
has started using phone/database tool
• In comparison the number of dogs per km of street is:
– Quick to estimate from the standard routes used in index of
abundance surveys
– Relates more closely to the dog problem
76. Conclusion
• Animal welfare:
– Body condition score
– Skin condition
– Lactating females
(more visible and
consistent to score
than litters of puppies)
– Disease incidence
(dog rabies cases)
– Euthanasia rates
– Dog population size
• Community
satisfaction
– Nuisance complaints
– Attitudes towards dogs
• Human health
– Dog bites
– Human rabies cases
77. Conclusion
• Methods of measurement
– Official records
– Community questionnaires/focus groups
– Dog population surveys
• Database and phone app
78. Whose responsibility?
“ An important job had to be done and Everybody
was sure that Somebody would do it. Anybody
could have done it, but Nobody did it. Somebody
got angry about that because it was Everybody’s
job. Everybody thought that Anybody could do it,
but Nobody realized that Everybody wouldn’t do
it. It ended up that Everybody blamed Somebody
when Nobody did what Anybody could have
done”
Make someone responsible and provide time
and budget
Good morning and thank you for joining us today. We’ll be talking through some aspects of evaluating ABC today – its quite a broad topic, so apologies if we do not cover everything. You’ll notice that we have a double act today – I will be presenting for the first part of the workshop and then will be handing over to my father – Lex Hiby. Please note there are handouts available with a list of resources relevant to this topic here – if there are not enough copies, or indeed if you would like an electronic copy, please do email me afterwards and I’ll send this on. I also want to encourage you to ask questions throughout this wrkshop – I will also be stopping at times to ask for your input to certain key questions – so please spk up, I hope this can truly be a 2-way conversation
So today we plan to cover the following. By way of an intro looking at the definition of evaluation and why we would ever bother delaying our normal work to play with data What indicators are available that can tell us something about the aspects of the world we are hoping to change I’ll then go through some methods of collecting data that will help us measure these indicators, in particular different methods of dog population surveying Before handing over to Lex to talk through a novel tool for conducting ongoing dog surveys during the running of an ABC programme – combining a mobile phone app and database of clinic data from the ABC operations to track dog population size and coverage of ABC. I should be quick to point out that this tool has been developed with funding support from HSI.
I wanted to start by just checking who in the audience is involved with running an ABC programme? If you can put your hands up if this is true for you. Can you keep you hands up if you evaluate that programme in anyway
This is one of many definitions of the word evaluation. This time from the Evaluation Trust. But from my perspective the most important reason to define and measure your goal is so we can learn to do our job better
This is one of many definitions of the word evaluation. This time from the Evaluation Trust. But from my perspective the most important reason to define and measure your goal is so we can learn to do our job better
You may ask – why evaluate? It takes up time I could be using to ABC more dogs. Who cares? First audience is our donors – who may be government, AWBI or municipalities or state gov, or may be an animal welfare donor - raise your hand if you are responsible for raising money for your organisation. Effective evaluation can prove to your donor that their money is making a difference and secure further commitment to donate – important when we are running ABC projects that have to sustained for the long-term Government or animal welfare organisations that have yet to launch ABC – good evaluation can inspire them to start and replicate the benefits Yourself and your staff, your volunteers, your CEO. – pride in your work and motivates the team to keep dedicated to ABC And finally your end beneficiary – the animals themselves, who I am sure if they could would ask – would value seeing the benefit of ABC, they are after all the ones that have to be caught, held, anaesthetised and undergo surgery. What difference will this make to their life and the lives of other animals But I think that the most important reason is to be from my perspective the most important reason to define and measure your goal is so we can learn to do our job better
This is an example from Colombo – you can see on this blue line the number of dog rabies cases, measured and reported by the municipal vet office – we could plot this against the years of culling and then against the years of mass vaccination – showing neatly what worked best
- Explored attitude towards dogs ‘now’ and ‘in the past’ (>4 years ago prior to when intervention started) Participants were asked open questions about roaming dogs in their area, both currently and in the past – facilitators then followed up with additional questions and participatory exercises to draw from each group a list of problems presented by roaming dogs – both in 2011 and before the intervention started. On this spider web diagram you can see the 9 most commonly mentioned problems caused or experienced by roaming dogs around the outside – the blue line indicates problems in the past and the red those mentioned in 2011. We can see that all 9 focus groups mentioned rabies and breeding or the presence of puppies as a problem in the past – however only 1 group mentioned this as a problem in 2011. We can also see that the total number of problems mentioned in the past – as indicated by the area delineated by the blue line - also decreased significantly by 2011 – as indicated by the smaller area outlined in red
We conducted a population survey like this in Colombo Sri Lanka using ‘wards’ the Colombo city municipality split into 47 wards and we counted dogs in 7 of those. Importantly there were maps available for these wards that allowed to ensure we were counting all the streets in the city
Need maps available for these official boundaries
Use Google Earth where maps are poor
Jack Jack to explain why they chose the Pink City and the time it takes to complete the counts
3
1
4
2
Never tried it but what about behaviour? Does this describe your dogs?
xxx
So I just want to share something with you that I learnt the hard way. Have a read of this quote. Ever experienced this? I most certainly have and specifically on this issue of evaluation. I have been part of workshops were over a period of days and using consultation with government representatives, community groups and local vets we developed goals and indicators that lit each of our respective fires. But it became a paper exercise because we didn’t follow this trough to make sure we had the funds, time and clearly identified person responsible to make these indicators work for us. So we weren’t able to do good evaluations and hence learn and improve – we continued to stumble on in the dark.