Participatory research uses approaches and methods that help guarantee the validity of data by examining research questions through multiple methods, e.g. triangulation., Participatory methods can also create a sense of ownership of the data by the community and are directly linked to carrying out actions.

In response, to the absence of action in traditional research, specialists in participatory research have developed a model that includes:
Giving priority to behaviours and attitudes that encourage real collaboration between community members and “external experts”.  These behaviours include respecting local knowledge, the willingness of external experts to let local populations be in charge of the process and the flexibility and the informal character of the use of these methods.
These methods allow community members to actively participate in the data collection process. Some of these methods use local materials (sticks, beads, pots, beans, etc.) provided by the community, which can then be placed on the ground.  It should also be noted that even in non-literate communities large pieces of paper and marker pens can be used.

3.1 - Action research

Action research, as defined by Stephen Kemmis and R. McTaggart (1992), has the essential feature of “trying out ideas in practice as a means of improvement and increasing knowledge about the curriculum, teaching and learning”.  The concept of action research has its origins in the work of social psychologist Kurt Lewin.  He developed and applied it in post-WWII America in a variety of group contexts, including integrated social housing; equalization of opportunities for employment; the cause and cure of prejudice in children; the socialisation of street gangs; and the training of youth leaders.

Action research is a participatory activity which encourages collaboration between researchers, practitioners and beneficiaries (subjects) of research.  As a participatory activity, action research brings these stakeholders together to work towards the improvement of their own educational system – in the case of education research.  It provides a way of thinking systematically and implementing critically informed actions where improvements are thought to be possible.  Effects of the actions taken are monitored and evaluated so the system may continually be improved.

Action research is ongoing and does not end, as conventional research often does, after the presentation of the research report.  As a spiral process, action research develops through the self-reflective spiral of planning, action (implementing), observing, reflecting and then re-planning, further implementation, observation and reflecting … and so the process continues.Action research keeps the professional researcher and other stakeholders constantly interacting in a dynamic process.  Action research thus brings together theory and practice, in contrast to conventional educational research which separates theory from practice.  Theory without practice is empty; practice without theory is blind (Kwame Nkrumah, 1966).

The purpose of conducting free lists is to help isolate and define relevant themes.  Free listing questions simply ask respondents to list as many items they can think of in a particular domain.  For example, in interviews with mothers about their children’s most recent illnesses, several mothers have mentioned fever.  You would like to know if mothers recognise and have names for different kinds of fevers.  In order to determine whether “kinds of fevers” is a meaningful theme for these mothers, you might ask: “Several mothers have told me that their children had fever.  Are there different types of fevers?”

Practice
Can you, reader, imagine a specific education-related context in which this technique could be used?

Strengths of free listing
Free listing is a useful first step in all research involving the definition of new domains.  It is the best way to ensure that the concepts and domain are culturally relevant.  Participants have little difficulty with this technique.

Weaknesses of free listing
Researchers need to be familiar with the culture and language of the respondents in order to identify appropriate domain names to start the listing process.

Pile sorting is a technique aimed at studying the relations among items within a domain (which have been identified through free listing or some other method). It is useful for discovering informants’ perceptions of the similarities and differences among items and to look at intercultural variation in how informants define domains.

Practice
Use local materials to quantify the importance given to each aspect of the subject being studied. It is followed by the composition and recomposition of piles until group consensus is finally reached.

Strengths of pile sorts
They can be done using large numbers of items, and are easy to administer.

Weaknesses of pile sorts
Unless items are pictures or actual stimuli, they cannot be used with non-literate informants. Simple hand analysis (i.e. non-computerised) of pile sort data is possible.  Computerised analysis is faster, but requires knowledge of an appropriate software programme.

This method consists in presenting a situation or a question and then asking the participants to write their responses on a card.  All the participants’ cards are fixed to a board/wall and can be seen by everyone.  The ensuing discussions can validate, modify, seek consensus etc. for the different responses on each card.  See Annex D for more information on this method.

Rating scales are a very popular technique for questionnaire data collection in the social sciences.  Scales can be created for any number of concepts or attributes, and items can be rated on a single conceptual scale or each may be rated on a series of scales representing a variety of concepts or attributes.  Scales can be presented numerically or graphically:

Example
“Circle the number that corresponds to the level of severity you would associate with HIV/AIDS”

0   1   2   3   4   5   6   7   (from least to most severe)

or

“Mark an ‘x’ on the line below indicating where on the line below you would rate HIV/AIDS in terms of severity”

Mild                                   Severe

Another approach, which can be used with non-literate respondents, is to use cards or some other visual stimuli (actual foods or medicines, for example), and ask the informants to place the objects in piles according to some pre-defined rating criteria.  For example, if you are interested in rating a number of illnesses according to their perceived severity, you might decide to use a 3-level rating system (very serious, moderately serious, not very serious at all), and say to informants: “I’m going to say the name of an illness and I would like you to tell me whether the illness is serious, moderately serious, or not very serious at all.”

Practice
Can you, reader, envisage a use for this technique in a specific context of education research?

Advantages of rating scales
Rating scales are easy to administer.  As the informant rates each illness, the researcher places the corresponding card or object in the appropriate pile.  Such visual stimuli make the task more fun for the informant and allows the informant to consider the relationship among items and to change their rating if necessary.

Weaknesses of rating scales
Rating scales can be extremely sensitive to response bias (the propensity of individuals to always use one end of the scale or a narrow range in the middle of the scale).  This can make it difficult to compare data between respondents.

3.6 - Rank order methods (complete and partial)

Rank order methods require informants to rank items (i.e. from most to least) in terms of a specific characteristic, for example, illness in terms of severity. Below are the two most common methods for rank ordering items:

Complete rank ordering methods usually require that informants be literate, and involve presenting respondents with a list of items which they are asked to order from “most” to “least” on some attribute by putting numbers next to each item.  Another method is to present each informant with visual stimuli or a stack of cards on which item names are written, and ask them to order the items or cards from “most” to “least” (“best” to “worst”, or whatever the attribute of interest is).
Partial rank ordering pairs each item with each of the other items (for a total of n [n-1] / 2 pairs).  These pairs of items are presented to respondents, who are asked to indicate which “more” is or “less” (“best” or “worst”, “most preferred” or “least preferred”, etc.).  This technique is also called a paired comparison.  A rank ordering is obtained by summing the number of times each item was chosen.

Strengths of rank ordering methods
The complete rank ordering technique produces a great deal of information in a short amount of time.  The method is ideal for studying differences among individuals. Paired comparisons are probably the most reliable method to use with non-literates when the number of items is small.

Weaknesses of rank order methods
The complete rank ordering technique can be tedious for non-literate respondents.  For paired comparisons, pre-testing is crucial for identifying the maximum number of pairs that informants will tolerate.  Some researchers have found that even as few as 15 pairs (6 items) can become tedious.

3.7 - Mapping and modelling

In this technique, a group of individuals from the study population are asked to draw or build maps and models of their community, using locally available materials (including: sticks, stones, grass, wood, cigarette packs, leaves, sand, etc.).  Maps or models can be constructed by different interest groups men, women, young, old, poor, wealthy etc.) working together in order to represent their particular viewpoints and interests.  Both the process of building the map and the final product facilitate the identification and exploration of topics of interest. There are many different types of maps and models that can be created:

Resource maps that represent health service coverage areas, villages, forests, fields, farms, garden, etc.;
Maps that illustrate different residential areas of a village;
Social maps which illustrate wealth rankings and household assets;
Health maps in which the health and welfare status of family members is marked on or by each house using various types of markers.  For example, it might indicate where malnourished children or TB patients are located;
Topical maps, which might show the soil quality of various fields, the location of water sources or irrigated fields, etc.

3.8 - Timelines

In this technique, group members are asked to recall local events and develop a community history which includes approximate dates. Timelines are especially useful for facilitating discussion about changes that have occurred over time (changes in health, crop output, environmental conditions, etc.).

3.9 - Observations

Interviews provide information about people’s beliefs, attitudes, values and reported behaviour.  Observations provide information about actual behaviour.  Direct observation of behaviour is useful because some behaviour involves habitual routines of which people may hardly be aware. In such cases, observation can reveal more information than other data collection methods.  Direct observation also allows the researcher to put behaviour into context, and thereby better understand it.

In research, observations can only be used as a reliable and unbiased source of information to the degree to which they are focussed and systematic.  The researcher should be explicit about what he or she wants to know, learn and understand, and how observations are to be collected and recorded.

Observations can be made either of people’s actual behaviours, or observations can be made of signs or indicators of behaviour.  Sometimes behaviour would be impossible to observe directly, and researchers must settle for observing indicators of that behaviour.  Observations can either be unobtrusive and non-reactive, or obvious and reactive. Unobtrusive observation, whereby people are observed without their knowing it, can involve serious ethical dilemmas.

3.10 - Formative Research

Formative research can help improve policy implementation.  Formative research is a highly participatory process-oriented assessment tool and can be adapted for macro-analysis of complex processes, to support systemic change.  It draws on case study approaches and uses qualitative research methods.  It promotes dialogue and develops human and institutional capacity.  Formative research can help improve educational practices.  It entails seeking answers to such questions as:

What is working?
What needs to be improved?
How can it be improved?

Figure 2: The Interactive process between program implementation and formative research

Figure 2: from K.Touré for ADEA Working Group on Education Sector Analysis, 2003

Tornes and Feldman argue that a main challenge for any reform is to make the various levels involved in the process communicate and relate to each other in ways that help meet the goals of the reform in question.  They propose that with a longitudinal, systems, approach to gathering data and communicating with the various groups of stakeholders during the reform process, the formative model is well suited to cope with the challenges of sectoral reform.  the focus in on comprehensive processes.  Unlike action research there is no assumption in the formative research model that practice must be redefined in line with research finding with the researchers playing an active part in initiating new practice.  Formative research produces findings to be used along with other sources of information, attempting to ensure that research-based knowledge is not seen as superior to or more true than the knowledge of policy makers, practitioners and other stakeholders.

Through formative evaluation, actors gain information and skills necessary to adapt programs plans to account for new factors and realities encountered during initial stages of implementation.

Figure 3: Relations among actors of the formative research process

Figure 3: adapted from Feldberg, Karen

3.11 - Terminology used in different participatory research approaches

Certain terms and acronyms are used to talk about participatory methods and processes. These include:

3.11.1  Rapid Rural Appraisal (RRP)
In RRP a multi-disciplinary team of researchers spend between 4 and 8 days in a community and involve all its members in all aspects of the data collection.  Generally, an RRP study produces a report which summarises the results.  Although this type of research is more participatory than traditional research methods, many practitioners have reservations as the participation is and isolated event. 

3.11.2  Participatory Action Research (PAR)
PAR methods such as diagrams and time lines are the same as in RRP.  The objective of PAR is to make the community responsible and not just in the data collection stage.  The vision is long-term in PAR and allows the community to analyses its own problems, define priorities and design a plan to solve these problems.  In PAR the role of “experts” from outside is to coordinate the process and offer technical assistance where necessary and in response to needs expressed by the community.  In PAR it is up to the community to direct the process; if RRP is a discrete, short-term process, PAR on the other hand has much wider scope and can last for several years.

3.11.3  Participatory Learning and Action
This relatively new designation is becoming increasingly popular.  Here, the accent is on actions to be taken based on the results, rather than considering data collection as an end in itself.  This ever-growing set of participatory methodologies and approaches integrates data collection techniques such as RRP and PAR but also other activities which strengthen and perpetuate local populations’ participation and control