Barriers to access to education

[ernestguevarra]

No description provided.

[claudiavidalc]

Dependent variables (Outcome): defined by access basic education, access education, ever attended school, no schools, mixed education - Pending to define how these variables were measured to identify assumptions and choose the most suitable to represent access to education

Independent variables:

1. Individual factors (access to education by age, sex)
2. Health services (access to education by , vaccine record, IYCF Counselling, ill disability, nutritional status, presence or absence of diseases [diarrea and fever])
3. Sociocultural factors (early marriage, displacement, access to preschool)
4. Structural factors (state, school fees, child working, other reasons?, no WASH, health insurance)

Potential analyses:

1. Descriptive: Frequencies 2x2 tables, chi square
2. Analytic: Multiple logistic regression

Next steps:

1. Understanding definitions of each variable (how they were obtained and what they capture), choosing the main outcome
2. Cleaning data
3. Descriptive and analytic analyses
4. RMarkdown report

[ernestguevarra]

Dependent variables (Outcome): defined by access basic education, access education, ever attended school, no schools, mixed education - Pending to define how these variables were measured to identify assumptions and choose the most suitable to represent access to education

Independent variables:

1. Individual factors (access to education by age, sex)
2. Health services (access to education by , vaccine record, IYCF Counselling, ill disability, nutritional status, presence or absence of diseases [diarrea and fever])
3. Sociocultural factors (early marriage, displacement, access to preschool)
4. Structural factors (state, school fees, child working, other reasons?, no WASH, health insurance)

Potential analyses:

1. Descriptive: Frequencies 2x2 tables, chi square
2. Analytic: Multiple logistic regression

Next steps:

1. Understanding definitions of each variable (how they were obtained and what they capture), choosing the main outcome
2. Cleaning data
3. Descriptive and analytic analyses
4. RMarkdown report

This looks like a really good plan Claudia! Well done!

I think this will be a good guide for the team when you start coding in our class on Monday! Looking forward to it!

[ernestguevarra]

Team @OxfordIHTM/rock-lee, just checking in to see how you guys are doing?

I have seen some code from you guys since two session ago but have not seen any commits and/or pull requests with new material.

How are things going? Need any help? How far have you gone with the steps above that you said you will be doing?

[ernestguevarra]

Team @OxfordIHTM/rock-lee, I saw code contributed to the rock-lee branch by @kap96 (a week ago) and @Rukomeza yesterday.

I think I am understanding where you are trying to do but not clear where you are headed. Start showing in code where you are headed by getting towards output/s that organise and present the point/result that you are getting and the answers that you want to share to your audience (which in this case is the Federal Ministry of Health in Sudan).

Another thing to note as well is to not forget to answer the original question which is "What are the barriers to pre-school education in Sudan?".

Finally, in as much as it is nice to see fancy analysis and models to show correlation and association, don't forget the "basic" and "simple" stuff as well - the data already has information from the mothers of reasons why their children are not in school. A "simple" or "basic" presentation/visualisation of these responses I think is very powerful beyond the model or the correlations mainly because it comes from the mothers themselves. Using that data acknowledges their responses and gives importance to this kind of information.

[ernestguevarra]

Team @OxfordIHTM/rock-lee, great to see you guys at work today.

I know you were able to get to a point where you have code to show results in plots etc. But I haven't seen a pull request yet. No problem at all and if you still want to work on this then I will wait until you PR. But if you want to move on and sort of keep this as it is now, let me know as well and let me know which branch has all the latest things you have done and I will take it from there.

If you want to see how the output report is currently looking right now, here is the current draft - https://oxford-ihtm.io/ihtm-hackathon-2024. Once you have a PR, then I can check that and get your stuff ready for merge to main and once merged, your stuff will get included into the report.

You guys have done well! Thank you for all the effort and time you have spent on this and on the module!

[babalao413]

Hie Ernest I hope this email finds you well. We committed and pushed our worked yesterday soon after class. I came and requested for your help with that when I hadn’t saved the work and you helped me right at the end of the session. I am not sure what may have gone wrong. All the codes are there on my RStudio. Would you be able to help on what may have went wrong and how to rectify it? Regards, Phill(on behalf of RockLee)

…

On Tue, 5 Mar 2024 at 22:22, Ernest Guevarra ***@***.***> wrote: Team @OxfordIHTM/rock-lee <https://github.com/orgs/OxfordIHTM/teams/rock-lee>, great to see you guys at work today. I know you were able to get to a point where you have code to show results in plots etc. But I haven't seen a pull request yet. No problem at all and if you still want to work on this then I will wait until you PR. But if you want to move on and sort of keep this as it is now, let me know as well and let me know which branch has all the latest things you have done and I will take it from there. If you want to see how the output report is currently looking right now, here is the current draft - https://oxford-ihtm.io/ihtm-hackathon-2024. Once you have a PR, then I can check that and get your stuff ready for merge to main and once merged, your stuff will get included into the report. You guys have done well! Thank you for all the effort and time you have spent on this and on the module! — Reply to this email directly, view it on GitHub <#6 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BC3JH45TWZ52KFOCPJNECUTYWZAT5AVCNFSM6AAAAABCZPXQRGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNZZG4ZTOMRSGU> . You are receiving this because you are on a team that was mentioned.Message ID: ***@***.***>

[ernestguevarra]

Thanks for letting me know. I agree, you committed and pushed. But I don't think you made a pull request.

No problem. I can find it from this end. I will get it onto the report now.

Well done!

[ernestguevarra]

Team @OxfordIHTM/rock-lee, thanks for the clarification. I have not processed your inputs from the rock-lee branch and did the pull request myself.

Some quick feedback:

• Good initiative to look for packages to help you with the task at hand. I think to some extent it helped you but at the same time it made you focus too much on the "presentation" of the results rather than figuring out what the actual results mean or whether the results make sense. As you can see in the updated report with your inputs included, the tables look nice (took 3 packages to get you to that table so it should be nice) but even with all that effort, It's hard to understand what the No and Yes are on the column panel as it is not labelled. I can see what those are because I can see your code, but the general audience will not. Another observation here is that it didn't seem to have become a realisation for you guys that with a survey with a sample of more than 139000 children aged 6 to 59 months old, your summary tables show so much of the so-called "missing values". It is important that you have at least flagged this and asked yourself, why is there so many NAs and that you either digged deeper or come to ask me why that is the case.

Remember that your topic is about education and that dataset covers children aged 6-59 months old. In Sudan, pre-school education is from 4 - 6 years old and basic education starts formally at 6 years old. If you paused to ask yourselves, you may have made this connection but just because of this, any child in the sample who is less than 4 years old will have a response of NA not because the data is missing or that they didn't respond to the question but because the question does not apply to them. So, the appropriate approach her would have been to filter out children less than 4 years (47 months and younger). I think your analysis would have been more meaningful with this in mind.

I can understand that the task was mainly a coding task and you focused on that. And I give you all full merit for this. I do hope that despite that, you don't lose focus on the question that is being asked.

• I think you did fantastic job with all the cross-tabulations and correlations you were want to test out. Some of those outputs are not on the report but You have them on your script. I do know that you were getting everything together in our last session and you were in a rush so understandable. But I really commend all of you for the good approach and thinking that went through to finding those associations and relationships.

• with more time, it would have been nice to have been able to get your analysis broken down by State. Sudan is a big country and correlations that you show for the whole country data is interesting but policies and programme implementation happens at the state level so any insight to what is happening at the state level means a lot more operationally than a highly aggregated country level result.

Well done everyone!