Home About us Solutions Case studies Testimonials Careers AOM Academy

Attitude surveys

An article by Neil Bentley

He uses statistics like a drunk uses a lamp post - more for support than illumination.

'Listening to staff' is usually regarded as a dispensable luxury during times of recession, and a symbol of enlightened management during a boom. As we move from recession to relative prosperity we can confidently expect to see more organisations expressing interest in introducing staff attitude surveys, and more research and consultancy companies blowing the dust from their offerings. It is therefore timely to recall some of the lessons learned in the last boom, and to speculate on how things might be different this time round.

The lessons

1 Surveys raise expectations

At the very least staff will expect a summary of the findings. (And if the survey reveals a belief among staff that senior management 'lacks vision', which surveys invariably do reveal, management may be forgiven for wishing to suppress the news!) More likely staff will expect not only feedback but action; but real and relevant action may not be that easy to implement. The paradoxical outcome may be that the act of 'listening' generates further dissension or scepticism.

2 Surveys in isolation seldom direct action

Survey data tend to highlight symptoms, rather than causes. For example, just knowing that 90% of employees feel that communications are poor doesn't tell you why they think they are poor or how they might be improved.

Furthermore, there is a risk of misplaced action. If employees believe that promotion is given to "blue eyed boys" whilst managers believe that they are scrupulously fair, they may not need to change the promotion system, only the perception of it.

3 Data problems

Designers of attitude surveys pride themselves on ensuring statistical significance. This can create a false sense of security. If the underlying data are unsound, no amount of statistical refinement will compensate.

The most fundamental flaw is the assumption that individual attitudes will sum to give predictions of collective behaviour. Psychologists, a long time ago, coined the term 'Group Think' to describe how groups will behave differently to how one might expect the individual members to behave.

A second problem is the way in which questionnaires can create opinions out of nothing. People will, on the whole answer a question if asked. Ask front line employees for their views on the company policy towards Europe, and you will get a view - even if they have never seen the policy. Up until that moment, they may have been entirely indifferent to the issue - now they have a statistically significant opinion ! At worst there is a "now you come to mention it" phenomenon in which the act of asking the question elevates its importance.

Even when the opinion is valid, it is difficult to be sure what you have got hold of. Did everyone interpret the question the same way? This kind of debate often surfaces when a result is less than favourable - and leads to an attack on the data rather than an attack on the problem.

Finally, everyone has experienced the frustration of ticking a box but wanting to qualify their answer. Many even waste their time and ink filling up the margin on computer read multi-choice score sheets!

New developments

1 Technology

The development of information networking and Groupware offers a streamlined mechanism for consultation. And as such the assessment of staff perception can be faster, more frequent and more akin to "business as usual", than the somewhat portentous methods of mass questionnaire.

2 'Attitude benchmarking'

Given the current popularity of benchmarking it is scarcely surprising that surveys should be offered by some companies in this guise. The use of standard instruments across the units of an organisation would appear to offer a form of benchmarking. Thus a Bank could compare perceptions of staff across its branch network, or a government department could compare the views of its staff in a number of regional offices. Such readings could form part of the evaluation of local management performance. Each organisation could be reasonably confident that enough consistency of issue and language exists internally to make the results reasonably comparable. Unfortunately no such confidence can attend the comparison of data gained from different organisations, and it is hard to imagine a single instrument that could probe the most relevant issues for any organisation.

Conclusions

It is our contention that attitude surveys should be used:

		Only in the context of a well considered communication strategy.
		Only when probable results can be forecast and management responses prepared in advance.
		Only when paper based (or electronic) methods are supplemented with prior qualitative research that reveals underlying themes and which is used to refine the questionnaire and focus it on relevant issues.

Most importantly surveys should only be used when there is a very good business reason - 'nice to know' does not qualify!

Attitude surveys

An article by Neil Bentley

The lessons

New developments

Conclusions

Other articles