:: Home :: Academic :: About :: Contact :: FAQs :: Site Map ::  
Client Login:Password:
Statistical Reasoning... Empowering Data-Driven Decisions Statistical Reasoning... Empowering Data-Driven Decisions
Design Develop Administer Analyze Model Publish :: Demonstrations ::   
:: RESEARCH DESIGN ::
  Research Design
  Sampling Design
  Questionnaire Design
     Instrument Pre-Testing

:: LOYALTY METRICS ::
Evaluating Customer Loyalty Metrics - Three Part Series

The loss of customer loyalty is alerting and sensitizing managers to its importance. Consequently, customer satisfaction measurement has been supplanted by the concept of customer loyalty, and so-called Customer Loyalty Indices [CLI] have emerged out of a need to better understand customer retention.

In specific, most CLIs are made up of behaviour-based metrics, such as "likelihood to recommend a product or service to others" and "likelihood to repurchase the product or service," and attitude-based metrics such as "overall satisfaction." The evaluation of such CLIs formed the basis for this three-part discussion.

Please contact SR should you wish to receive more information when this article becomes publicly available.
 
:: Design|Questionnaire Design ::
:: Request a Bid ::
Research DesignSurvey DevelopmentProject AdministrationData AnalysisMarketing ModelingPublish in the SR Journal™
Statistical Reasoning takes a reverse engineering approach to
     questionnaire design by envisioning how the data will be used.

Given that the information gaps and research objectives have been identified, defined, and delimited, the next step is to design a questionnaire capable of producing the data required to answer the research questions.  The survey questionnaire essentially operationalizes the research objectives through a set of standardize questions and responses.

Proper questionnaire design requires planning what needs to be measured, formulating the questions to obtain needed information, deciding on the order for asking the questions and the layout of questionnaire.

Statistical Reasoning's approach to questionnaire design is to reverse engineer the process.  We begin by asking what data is required and in what form.  By envisioning how the data will be used, SR ensures that the questionnaire will be designed appropriately and in its most complete form.  In other words, that all the right questions are asked and asked in a manner that generates the best type of measurement for analysis purposes.

For instance, rather than asking whether the respondent is merely satisfied or dissatisfied (nominal scale) with the company's service, it would be more prudent, informative, and accurate from an analysis standpoint to ask the respondent to rate the extent of his/her satisfaction or dissatisfaction (interval scale).

Additionally, SR's philosophy is to get as much information from the survey as feasibly possible without alienating the respondent.  In the above example, SR would propose a follow up question to elicit insights as to possibly what aspects contribute to the respondents satisfaction or dissatisfaction.

Indeed, SR has a wide range of experience in various content areas, including customer satisfaction, loyalty, employee/job satisfaction, and service quality.  In fact, SR's proprietary customer measurement scale includes not only behavioural and attitudinal proxies, but also captures elements of brand trust and commitment, for a truly holistic perspective of your customers' loyalty (see also the loyalty metrics discussion to the left).

If needed, a round of pre-testing will be conducted with a sub-set of the respondent group to ensure the integrity and usability of the survey instrument. Click Here to Learn More

Click here to learn more about the features and capabilities of SR's online surveys. Click Here to Learn More

Click here to learn more on how Statistical Reasoning can help you manage the online survey project from delivery of the email invitations to delivery of the final data file, and everything in between. Click Here to Learn More

@ :: MR INSIGHTS ::
A scale’s level of measurement is defined as the amount of information embedded in the data it yields.  Measurement scales can have one or more of these characteristics:
  • Description: unique label or descriptor
  • Order: relative ‘sizes’ of descriptors
  • Distance: absolute differences between descriptors
  • Origin: unique beginning or true zero point

And of course, higher levels of measurement contain more information.  That is, a Nominal Scales (categorical, discrete) has only the characteristic of description, but not of order, distance or a true zero point.
Ordinal Scales (rank-order) have the added characteristic of relative 'size' and order but not distance or a true zero point.
Interval Scales have equal intensity intervals between scale points in addition to the qualities of the Nominal and Ordinal Scales.  Ratio Scales have all four characteristics.

Ultimately, the amount of information embedded in the data determines what we can or cannot say about construct being measured.  Moreover, the type of scale determines, in part, the statistical analysis techniques that can be used.

A firm grasp of measurement scales can make all the difference in for a well-tuned and usable survey instrument.

   
Contact us to discuss your project: info@statisticalreasoning.com.:: Back to Top ::
  :: Home :: Academic :: About :: Contact :: FAQs :: Site Map ::
:: Design :: Develop :: Administer :: Analyze :: Model :: Publish :: Demonstrations ::

Best viewed with Microsoft Internet Explorer at 800 x 640 resolution or higher with Javascript enabled
© 2003 Statistical Reasoning Research & Analytics, Inc.  All Rights Reserved.  [Privacy Policy]