Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

SOST70042 Structural Equation and Latent Variable Modelling, 2022-23

Assessment 3: Sex differences in maths self-efficacy

(50% of course grade)

You will be using data from the teleprism study (http://teleprism.com), which surveyed school pupils and their families about maths education.

The data are from pupils in classes in schools. You will use binary items maths6a, maths7a, maths11a, maths14a, and maths15a as observed indicators of latent maths self-efficacy (1 = Agree   with item, 0 = unsure or disagree with item.) See the end of this document for a full codebook of the variables.

Use the file “teleprism_sem_for_assessment2023.R” to start your analysis. This file contains syntax for some basic data cleaning and variable creation.

Use the data to address the research questions below, using appropriate models from the lavaan library in R. Write a short (max 1,500 words) report that answers the research questions.

Your report must include an appendix at the end that contains the R code you used to edit the data (if necessary) and fit your models. This appendix will not count towards your word count. There is a 10-mark penalty for not including an R appendix.

Research Questions

1.    How good are the measurement properties of the five-item scale of latent maths self- efficacy? (discuss descriptive statistics and model parameters) [20 marks]

2.    Is latent maths self-efficacy different in pupils of different sex?  [30 marks]

3.    Discuss whether your results to 2. above may have been confounded by maths ability (math_ab) per class? If needed, present additional model results to address this confounding [50 marks]

Guidance

0.    Use SINGLE- LEVEL IRT MODELS to address ALL questions above. (Lavaan cannot fit multilevel IRT models.)

1.    Organise your report as you see fit to best communicate to me how your data and models answer the research questions, but you must include an R appendix (see above) and a        References section if you cite literature in your report.

2.    1,500 words is the limit for your report. Your report may contain fewer words, but not more.

3.   You may present up to four illustrations in your report (illustrations are Tables and Figures, e.g., you could use two tables and two figures, or four tables but no figures, and so on.)

4.   The Tables, Figures, Reference section, and R appendix do not count toward the word count.

5.   The number of models you fit does not have to equal the number of research questions above. You may be able to address more than one research question with one model or need more than one model for another question.

6.   You may amend the dataset as you see fit, but you must ensure that any data cleaning or data manipulation is clearly described in your code and justified.

7.   Your grade will reflect how well you have addressed the research questions.

8.   A good answer will include consideration of the important assumptions made in the modelling/analysis, and whether these assumptions have been met or violated.

Codebook

Variable name Variable label / content Response