Customer Satisfaction Pilot Studies and Analysis

Dispersion of Responses (Continued)

Figure 3: Distribution of Employer ACSISAT Scores

figure 3

The employer distributions (Figure 3) indicate a different distribution pattern. Only two of the four states have large numbers of highly positive ratings. No state has the highest frequency of responses in the highest rating category. There are many more moderate ratings. In State B and State D, there are also a small but significant number of very low scores, suggesting there might be a geographic region or group of employers where services are particularly unsatisfactory.

Standard Deviation.

To present the pattern or distribution of numbers with a single numeric indicator, one can take all the survey question responses for a given State or customer group and calculate the Standard Deviation. The Standard Deviation indicates how scores fluctuate on either side of the average score. The State E ACSISAT average is 72.08 with a standard deviation of 27.5. This calculated standard deviation means that approximately 68% of the customer responses were between 44.58 and 99.58 (i.e., 72.08 plus or minus 27.5).

Strengths. The standard deviation is easily calculated in MS Excel and most spreadsheets. This one number provides some important information to complement the average. A standard deviation indicates that the two sets used as an example above ({6, 6, 6, 6} and {2, 2, 10, 10}), which had identical averages are very different. For the first set the standard deviation is 0, indicating that all the elements of the set occur at the mean or average. For the second set, whose mean is identical to the first, the standard deviation is 4.6, indicating that the elements of the set are widely dispersed about the average. If these two sets of numbers represented customer satisfaction scores, they would depict two very different pictures of satisfaction and present the service providers with very different issues to address. In the first group (all responded 6), no one feels strongly positively or negatively about the service. The service providers might gather additional information to determine why all customers respond in a lukewarm manner to their service. In the second group ({2, 2, 10, 10}), half have a strongly negative response to the service and half have a strongly positive response to the service. The service providers might gather additional information to determine why half their customers are very displeased and whether different groups treated differently. What gaps in service can be addressed so that fewer customers rate their satisfaction so low?

Weaknesses. A standard deviation is not easily understood. An individual is not likely to intuitively understand a standard deviation unless he or she has studied statistics. Even those who have taken a statistics course will often not understand the importance of the number and how it might inform interpretation of results (they may only remember the pain of hand-calculating the number). These difficulties with understanding the standard deviation may mean that important information regarding dispersion of responses around the mean might be overlooked.

Primary Audience
Standard deviations with averages are appropriate for all audiences when accompanied by a detailed explanation. This required explanation may limit the degree to which the audiences use the information and so limits the likely audience to those with a stake in a more detailed understanding of the numbers, such as operational managers and staff.

Continue>>
 
 
 
Home | About | Resources | Services | Forums