Write an academically referenced paper that you could present to Sfinx Management as a consultant, in which you answer the following questions about
Write an academically referenced paper that you could present to Sfinx Management as a consultant, in which you answer the following questions about the Sfinx problem, while using the DMT Case Study Sfinx Online.xlsx, and Sfinx SPSS data:
1. Conduct exploratory factor analysis on the potential scale made up of the Customer service items: X6-X18, in order to examine whether it is actually unidimensional or not (make sure that any negatively worded items have been reversed before proceeding with this analysis). For each analysis report the following: Kaiser-Mayer-Olkin Measure of Sampling Adequacy, Bartletts Test of Sphericity, total variance explained, eigenvalue(s) of retained factors, and save and NAME the factors if there is more than one.
2. Assess the emerging factor loadings for each scale item and decide whether any items are performing poorly and should therefore be dropped from further analysis. With the remaining items estimate Cronbach alpha coefficient and item-total correlation per factor. Report the reliability (Cronbach alpha) for each factor you discovered. Also, based on the item-to total correlation analysis decide whether any other items should be dropped from each scale.
3. Calculate average scales (composite scores) for the customer service factors that you discovered in Question 1. Use all items that were retained following your exploratory factor analysis, and leave-one-out or item-to-total correlation procedures. For each summated scale report the average score and standard deviation and give it a meaningful name, dependent on the variables that load on it.
4. Using the average scales created above, perform test analyses (which one is appropriate?) to examine whether there are any significant differences in the levels of X19 Customer service :
a) males and females
b) Customer type,
c) Payment method
5. Report exactly what tests you did, what their null hypotheses2 are, what p-values you find, and whether you reject the null Hypotheses or not, and what then is the conclusion in non-statistical jargon.
6. Draw a conceptual model of demographic variables X1-X5 and the customer service quality factors you found above influencing X19 overall customer satisfaction. Do you think you should use moderation?
7. Estimate a regression for that Conceptual model and distil the essential factors (what do you base essential on)?
8. Draw your resulting conceptual model,
9. For each regression model,
a. specify if you chose for simple or multiple regression,
b. assess the level of variance explained and
c. the significance of beta coefficients.
10 Discuss the main conclusions that can be derived from this analysis, regarding the influence of the service factors on customer satisfaction.
11 Given the Cost Sheet in the Sfinx MS Excel file, which explains which service quality variable improvement costs how much, which variables investment would you change? What would be the effect? Would you stay Budget-Neutral? (Note: the coefficients are only valid in a range around the current data, so cranking one of the variables from 3 to 10 would probably not have a linear effect on the DV.)
Submissions should be a report you could present to managers (interested in arguments for conclusions and recommendations) in single spacing, APA style referencing.
Table Of Contents
1 Introduction 2 Problem Statement 3 Research Questions 3.1 Minor Research Questions 3.2 Major Research Questions 4 Data Analysis Procedure 4.2 Market Demographics 5 Digital Media Approach 6 SWOT Analysis 7 Value Equation 7.1 Value Equation With TM 8 Marketing Mix 9 Conclusion 10 Reference list 11 Appendixes Introduction
The strong link between how happy customers are and the success of a business is widely recognized. The happiness of customers plays a vital role in ensuring a business thrives, affecting whether customers keep coming back, their loyalty to the business, and their likelihood to recommend the business to others. To measure and improve customer satisfaction in a numerical way, businesses are turning more and more to methods of data analysis. For example, factor analysis is used to discover the key elements that affect how satisfied customers are. The Cronbach's alpha test is crucial for checking how reliable the measurement tools are for gauging customer satisfaction, making sure the data is consistent. Likewise, the Kaiser-Meyer-Olkin (KMO) test checks if the data is good enough for factor analysis, helping confirm the analysis is valid. Regression analysis is also key for figuring out how customer satisfaction relates to different outcomes for the business, helping companies forecast how changes in satisfaction might affect them. By using these data analysis tools, companies can gain useful insights from their data, leading to better decisions that boost customer satisfaction and, in turn, the success of the business.
Problem Statement
The central issue addressed in this data analytics assignment is determining the influence of various independent variables, namely quality, technical support, advertising, and competitive pricing, on the dependent variable, which is customer satisfaction. This problem statement underscores the necessity for a detailed examination using data analytical techniques to unravel how these distinct factors contribute to the overall satisfaction of customers. Employing statistical methods such as regression analysis will enable the quantification of the impact each independent variable has on customer satisfaction. Furthermore, correlation analysis will be utilized to assess the strength and nature of the relationships between these variables. The outcome of this analysis is expected to provide invaluable insights, allowing businesses to make informed decisions on how to optimize these variables to enhance customer satisfaction, thereby achieving better business outcomes and improving customer-centric strategies.
Research Questions
Minor Research Questions :Does the impact of customer satisfaction with past purchases on perceived quality differ based on the type of product purchased?(
To what extent does the influence of delivery time on customer satisfaction with the product vary depending on the customer's individual delivery preferences?
Does the moderating effect of price on the relationship between customer satisfaction with past purchases and customer satisfaction with the product change based on the income level of the customer?
Major Research Questions :To what extent does customer satisfaction with past purchases from Sphinx directly and indirectly influence customer satisfaction with the product, and how do these relationships vary based on different customer and product characteristics?
Data Analysis Procedure
4.1 Correlation Analysis
-653415000 INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/b86a0a5c-b5b3-4022-99d5-704aa65f2994" * MERGEFORMATINET
The table shows the correlation coefficients between various customer satisfaction measures and other variables, including perceived quality, advertising, price, and delivery time. It appears that there are statistically significant positive correlations between most of these variables and customer satisfaction. For example, Customers tend to be more satisfied with their previous purchases from Sphinx when they believe the quality is high, as shown by a strong positive link (0.500) between quality perception and satisfaction. On the other hand, as prices go up, customer satisfaction tends to go down, which is indicated by a noticeable negative relationship (-0.208). When it comes to how much Sphinx is seen to market and offer new products, the connection with customer satisfaction is quite small but still positive (0.071). Additionally, there's a robust positive connection (0.522) between customers' satisfaction and their belief that Sphinx manages the ordering and billing process well and without errors.
4.2 Factor Analysis
The table shows that the data is good for factor analysis because the Kaiser-Meyer-Olkin (KMO) measure is .694, which is higher than the minimum acceptable level of .5. Also, the Bartlett's test is important (with a p-value less than .001), which means the data has enough variety for factor analysis to be useful. These tests tell us that it's okay to go ahead with factor analysis on this data set.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/8a807d92-ebbc-4cd2-8f77-d29c2200a0a8" * MERGEFORMATINET
The table shows that the data is good for factor analysis because the Kaiser-Meyer-Olkin (KMO) measure is .694, which is higher than the minimum acceptable level of .5. Also, the Bartlett's test is important (with a p-value less than .001), which means the data has enough variety for factor analysis to be useful. These tests tell us that it's okay to go ahead with factor analysis on this data set.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/3b229682-364b-4bb9-9616-7c493961302b" * MERGEFORMATINET
The initial factor explains 36.77% of the variation in the dataset, the second factor explains 27.29%, and the third factor is responsible for 12.66% of the variation. Combined, the first and second factors account for 64.06% of the total variation. it is often beneficial for the first few factors to explain a significant portion of the variation, but the exact amount considered significant can change based on the goals of the research and the number of variables being analyzed.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/3d112abe-0e7c-4af4-9827-154e628419d6" * MERGEFORMATINET
The first two factors stand out because they have much higher eigenvalues compared to the others. Also, if you look at the scree plot, which is a graph that shows the importance of each factor, there's an obvious bend or "elbow" after the second factor. This suggests that the factors after the second one don't add as much new information.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/ADjfeTeTb3sC6V34NLo2qK8q5mkH0Nh1-SqrtjT7OQx4P3XwYRQCwj49DOISGm8uOV-JSMXBQX_enZu7pJObylC8jSaSsaP6OBQ2U0Eb7MHgiJo0ju2PNe-TxuduFZSwAKGdrXqGHOLEJD7Q0oJCiTVf2s2rZSAQ0FHi8_INA4iCrNIS1u20LyurmAhzEzcbXqQdJBUEXPc-nwD3j5rUQoJ0YIveXORSs2Pvb_CwTSh84PoPRSweZu4DJq2p3FCLb2ZbnBvSu_Q7GBdhIvLNgMwsMAEEGIGAmA=d?authuser=1" * MERGEFORMATINET INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/7b185066-b6ee-42ec-beb2-6ed9cb99b070" * MERGEFORMATINET
The results suggest that there are three underlying factors that explain a significant portion of the variance in the customer satisfaction data. These factors are related to perceived quality and efficiency, price and website usability, and perceptions of advertising and order tracking.
4.3 Cronbach Alpha
Factor 1 :
The Cronbach's alpha scores for the three customer satisfaction metrics fall between 0.780 and 0.869, according to the table. This range indicates that these metrics have a satisfactory level of internal consistency, implying that they are fairly reliable in assessing what they are intended to measure. However, it should be noted that higher scores would point to even greater dependability.
Factor 2 :
The Cronbach's alpha numbers in the table go from 0.556 to 0.840. These values measure how well the survey questions work together within a certain topic. For the satisfaction with the online order tracking system (0.703) and how people see Sphinx's ads across different media (0.840), the scores are within an okay range. This means these areas are consistently measured. But, the score for how people view the website of Sphinx, including its appearance and how easy it is to use (0.556), is below the generally advised level of 0.6.
Factor 3:
The Cronbach's alpha score for how customers perceive the quality of the products they receive, which includes the external and internal packaging as well as the product quality itself, is 0.477. This score is a statistical measure used to determine the consistency of survey responses related to these aspects of product quality.
4.4 New Independent Variables Weighted Mean
Cretaing the new indepented variables for the each factor by calculating the weighted mean and proceeding with the correlations analysis.
Executing correlations analysis with the new independent variables against the dependent variable.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/b0874f85-9347-49f2-b381-81603c052a0f" * MERGEFORMATINET
Using a 10-point scale, there's a positive relationship (0.614) between how happy customers are with their previous buys from Sophin and their satisfaction with the product itself. This implies that if customers are happy with their past purchases, they're likely to be happy with the product too. Additionally, there's a positive link between being satisfied with the product and being happy with how quickly it's delivered (0.437) as well as the price of the product (0.602). This indicates that customers who like the product tend to also be pleased with its delivery speed and cost.
4.5 Estimated Conceptual Model
The estimated conceptual model is created from the new independent variables is based on the customer satisfaction with past purchases from Sphinx, which is calculated as a weighted mean. Additionally, customer satisfaction with past purchases is thought to have a positive indirect effect on customer satisfaction with the product, mediated by perceived quality, delivery time, and price.
4.6 Hypothesis
4.6.1 Gender:
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/fc5621e0-f868-4799-a0cc-eff014ee2f62" * MERGEFORMATINET
Null Hypothesis : No effect or No difference
Alternative Hypothesis : Effect or difference
Significance Level (alpha) : 4%
The null hypothesis cannot be dismissed. While descriptive statistics reveal a marginally higher average satisfaction score for women compared to, this disparity is not statistically meaningful. The ANOVA findings, coupled with the descriptive statistics, do not support the notion that customer satisfaction varies between genders. This suggests that individual differences within each gender category contribute more to the overall variation in customer satisfaction scores than the differences between the gender categories.
4.6.2 Customer Type
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/53836b4c-2096-46e4-a5b2-b7224ad7ce59" * MERGEFORMATINET
Null Hypothesis : No effect or no difference
Alternative Hypothesis : Effect or difference
Signifance Level : 4%
The data indicates that long-term customers of Sphinx, those who have been with the company for over five years, exhibit the highest average satisfaction level. This group is followed by customers with a tenure of one to five years , and the newest customers, with less than one year with Sphinx, showing the lowest satisfaction. The ANOVA findings reveal a significant variance in satisfaction among these customer groups. The disparity in satisfaction across different types of customers surpasses the variance within each type. This statistical evidence supports the acceptance of the alternative hypothesis over the null hypothesis, indicating a notable difference in satisfaction levels across the various customer segments.
4.6.3 Payment Method
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/1e0135bc-f21e-474a-8c39-801bd0853b9f" * MERGEFORMATINET
Null Hypothesis : No effect or no difference
Alternative Hypothesis : Effect or difference
Signifance Level : 4%
The analysis revealed that customers who paid by cash on delivery reported higher satisfaction compared to those who paid with credit cards. The statistical results indicated a significant difference in satisfaction levels based on the chosen payment method. The effect size (eta-squared) suggested that the method of payment accounts for approximately 27% of the variation in customer satisfaction scores. Given that the p-value is below the conventional threshold of 0.05, the null hypothesis is rejected, supporting the alternative hypothesis that there is a meaningful difference in satisfaction between the payment methods.
The newly created variables, has a positive relationship with satisfaction regarding the current product. This indicates that customers who were happy with their previous purchases are likely to be happy with their current product too.In factor analysis, it seems that the factors might be linked to perceived quality and service efficiency. The satisfaction with past purchases could be reflecting these aspects, thereby influencing satisfaction with the current product.This idea fits well with what we know in marketing - if people are happy with their previous buys, they tend to trust the brand more, see the quality in a better light, and are more likely to buy again, all of which can boost their current product satisfaction. So, with these correlations, the possible connections with important factors, and the backing of marketing theories, this new variable of past purchase satisfaction, calculated as a weighted average, seems promising for more detailed studies on customer satisfaction. But remember, this is just a theory for now, and more research is needed to fully understand these relationships.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/83c11a76-ce55-4180-bae1-e46dc2948699" * MERGEFORMATINET
4.7 Regression Analysis
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/289dc8ae-cb46-4d35-be6b-2d744beaa9e6" * MERGEFORMATINET
Whenperceived quality, delivery time, and priceare added to the model, the explained variance (R-squared) increases to42.3%.Perceived qualityhas asignificant positive effecton customer satisfaction with the product ( = 0.244, p-value < 0.001).Pricehas asignificant negative effecton customer satisfaction with the product ( = -0.117, p-value = 0.029).As the price of the product increases, customer satisfaction tends to decrease.This aligns with the concept of price sensitivity, where customers are generally less willing to pay more for a product.
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/47adaaf7-b48e-4c89-9c91-d6f7b14b64d3" * MERGEFORMATINET
The data shows that being happy with past purchases significantly boosts satisfaction with the product ( = 0.510, p-value < 0.001). This means if satisfaction with past purchases goes up by one point, satisfaction with the product goes up by 0.510 points, when we don't change anything else.
The analysis also indicates that how customers perceive the quality ( = 0.244, p-value < 0.001) and the price ( = -0.117, p-value = 0.029) significantly influence their satisfaction with the product. However, the relationship with price is negative, suggesting that when prices go up, satisfaction with the product usually goes down.
The model is pretty good at explaining how satisfied customers are with the product, accounting for 42.3% of the changes in satisfaction (R-squared = 0.423).
INCLUDEPICTURE "/Users/ullash/Library/Group Containers/UBF8T346G9.ms/WebArchiveCopyPasteTempFiles/com.microsoft.Word/2a709f51-42bc-454b-a319-91aab3267c9c" * MERGEFORMATINET
Satisfaction with past purchases plays a key role in maintaining customer loyalty. A detailed analysis using SPSS for multiple linear regression revealed that the model could explain 70.9% of the variance in customer satisfaction (F(5, 94) = 45.915, p <.001). Examining specific variables, it was found that customer type ( =.615, p <.001), age group ( =.242, p =.001), and payment method ( =.496, p <.001) were significantly correlated with customer satisfaction. However, gender ( = -.034, p =.544) and region ( =.104, p =.158) did not show a significant relationship with satisfaction levels.
The positive coefficients for customer type, age group, and payment method suggest that as these factors increase, so does customer satisfaction. The model, with an adjusted R-square of .694, indicates that these demographic factors explain approximately 69.4% of the variance in customer satisfaction.
The analysis suggests that factors such as customer type, age, and payment method are crucial in determining satisfaction, while gender and region appear to be less influential. These insights should guide businesses in tailoring their marketing and service strategies to align with the significant demographic factors to enhance overall customer satisfaction. While the study shows that gender and location are not major determinants in this scenario, it's essential to recognize that these factors might still play a role in different contexts or industries. The model's strength lies in its ability to account for multiple variables, providing a comprehensive view of what influences customer satisfaction.
In summary, the most significant indicator of current product satisfaction is the customer's satisfaction with previous purchases. However, the perceived quality and the price of the product also significantly impact their satisfaction. These insights are valuable for developing strategies to enhance customer satisfaction throughout various phases of the customer journey. This includes creating positive previous experiences, making sure the product quality aligns with customer expectations, and being mindful of customers' price sensitivity.
Conceptual Model
Conclusion & Recommendations
In conclusion, Sphinx's strategic focus on developing new products or services based on customer tenure, alongside a loyalty program, is poised to elevate customer satisfaction and strengthen brand loyalty. By prioritizing customer type, gender, and payment method over age segmentation, Sphinx aligns its resources with key factors that drive customer satisfaction, setting the stage for enhanced customer experiences and business growth.
The organization is encouraged to unveil a new product or service that caters to the duration of customer engagement with the company. Given the significant findings, initiating a loyalty scheme to boost overall satisfaction and foster enduring connections with the brand is a well-supported strategy.The data analysis suggests that segmenting customers based on their age group does not substantially influence their satisfaction levels. As such, it would be prudent for Sphinx to eschew age-based segmentation in favor of more impactful criteria.
An assessment of various independent factors and their linkage to the dependent variable X19 reveals that these factors are aligning with market benchmarks and do not require urgent adjustments.To enhance investment returns and achieve strategic goals, the organization should prioritize areas such as customer type, gender, and payment method. These dimensions have demonstrated influence over customer satisfaction levels and could inform the development of new products or services.
By concentrating on these strategic areas, Sphinx can more effectively allocate its resources towards elements that significantly affect customer satisfaction, thereby augmenting the customer experience and potentially boosting brand loyalty and revenue growth.
References
Appendix
Individual assignmentDescription individual assignmentWrite an academically referenced paper that you could present to Sfinx Management as a consultant, in which you answer the following questions about the Sfinx problem, while using the DMT Case Study Sfinx Online.xlsx, and Sfinx SPSS data:
1. Conduct exploratory factor analysis on the potential scale made up of the Customer service items: X6-X18, in order to examine whether it is actually unidimensional or not (make sure that any negatively worded items have been reversed before proceeding with this analysis). For each analysis report the following: Kaiser-Mayer-Olkin Measure of Sampling Adequacy, Bartletts Test of Sphericity, total variance explained, eigenvalue(s) of retained factors, and save and NAME the factors if there is more than one.
2. Assess the emerging factor loadings for each scale item and decide whether any items are performing poorly and should therefore be dropped from further analysis. With the remaining items estimate Cronbach alpha coefficient and item-total correlation per factor. Report the reliability (Cronbach alpha) for each factor you discovered. Also, based on the item-to total correlation analysis decide whether any other items should be dropped from each scale.
3. Calculate average scales (composite scores) for the customer service factors that you discovered in Question 1. Use all items that were retained following your exploratory factor analysis, and leave-one-out or item-to-total correlation procedures. For each summated scale report the average score and standard deviation and give it a meaningful name, dependent on the variables that load on it.
4. Using the average scales created above, perform test analyses (which one is appropriate?) to examine whether there are any significant differences in the levels of X19 Customer service :a) males and females
b) Customer type,
c) Payment method
5. Report exactly what tests you did, what their null hypotheses2 are, what p-values you find, and whether you reject the null Hypotheses or not, and what then is the conclusion in non-statistical jargon.
6. Draw a conceptual model of demographic variables X1-X5 and the customer service quality factors you found above influencing X19 overall customer satisfaction. Do you think you should use moderation?
7. Estimate a regression for that Conceptual model and distil the essential factors (what do you base essential on)?
8. Draw your resulting conceptual model,
9. For each regression model,
a. specify if you chose for simple or multiple regression,
b. assess the level of variance explained and
c. the significance of beta coefficients.
10 Discuss the main conclusions that can be derived from this analysis, regarding the influence of the service factors on customer satisfaction.
11 Given the Cost Sheet in the Sfinx MS Excel file, which explains which service quality variable improvement costs how much, which variables investment would you change? What would be the effect? Would you stay Budget-Neutral? (Note: the coefficients are only valid in a range around the current data, so cranking one of the variables from 3 to 10 would probably not have a linear effect on the DV.)
Submissions should be a report you could present to managers (interested in arguments for conclusions and recommendations) in single spacing, APA style referencing.
Deliverables: Course LMS upload with
A maximum 3000 words long single spacing text reportfor your boss (i.e. normal managerial language!), in preferably ODT, Word, or if the previous are not available PDF format, with
a small intro,
a description of the research problem and research questions,
description of the data preparation process
descriptive statistics (graphically and numerical) of the main variables, make sure the visuals make sense and clarify matters.
description of whatproceduresyou did and why they are appropriate,
their results,
interpretations of the results
conclusion and recommendations
references APA style (10 minimum).
Upload any workout files, twb(x) or xlsx, and publish AzureML models, copy the link into the word document as reference! andshareboth the link and the model with me (v.feltkamp@maastrichtuniversity.nl).
lit review table with coding and URLs of the papers found (doi being preferred).
Tableau file for the visualizations andOrange3, AzureML (link), Realstatistics or other program analysis output as appendix, zip all this together
your data if you collected data.
In case of submission problems, timing is flexible but the submission method is not.In no case will an email submission be allowed. Discuss problems with the lecturer at their email address, or in the course discussion forum.
Assessment rubric individual assignmentPlease only change this section after discussion with the Academic Course Coordinator. In case you adapt this assessment rubric, it also needs to be changed in Moodle as all grading is done directly in Moodle. In this case, please contact the responsible programme administrator for support.
DD Weigh-tingAssessment Criteria Expected performance standards
1 point (very poor) 2 points (unsatisfactory) 3 points (satisfactory) 4 points (good) 5 points (outstanding)
DD1 20% no effort Description of applicable methods not complete or faulty Basic description of methods methods definitely applicable, but descriptions not precise Clearly describes methods that are applicable and useful
DD2 20% no effort application of the methods has severe problems. Basic application, with some errors application of the methods are mostly but not totally correct clearly describes a correct application of the methods chosen.
DD3 20% no effort argumentation is weak, lacks logic, and recommendations do not follow from argumentation Argumentations basically correct. argumentation ok, recommendations dont follow from them or argumentation weak and recommendations make sense argumentation inevitable, recommendations totally to the point and practical
DD4 20% no effort write-up not easy to follow, formatting errors severe. Writeup readable. Language mostly ok but there are some problems with English, or precise phrasing. Write-up uses the correct Statistical phrasing without inventing confusing expressions.
DD5 20% no effort has not made much of an effort to self educate on the topic and methods. Some self study applied, but lacunas exist. Uses literature to find most of the appropriate procedures. extensive literature use and goes way beyond the concepts of the course to learn the appropriate procedures in detail.
Write an academically referenced paper that you could present to Sfinx Management as a consultant, in which you answer the following questions about the Sfinx problem, while using the DMT Case Study Sfinx Online.xlsx, and Sfinx SPSS data:
1. Conduct exploratory factor analysis on the potential scale made up of the Customer service items: X6-X18, in order to examine whether it is actually unidimensional or not (make sure that any negatively worded items have been reversed before proceeding with this analysis). For each analysis report the following: Kaiser-Mayer-Olkin Measure of Sampling Adequacy, Bartletts Test of Sphericity, total variance explained, eigenvalue(s) of retained factors, and save and NAME the factors if there is more than one.
2. Assess the emerging factor loadings for each scale item and decide whether any items are performing poorly and should therefore be dropped from further analysis. With the remaining items estimate Cronbach alpha coefficient and item-total correlation per factor. Report the reliability (Cronbach alpha) for each factor you discovered. Also, based on the item-to total correlation analysis decide whether any other items should be dropped from each scale.
3. Calculate average scales (composite scores) for the customer service factors that you discovered in Question 1. Use all items that were retained following your exploratory factor analysis, and leave-one-out or item-to-total correlation procedures. For each summated scale report the average score and standard deviation and give it a meaningful name, dependent on the variables that load on it.
4. Using the average scales created above, perform test analyses (which one is appropriate?) to examine whether there are any significant differences in the levels of X19 Customer service :
a) males and females
b) Customer type,
c) Payment method
5. Report exactly what tests you did, what their null hypotheses2 are, what p-values you find, and whether you reject the null Hypotheses or not, and what then is the conclusion in non-statistical jargon.
6. Draw a conceptual model of demographic variables X1-X5 and the customer service quality factors you found above influencing X19 overall customer satisfaction. Do you think you should use moderation?
7. Estimate a regression for that Conceptual model and distil the essential factors (what do you base essential on)?
8. Draw your resulting conceptual model,
9. For each regression model,
a. specify if you chose for simple or multiple regression,
b. assess the level of variance explained and
c. the significance of beta coefficients.
10 Discuss the main conclusions that can be derived from this analysis, regarding the influence of the service factors on customer satisfaction.
11 Given the Cost Sheet in the Sfinx MS Excel file, which explains which service quality variable improvement costs how much, which variables investment would you change? What would be the effect? Would you stay Budget-Neutral? (Note: the coefficients are only valid in a range around the current data, so cranking one of the variables from 3 to 10 would probably not have a linear effect on the DV.)
Submissions should be a report you could present to managers (interested in arguments for conclusions and recommendations) in single spacing, APA style referencing.
