PSYCHOL7332: Two hundred fifty students complete an 8-item survey of Attitudes Toward Statistics: Psychological Testing Assignment, UoA, Singapore

Two hundred fifty students complete an 8-item survey of Attitudes toward Statistics. The survey is measured on an a5-point scale. Five of them measure the value of learning statistics and three of them measure the difficulty of learning statistics.

Items

• V1  Statistics should be a required part of my professional training.
• V2  Statistical skills will make me more employable.
• V3  I use statistics in my everyday life.
• V4  Statistics is irrelevant in my life. ^R
• V5  I will have no application for statistics in my profession. ^R

1. D1  Statistics is a subject quickly learned by most people.
2. D2  Statistics is a complicated subject. ^R
3. D3  Learning statistics requires a great deal of discipline. ^R

Part I

• Is there enough correlation to conduct factor analysis? Why or why not? Provide relevant statistics to support your answer.
• A one common factor model using maximum likelihood estimation was fitted. Comment on the goodness of fit of this model. Provide relevant statistics to support your answer. Consider factor loadings >0.3 as salient
• Now, an EFA model with two common factors using maximum likelihood estimation was fitted. Basic on the given output, comment on the goodness of fit of this model. Provide relevant statistics to support your answer. Consider factor loadings >0.3 as salient (non-ignorable). Use Promax rotation if you need to. How will you label the two factors? Are there any items that should be removed/revised based on this analysis?
• Based on results from (c), a CFA model with two independent clusters based on test content was fitted. Comment on the goodness of fit of this model. Provide relevant statistics to support your answer.
• Compare the results from (b) and (d), which model you think best represents the data
• Based on the model fitted in (d), compute the coefficient omega for each cluster of items and compute the correlation between the two cluster sums. Why is this correlation smaller than the correlation between the two factors?
• Based on the model fitted in (d), find the validity coefficients for each cluster sum.
• Find the correlation (1) between the first cluster sum and the second factor; (2) between the second cluster sum and the first factor.
• Based on results from (g) and (h), arrange the correlations in a convergent-discriminant validity matrix. What can you conclude about the convergent and discriminant validity of the present scale?