Power and Evaluation

It could be argued that research on power and evaluation is still relatively underdeveloped – for instance, Pawson and Tilley (1997, p.20) highlight a ‘failure to appreciate the asymmetries of power’ within the evaluation context. Yet, many of the issues in evaluation (utilisation of, access) appear to arise from issues relating to power. For this reason, I briefly want to consider Yukl’s, French and Raven’s, and Greene and Elfrers’ theories of power and apply this to the evaluation context. 


Burton and Thakur (1995, p.354) define power as ‘the ability or the potential ability of a person or a group to influence another person or group’. In this way there are two observable parties the the influenced, and the influencers. French and Raven concur, stating that ‘the phenomena of power and influence involve a dyadic relation between two agents’ (2001, p.61). Abma and Widdershoven’s definition also suggests multiple parties in a power dynamic:

Every social relationship involves power. Power refers to the possibility of letting someone do something he or she would not do otherwise. Power is thus relational, and parties in a power relationship are tied to each other by mutual dependency (Abma and Widdershoven, 2008, p.212).

In the context of the evaluation relationship, the primary parties are the evaluator and evaluated (initiative, project, authority), however, other parties also exist (for example the community or beneficiaries).

Various origins and forms of power are presented within available literature. French and Raven (1959 in Moorhead and Griffin, 1989, p.358) identify five power bases from which power can originate: ‘legitimate’; ‘coercive’; ‘reward’; ‘expert’ and ‘referent’. Similarly, Greene and Elfrers’ (1999, p.178) boast analogous categories but extend to include ‘connection’ and ‘information’ power. Yukl (1998) offers three origins of power; broadly correlating with those proposed by French and Raven, and Green and Elfrers, these are ‘position power’, ‘political power’ and ‘personal power’. The power bases typified by French and Raven, Greene and Elfrer, and Yukl are applicable to the relationships and interactions occurring through evaluation. The table below provides a summary of these power typologies and their relevance to evaluation.

Characteristic Power type Evaluation influence
Authority, delegated, formal, within organisation Legitimate Power (Green and Elfrers, French and Raven)Position Power (Yukl)Political Power (Yukl) An evaluator can be given authority to exert power over others perhaps by directing resource. Politically evaluator knowledge and findings means that their position is well regarded.
Control sanctions Coercive Power (Green and Elfrers, French and Raven)Position Power (Yukl) An evaluation function can instigate the most punishing sanction if used for governance purposes. Evaluation findings can ultimately prevent the continuation of a programme.
Control reward Reward Power (Green and Elfrers, French and Raven)Position Power (Yukl) The reporting of positive results may lead to rewards for the programme concerned.
Specialist skills/knowledge Expert Power (Green and Elfrers, French and Raven)Information Power (Green and Elfrers)Personal Power (Yukl) The skills possessed and learned throughout the evaluation experience provide an evaluator with information or expert types of power.
Social need to be liked to influence others, personality  Referent Power (Green and Elfrers, French and Raven) This may present difficulties for an evaluator with issues of independency clashing with personal needs.
Networks – both internal and external Connection power (Green and Elfrers)Political Power (Yukl) Evaluators should become well connected within the programme setting maintaining good relationships with members of these networks. There are many power consequences if these are maintained, or oppositely neglected.

Position Power and Evaluation

Yukl’s notion of ‘position’ power refers to the authority that is delegated to stakeholders within an organisation, mirroring French and Raven’s (1959), and Greene and Elfrers (1999) concepts of ‘legitimate’, ‘reward’ and ‘coercive’ power. Position power gives the possessor (in this case evaluator) formal authority, status and control, over operations and other stakeholders, to a specified extent.

In the case of an evaluator, there could be said to be position power, with their work being capable of influencing the continuation of a programme. This position power gives the evaluator formal authority to act on behalf of other stakeholders according to Clarke (1999, p.26), who asserts ‘evaluators have a moral responsibility to act as advocates for powerless stakeholder groups’.

There is also a risk that position power is wrongly exerted, creating friction and conflict amongst stakeholders (Weiss, 1972; Gordon, 1991; Davies, 1999; Sell et al, 2004; Abma and Widdershoven, 2008). Formalising the power structure, including where an evaluator sits among this may be worthwhile; as such position power may also refer to a role within a political structure (Ridgeway, 1991; Sell et al, 2004):

Within an organisation the allocation of status assists in operational and power structures. The formalisation of power through such a structure/status has been found to assist the authority of power players to be accepted (Sell et al, 2004, p.47).

Evaluators should not necessarily be seen as the subordinates of managers, commissioners or other stakeholders, although and here lies a potential power struggle which needs to be managed for a healthy stakeholder relationship. Poor clarity of the role and purpose of evaluator only adds to this issue: ‘it is useful to realize that clients may have specific ideas about what evaluation means and how an evaluation is supposed to be conducted’ (Stecher and Davis, 1988, p.22):

The key question for any local evaluator remain the same, how does one engage with such networks in which issues of power and difference often remain unacknowledged (Diamond, 2005, p.179).

Understandably, there is opportunity for an evaluator and another delegated power-player to disagree. An imbalance of ‘knowledge’ and power between funders, initiatives and participants may cause complications (McKie, 2003, p.321; Huberman, 1990; Lincoln, 1994). Clarke (1999, p.15) remedies this by suggesting that ‘it is essential at the outset that an evaluator obtains a clear understanding as to what the client requires from the evaluation and develops the evaluation research design accordingly’. The type of evaluation being conducted influences greatly the importance of power in the evaluator/stakeholder relationship.

Reward Power and Evaluation

Coercive and Reward power allows the holder to impose sanctions or entice subordinates with rewards. Coercion in a power context has been discussed by a host of exchange and political theorists (Lasswell and Kaplan, 1950; French and Raven, 1959; Molm, 1997; Sell et al, 2004). Generally (in the UK), bureaucratic mechanisms such as processes, procedures or legislation limit the allocation of rewards and sanctions (Burton and Thakur, 1995). The perceived legitimacy of coercive power may also affect its degree of acceptance, if it is felt that power has been exerted unfairly then conflict may occur.

A sense of forcing or getting another to do a certain task, in a certain way or for certain purposes is fraught within evaluation politics (Frohock, 1974; Mohan and Sullivan, 2006). It may be an evaluator attempting to have their findings used within the policy decision-making process, or programme staff trying to get evaluators to report favourably on the programme in question.

Expert Power 

Expert power considers the specialist knowledge or skills of the holder giving them power and control over others. Evaluators may seen to have expert power, as experts in evaluation research. Burton and Thakur (1995, p.357) also explain that a compromise between the expert and managers can sometimes be overcome by ‘power sharing’, ‘when a manager is aware that subordinates possess significant experience, it is common for the manager to allow these subordinates the exercise of considerable power’. 

But you forgot Referent and Connection Power! 

Referent power and connection power relate to relationships, networks and social needs. These are interesting areas and bring into play aspects of evaluator independence. Whilst I have researched a great deal on these and written about them I would like to conduct more research to look at them in greater depth – watch this space.