research design and data collection tools
.References Style(APA Ã¢â‚¬â€œHarvard-Chicago-Others): harvard
This has high mark, and total 6 questions, using spss.
You can log on utsonline,get more documents, I you need
Subject: research design and data collection tools
The following is q3-q6 guide:
Ã¢â‚¬Â¢ Question 3: EFA and Reliability Analysis
You need to run EFA (Exploratory Factor Analysis) for each construct, one at a time first in order to purify the measure and assess convergent validity. That is, to see (1) whether the items for the construct are loaded on one factor as expected and (2) whether each item's factor loading is substantially big (i.e., > 0.5, preferably >0.7). In addition, you need to check the overall KMO and individual KMO values are greater than 0.5. If the above conditions are satisfied, then you have evidence in support of convergent validity of the multi-item reflective measure. You may need to drop non-performing items and run the EFA again.
After you run EFA for each construct, you need to run EFA for all the conceptually related constructs together in order to further purify the measure and assess discriminant validity. That is, to see (1) whether the items are loaded on their respective constructs as expected and (2) whether each item's factor loading (not cross-loadings) is substantially big (i.e., > 0.5, preferably >0.7). In addition, you need to check the overall KMO and individual KMO values are greater than 0.5. If the above conditions are satisfied, then you have evidence in support of discriminant validity of the multi-item reflective measure.
Finally, run reliability analysis for each factor, one at a time. Do reverse-coding, if necessary (when the factor loadings for the factor have opposite signs) prior to running reliability analysis.
Question 4: Positining Map
Using notes given in Week 03 to run multidimensional scaling (MDS) to get the map. Alternatively, use exploratory factor analysis (EFA) to get the map.
Question 5: Traditional Conjoint and Best-Worst Conjoint
For Question 5 Part A: Traditional Conjoint, just rank-order your 16 bundles and enter the ranking in SPSS. Follow my handout notes to run Linear Regression Model (with effects coding) to get the results.
For Question 5 Part B: Best-Worst Conjoint, you need to fill out your best-worst conjoint survey first and then count the number of times each bundle has been selected as Best and Worst. Based on these Best and Worst counts, you can compute BMW scores for the 16 bundles and enter these BMW scores into the SPSS data file used in Question 5 Part A. Now use BMW as the depenent variable to replace Utility and follow the same subsequent procedure to get the results.
Question 6: BMW Score Computation
You can use Week 05 notes to calculate the BMW scores for each individual and aggregate over your sample. Alternatively, you can adapt my SPSS syntax to compute the BMW scores for your sample.