Description : For this final question set, you will be analyzing data from the SACMEQ assessment of student performance in Kenya. Data were collected from 6 ^th -grade students using a combination of stratified and cluster randomized sampling approaches. Listed below are a number of quantitative research questions for you to answer, using various statistical tests that we’ve covered during the semester. It will be up to you to determine which tests are best suited to answer each of the research questions. Provide your reasoning for choosing each statistical test. To answer each of the questions adequately, provide narrative responses, including correctly-formatted APA reporting of results, as well as a brief discussion as to the importance of the findings. Provide an adequate assessment of the required assumptions for statistical tests unless otherwise specified.

Research questions:

1. Is there a significant relationship between mother’s level of education (MOMED) and father’s level of education (DADED)? If so, describe the nature of the relationship.
2. Are girls more likely to be absent from school (ABSENT) if they have a male vs a female reading teacher (RSEX)?

1. Filter the dataset using "Select Cases" (using the SEX variable) so that only female students are included in the analysis
2. Carry out a statistical test to determine differences in absences (ABSENT) for students of male and female teachers
3. Remember to remove the female filter for subsequent analyses

1. Are there differences in reading class sizes (RCLSIZE) by school location (SLOCATION)? If there are significant differences, what is the nature of the differences, and where do they exist?
2. Is there a difference in student reading performance (READSCORE) between public schools and private schools (PRIVATE)? Note: Ignore the assumptions for the statistical tests in this question.

1. Run a statistical test to determine whether there is a performance difference between public and private schools. If so, how large is the difference in public and private school performance?
2. Does any public/private (PRIVATE) reading (READSCORE) difference change after controlling for student socioeconomic status (SES) and sex (SEX).

1. Describe the results and the nature of any change. How do you interpret these results?
2. In your view, what is causing any observed changes in these two models (i.e., why would you see a change after controlling for SES and SEX)?
3. Discuss the relative importance of all independent variables in predicting a student’s reading score.

1. What do you interpret about the effectiveness of these two models in explaining the outcome?

1. Within the dataset, there are 14 items that measure different household tasks that participants carry out (TASK1 – TASK14). Perform exploratory factor analysis and reliability analysis on these items to determine whether they measure any latent variables (i.e., constructs). Note: ignore the assumptions for this statistical test .
   • How many reliable factors are represented within these items?
   • Report the information or results of the factor analysis and reliability analysis that led you to your decisions.
   • For any reliable factors, look closely at their respective items (i.e., the wording of the actual questions/descriptions of tasks 1-14) and provide an explanation as to what construct(s) these items are measuring.
   • Create composite variables for each of these constructs, and describe the process you used for creating these composite variables.
2. Which variables in the dataset best predict student math performance (MATHSCORE)?

Note: Ignore the assumptions for the statistical tests in this question.

1. Looking at the dataset, which variables do you expect might have predictive power on student outcomes in math performance? Select a handful of variables (4 to 10) from the dataset, and assess their combined ability to predict math performance. Include any composite variables that were created in question 5. Do not include the private school variable (PRIVATE) in the model.
2. Be sure to check how each variable is coded and labelled. Keep in mind:

1. Dichotomous nominal variables included as independent variables in regression models should be coded 0/1.
2. Nominal variables with more than two categories cannot be included in regression models. If you want to include a nominal variable, it must be converted into dichotomous dummy variables.
3. You can treat any ordinal variables as continuous for the purposes of this question.

1. Does this set of variables (i.e., the model) offer any predictive power with respect to student math performance? If so, is it small, moderate, or large?
2. Which variables (two or three) are the most important predictors? Provide your interpretation of the results related to these predictors.
3. Describe one or two non-significant predictors. Are there any you were surprised by?
4. Provide (copy and paste) the following tables from your SPSS output: Model summary; ANOVA; Coefficients.
5. Provide your interpretation of these findings

1. Is there a difference in the provision of school-provided free meals (FREEMEAL) across geographic regions (SLOCATION)? If there are significant differences, what is the nature of the differences, and where do they exist (i.e., in which regions are schools more and less likely to provide free meals to students)

Price: $41.52

Solution: The downloadable solution consists of 17 pages, 2452 words and 1 charts.
Deliverable: Word Document