Method for prioritising and ranking

How do we Evaluate Programmes?

Data Analysis methods

 

 

used in PPA

 

Ranking or scoring means putting things in order and it reveals differences

within a population. It helps to identify the main problems or preferences of

36

people, and the criteria they use when deciding in what order to place things.

Further, it enables the priorities of different people to be compared.

Ranking exercises can be used in interviews or on their own and they can lead to

more direct and revealing questions (for example, Why is corruption a more

serious problem than illiteracy?). Tools such as ranking are very useful especially

if they are used to complement semi-structured interviewing. They may be used

either as part of an interview or separately. Pair-wise ranking, for example, helps

identify the main problems or preferences of individual community members, and

their ranking criteria, and enables the priorities of different individuals to be easily

compared.

Ranking can also be useful for collecting sensitive information, especially on

income and wealth. Informants tend to be more willing to provide relative values

regarding their wealth than absolute figures.

There are several types of ranking. The most common ones are:

(1) Preference ranking – Where people vote to select priorities.

(2) Pair-wise ranking – Where a matrix is drawn to compare which is the

preferred of two options.

(3) Direct matrix ranking or scoring – A way of identifying criteria for

choosing certain objects. It can be used as a means of understanding the

reasons for local preferences for such things as tree species or crop varieties.

The criteria are likely to change from group to group. Women and men may

also use different criteria.

(4) Wealth ranking – Where the communities classify themselves into various

categories of well-being, giving reasons for classification as well as how

people move into or out of particular categories.

5.9.1 Preference Ranking

The main objectives of preference ranking are:

To determine the main preferences of individuals and groups within a set of

items.

To compare the priorities of different groups.

Steps of Preference Ranking

(1) Choose a set of problems or preferences to be prioritised. This could be, for

example farming problems or preferences for domestic animals.

(2) Ask each respondent to give you her/his favoured items in this set, in order of

priority. Although there is no limit to the number of items each respondent can

handle experience shows that about 6-7 items would easier to manage. Beyond

this the exercise becomes cumbersome for most respondents.

(3) Repeat the exercise for several respondents

(4) Tabulate the responses.

37

 

5.9.2 Pair-wise Ranking

This is one form of preference ranking. Pair-wise ranking allows the PLA

team to determine the main problems or preferences of individual community

members, identify their ranking criteria, and compare the priorities of different

individuals.

Steps in pair-wise ranking:

(1) Choose a set of problems, or preferences, to be prioritised.

(2) Choose with the help of participants (or from previous discussion or from

a key informant), a reasonable number of the most important items in this

set5.

(3) Note down each of the selected items on a separate card.

(4) Place two of the cards in front of the respondent/informant and ask

him/her to choose the bigger problem (or more favoured preference)

giving reasons for the choice. Mark down the response in the appropriate

box in the priority ranking matrix.

(5) Ask whether the other of the two problems/preferences is in any respect

more important/more popular than the first. Note down the criteria in the

ranking criteria matrix.

(6) Present a different pair and repeat the comparison.

5 Although there is no specific limit to the number of items one may work with, a number of PLA

practitioners find large numbers of items rather cumbersome to use.

38

(7) Repeat steps 4-6 until all possible combinations have been considered

(and all boxes of the matrix have been filled).

(8) List the problems/preferences in the order in which the respondent has

ranked them by sorting the cards in order of priority.

(9) Check with the respondent whether any important problems/preferences

have been omitted from the list. If there are any, place them in the

appropriate position in the ranking table.

(10)As a useful cross-check to the responses, complete the ranking session by

asking the respondent about the biggest problem (or most favoured

preference) in the list (e.g., “If you could grow only one crop variety,

which one would you choose?”

(11)Repeat the pair-wise ranking exercise for a number of individuals, and

tabulate their responses.

Table 2: Example: Pair-wise ranking of selected crops

Millet G/Nuts Simsim Peas Beans Score Rank

Beans Beans Beans Dura Beans XXX 3 2

Peas Peas Peas Dura XXX XXX 2 3

Dura Dura Dura XXX XXX XXX 4 1

G/Nuts G/Nuts XXX XXX XXX XXX 1 4

Millet XXX XXX XXX XXX XXX 0 5

5.9.3 Direct Matrix Ranking

Direct Matrix Ranking allows the PLA team to identify lists of criteria for a

certain object. It allows the team to understand the reasons for local

preferences for such things as tree species, or crop varieties. The criteria are

likely to change from group to group, and women may have different criteria

for certain trees from that of men.

Process guidelines:

(1) Ask community members to select a set of items/objects that are important to

them and which they want to rank (Examples: animals, tree species, crops,

fruits, etc.);

(2) List the most important items (not too many);

(3) Find out criteria by asking:

“What is good about each item? What else? (Continue until

there are no more replies)”.

“What is bad about each item? What else? (Continue until there

are no more replies).

(4) List all the criteria. You may wish to turn “negative” criteria into “positive”

ones by using the “opposite” expression, for example “vulnerable to pests”

becomes “resistance to pests”.

(5) Draw up a matrix. For each criterion ask which object is best:

- “Which is best, then next best?”

- “Which is worst, then next worst?”

- Of the two remaining ask, “Which is better?”

39

 

5.9.4 Wealth (or well-being) ranking

This tool can be used to investigate perceptions of wealth differences and

inequalities in a community, to discover local indicators and criteria of wealth

and well-being, and to establish the relative wealth of households in the

community. It can be useful if, for example, one is interested in identifying

who the poor people are and how they may currently be targeted.

Wealth ranking is done by making a list of all households and asking different

people to sort them into categories according to their own criteria of ‘wealth’.

The term ‘well-being’ is often used, since perceptions of wealth usually

include non-economic criteria. Often only three categories are needed: the

poorest, middle and richest (or much better-off). However there is no “right”

number of categories, and communities should always be left to determine

their own categories using their own criteria.

Steps in Wealth Ranking

(1) Identify a group of people who are knowledgeable about the wellbeing

of the community and who represent different categories and

shades of the population (men, women, youth, educated, non-educated,

etc). These will serve as your community representatives.

(2) Explain the purpose of the exercise (for example to understand more

about the different categories of well-being or wealth in a community

with a view to recommending action or targeting).

(3) Ask the community representatives to list on cards all households in

their village. Only one name should be written per card. Remind the

40

team that no household should be missed in the listing (including those

for people who may be single, widows, or those without a home).

(4) Begin with a general discussion on poverty: how it is perceived in the

village, who is affected by it, how it has been changing over time, and

how it affects different people differently.

(5) Ask the team you selected to identify the main categories of well-being

in the village, and to spell out the specific characteristics for each of

the categories. A start could be three or four categories.

(6) Explain how the exercise will be carried out: one person reading cards

of household members and others assigning the card to a given

category.

(7) Follow up the assigning of cards with the question: “Why is s/he in this

category and not in another?” As the reasons unfold, add to your list of

characteristics. As the exercise progresses it may dawn on the

community that there are actually more categories than they had

previously defined. Add more categories as the need may arise.

(8) After all the cards have been assigned, ask the community members

whether they wish to make any changes to their allocations. If they do,

find out REASONS WHY. Following this ask them to count the cards

in each category.

(9) Find out whether in the last 5-10 years (period may vary) some people

have moved from one category to another – “upwards” or

“downwards”. Establish the nature of movement (sudden? Slow?

Other?) and the reasons for the movement.

(10) Establish with the community whether more people have been

moving “upwards” or “downwards” and the reasons why.

(11) Ask one of the community members to transfer the results of

the exercise to a large paper. A pie-chart representing different

categories of well-being may be one way of representing the outcome

of the exercise.

Things to note:

(1) Well-being analysis can be a sensitive exercise. Teams should way the pros

and cons before embarking on its use. Good rapport with the community is

always important for a successful well-being exercise.

(2) A lot of “side discussions” take place during the process of well-being

ranking. These are very important as they tend to bring out peculiar or

“uncharacteristic” issues and comments. Listen very carefully, especially, to

the silent voices.

(3) In undertaking well-being analysis you may find it necessary to separate men

from women, as there are sometimes likely to be different perceptions of

indicators of poverty.