CONTEMPORARY RESEARCH MANAGEMENT II (METHODS AND MEASUREMENT)

CONTEMPORARY RESEARCH MANAGEMENT II (METHODS AND MEASUREMENT)

 

Paper details:

this is a type of report with questions to fill in 1. Define and explain the following phenomena, using examples when possible. a. Item-total correlations b. Coefficient alpha c. Item discriminability (or item discrimination) d. Endogeneity e. Standard error of measurement f. Factor loading g. Multi-trait multimethod matrix h. Adjusted R2 i. Metric invariance j. Correlated errors 2. Find a measure for the following variables and provide a rationale for your measure. Indicate how you would validate this measure or provide validation evidence (e.g., cited from previous sources) and how you would use it in a study. Note that this question is more open-ended than others. You are welcome to select what is widely considered the best measure of a given construct or discuss how you would create and/or validate a new measure. You are also welcome to select more than one measure and use them as indicators of a latent variable; realizing that a single measure may be flawed but that a composite of variables or a latent variable may compensate for flaws in individual measures. What I’m looking for here is that the chosen measure(s) are justified given the context (how you intend to use them) which might include also discussing the advantages and disadvantages. a. Core self-evaluations b. Firm-level performance 3. What are some advantages and disadvantages of archival data? How might you provide evidence for the validity of a measure derived from archival data? 4. What are the advantages and disadvantages of using survey data (self- and other-reports)? How might you provide evidence for the validity of a measure derived from survey data? 5. Assume that you are working on a study of a relatively new construct, willpower, which is defined as the ability to overcome internal and external obstacles in pursuit of a goal. Reviewer 1 on your paper has suggested that your construct is not unique from the construct of grit, which is defined as the “ability to work strenuously toward challenges and maintain effort and interest over time despite failure, adversity, and plateaus in progress”, and thus, has asked that you provide evidence that grit is unique. Reviewer 2 has asked that since your construct is relatively new that you conduct a cross-validation study of the construct. The editor has suggested that the paper will be accepted if you can address these two points. Please discuss how you would solve the issues that the Reviewers pose. 6. Compare and contrast the evidence necessary for content-related, construct-related, and criterion-related validity. 7. Name at least 2 underlying assumptions of Classical Test Theory 8. Describe the ladders of power and how it is used to transform a distribution of data. Be sure to note under which conditions it would be applied (Note: I am not asking you to list the specific name of the reexpressions). 9. How does reliability affect the statistical power of a test? 10. Write out the formulas for the following: a. Classical test theory b. Reliability 11. Methodologists have long criticized the proliferation of organizational theories and declared that too many of these theories are ambiguously specified, have been subjected to weak tests, and do not focus on important questions. Explain these critiques and develop recommendations for 1) individual researchers, and 2) academic institutional gatekeepers that would foster the development of strong theory and strong theoretical tests. 12. Assume that you surveyed 300 employees, asking them to report on the following variables: • job satisfaction (5 items), • leader-member exchange quality (LMX; 6 items) • organizational citizenship behavior (OCB; 8 items) • counterproductive work behavior (CWB; 8 items) • task performance (5 items) You test a model whereby job satisfaction mediates the relationships of leader-member exchange quality with OCB, CWB, and task performance. You test that model via path analysis with observed variables. There are paths from LMX to job satisfaction, and from job satisfaction to OCB, CWB, and task performance. There are no paths from LMX to OCB, CWB, and task performance. (a) After analyzing your model, you get the following fit statistics: chi-square = 280.25 (p < .05), CFI = .89, RMSEA = .11, AND SRMR = .08. Does this model fit the data well, and why? (b) What model modifications could you include to improve model fit that are theoretically justifiable? In addition to the modifications that you make, your software suggests that you add a path from LMX to task performance to further improve model fit. Discuss why or why not you would include this path. (c) When you submit this paper for review, a reviewer makes the following comment: “I noticed that job satisfaction and leader-member exchange quality are highly correlated. The correlations among your outcome variables are also somewhat high. I therefore question the discriminant validity of your constructs and would therefore like to see some empirical evidence of their distinctiveness.” Using the data that you have, what could you do to address this reviewer’s comment? Write a one-paragraph response to this reviewer outlining the steps that you took, including hypothetical results that would be supportive of the discriminate validity of your measurement model. (d) The same reviewer makes the following comment: “I also question the proposed causal order of your model. It seems equally plausible that people who engage in good behaviors will develop higher quality leader-member exchange relationships and be more satisfied with their jobs.” Again, using only the data you have, are there any analyses that you can do to assuage this reviewer’s concern? Write a one-paragraph response to this reviewer outlining the analyses you conducted with hypothetical results that would be supportive of your particular model. (e) The editor requests that you collect additional data to address the issue of causality. Describe the new study that you would conduct to provide empirical evidence for the internal validity of your model. Explain why this study meets the criteria needed to demonstrate causality.