The current value obtained from the data collected is not a valid summary as the number assigned to the various responses may not be considered indicative of the exact satisfaction scale. For instance, the difference between the strongly disagree and disagree is the same as that between disagree and not sure. This is rather a convenient way of simplifying the given data which does not accurately capture the satisfaction level of the consumers. The weights assigned to each of the response should be in line with the utility derived by the customer rather than some random numbers driven by convenience of computation (Hillier, 2006). Further, in the given case, it may so happen that the same score is available for a host of combinations of responses of customers which makes it difficult to draw any conclusion. A better mechanism to present the given data is through assigning of numerical scores which capture the utility from consumer’s end to each of the responses. Then the same technique should be applied, however the number obtained would be more representative of the actual responses of the customers (Swarup, Gupta & Mohan, 2010).
No, the data collected in this mechanism i.e. through website would not be reflective of all the customers of the store. This is because it is highly likely that young customers who are more tech savvy are over presented in such surveys while the older generation which prefers paper based surveys may be underrepresented in such surveys. The actual accuracy of the results in this case would depend on the sample filling up the online survey being representation of the actual customer base of the store (Lieberman et. al., 2011). There is only a very thin possibility of this happening and typically there would be misrepresentation of various sections. Also, in order to draw some meaningful conclusions, the company should allow for enough responses to be collected through the online survey and also focus on the demographics of the participant. This would provide a fair idea to the company as to whether the sample of customers filling up the questionnaire is a valid and reliable sample of the actual customers (Hastie, Tibshirani & Friedman, 2001).
1. The type of data would be nominal since the variables are serving the purpose of labelling (male or female) and do not capture any particular quantitative value (Swarup, Gupta & Mohan, 2010).
2. The Fahrenheit thermometer would be an example of interval data scale since the various values are represented on a numeric scale in a proper order and further the differences between two given values is also known. However, the pivotal aspect is that Fahrenheit scale does not have a true zero and hence it is not considered as ratio data (Hillier, 2006)
3. The Kelvin thermometer would be an example of ratio scale since unlike in the previous case, a true zero does exist for Kelvin thermometer. Besides the other conditions are also met such as various values being represented in a proper order on a numeric scale and further the differences between two given values is also known (Hastie, Tibshirani & Friedman, 2001).
4. The number of items bought would also be an example of ratio scale as true zero is defined, data would be numeric and arranged and also comparison can be made.
5. Bank account balance is an example of ratio scale as true zero is clearly defined which implies that there is no money in the account. Also, the data would be in numeric form and further comparisons may be made between values that can be arranged in a particular format (Hillier, 2006).
A descriptive non-experimental study in the given context can be done by giving the players of the local team with orange juice over four days preceding the weekend with a frequency of thrice daily and then observing if their performance has indeed significantly improved over their performance during the last weekend (Lieberman et. al., 2011).
A quasi experimental study in the given context can be performed by dividing the local team players randomly into two groups where one group is the intervention group and the other is comparison group. The intervention group is given orange juice over four days preceding the weekend with a frequency of thrice daily while the comparison group is not given orange juice. The performance of these two groups is then compared over the weekend so as to ascertain whether orange juice is indeed effective (Hastie, Tibshirani & Friedman, 2001).
A experimental study would involve making two groups with comparable baseline. Hence, the intervention group and the control group both should have players with matched ability and performance. The intervention group is given orange juice over four days preceding the weekend with a frequency of thrice daily while the control group is not given orange juice. The performance of these two groups is then compared over the weekend so as to ascertain whether orange juice is indeed effective (Hillier, 2006).
The experimental study is the most superior option since unlike quasi experimental study it accounts for the incorrect grouping where there may be difference between the ability and performance of players even before orange juice is given. Further, experimental study is superior to non-experimental study as it is more controlled and would amount to minimal influence of other factors such as training and weak opponent (Hastie, Tibshirani & Friedman, 2001).
Hastie, T, Tibshirani, R & Friedman, J 2001, The Elements of Statistical Learning, Springer Publications, New York
Hillier, F 2006, Introduction to Operations Research, McGraw Hill Publications, New York
Lieberman, FJ, Nag, B, Hiller, FS & Basu, P 2011, Introduction To Operations Research, Tata McGraw Hill Publishers, New Delhi
Swarup, K, Gupta, PK. & Mohan, M 2010, Operations Research, Sultan Chand & Sons, New Delhi