The number of responses required for your assessment to be representative (of the overall population you are surveying) is dependent on your sampling method that your organisation uses and total population size.

The following guidance is offered based on our experience:

• Gather as much data as you have time for or until the “data runs clear” i.e. no more new information is emerging, whilst avoiding survey fatigue
• Data can be sampled from large populations to statistical standards (see table below on sample sizing) if this is an important frame of reference in your organisation
• Try to get at least three “cuts” of data wherever possible from senior management, middle management/Change Agents/Project Teams and frontline workforce (Note: Our experience is that the view of the frontline workforce is twice as likely to be accurate as the view or senior managers in terms of identifying implementation risks. Weightings of 1 = senior managers, 1.5 = middle managers and 2 = frontline may be used in any averaging any data.)
• Most Sponsors accept data from a sample of around 50 people as authentic – irrespective of statistics.
• If survey fatigue is managed appropriately, sample more than you need to allow for a lower response rate than anticipated.
• Consider whether it is actually of more value from an involvement and political point of view to sample entire populations rather than to request input from a whole population.
• Use the assessment introduction to explain exactly why that person is being asked to take part, involvement and the reasons why you want them involved are important to communicate.
• Sharing the overall results afterwards may assist in engagement, particularly when following up with a workshop to gain qualitative inputs.

Review this guidance before deciding on the number of respondents you need, based on the recommended sample sizes (for your population) to obtain a representative result for your total population.

Consider survey fatigue/lethargy

Be aware to try and avoid survey fatigue/lethargy by communicating with your colleagues who are licenced to use Changefirst Diagnostics in advance of sending multiple requests in a short period of time to the same populations.

Scheduling with colleagues before surveying is important (you may want to maintain a list of assessments with a schedule and review that prior to any sending of more requests as this could assist in monitoring aged data and avoid participants being overloaded across multiple projects).

Representative Sample Sizing

The sample sizing is based on:

• sampling error = +/- 2
• confidence level = 95%
• estimated Standard Deviation = 10