Advanced Statistics Written by Joanne Birchall from Rainbow Research
This article covers the following:
· Factor Analysis
· Cluster Analysis
· Regression Analysis
Descriptive data is important and reporting the percentage “score” on each key question during debrief presentations or in reports is extremely useful, however it is the most basic level of analysis. Conducting further analysis on the findings could help not only assess the data at a more detailed level, but also interpret the strategic messages contained within it. Further analysis allows you to dig deep and hone in on specific areas within the data resulting in investment focus in areas of high return. Such analysis techniques include:
You may have questioned your respondents on many varied areas relating to your business and knowing where to start in terms of analysing and interpreting your data may be minefield. For example you may have data for a battery of 20 plus attitude statements, and not know where to start in terms of interpreting the findings. Factor analysis allows you to boil down the data into recognisable and manageable ‘themes’ e.g. loyalty, image, service, price, value for money etc. which helps to identify a “birds-eye” view of the data and significant patterns within it. You can then use these ‘themes’ to continue with your analysis in a more structured and easily interpretable way.
Cluster analysis helps identify ‘segments’ or ‘clusters’ from your sample with particular attitudes, behaviours or demographic profiles e.g. younger enthusiasts who are innovators and early adopters of new concepts and are driven by branding/image; or older price driven customers whose propensity to spend is driven by perceptions of value for money; or various integrated levels of employee satisfaction and commitment i.e. apostles, terrorists, mercenaries and hostages.
Again, a battery of attitudinal and / or behavioural statements relating to the topic of interest, are presented to respondents on and agreement/disagreement scale. Their responses are analysed, in effect by assessing correlations between their responses to assess which of several categories each respondents falls into in terms of their behaviours / attitudes towards a product, service, their morals, spending capacity etc. At this stage we know how groups of respondents’ attitudes / behaviours differ. Each respondent is then allocated a cluster number (there are usually around 3 to 5 clusters of people) and the data is then run against demographic profile data (gender, age, socio-economic grades etc.) to identify who these people are with differing attitudes / behaviours. This technique is ideal for identifying target audiences for marketing purposes.
Strategic Priority Analysis
The main purpose of strategic priority analysis is to help identify priorities for action in order for the business to provide the ideal products / services to customers, grow, become more profitable, diversify etc. This technique allows businesses to identify where they are leveraging their strengths and where they are suffering from key weaknesses according to their customers / client / employees e.g. which of perceptions of service, variety of product type or price is the stronger driver of customer satisfaction? or which of salary, management support or corporate communication drives employee loyalty within the company?
Respondents are presented with a list of different aspects of service and / or product information which they rate in terms of satisfaction / dissatisfaction. The analysis identifies which of these issues cause an increase or decrease in overall satisfaction levels (the issues can be ranked in order of impact or importance to customers)
If you feel any of these techniques would be useful to you, or if you would like some further advice on general analysis of your data please get in touch with Rainbow Research on +44 (0) 1772 743235.