What probability model should I use to estimate the likelihood of achieving tertiary education?

[Ajay Joshi]

So, I am currently writing a dissertation for the effect of equal educational opportunities on the significance of the relationship between parental occupation and likelihood of achieving tertiary education.

My dependent variable is a binary variable that measures if individual has achieved a tertiary education level. I have two parental occupation variables, one for each parent and then two variables that measure the equality of education by asking individuals if they feel they can achieve the level of education they seek and the other asking if they believe any individual within that country can achieve the level of education they seek.

My sample size is around 800.

I was considering using a LPM model, but not sure if this is the best out of the three?

[Maarten Buis]

Remember, in statistics "significance" is not a synonym for "importance". So I suspect that you are interested in how the extend to which educational opportunities are equal influences the size of the effect of parental occupation on the probability of attaining tertiary education.

You only mentioned the Linear Probability Model (LPM). What other two are you considering? There are many more, as the number of ways of analyzing data is only bounded by your creativity.

Like any other model a LPM is a simplification of reality. That is after all the definition of a model. Simplifications will work in some situations and not in others. So that is the answer to your question: It may or may not work. Just carefully diagnose your model, and see if works. Look at alternative methods and see if the results are equivalent. The parameters often, but not always, have very different numerical values, but the substantive conclusion your draw from these models should be the same. When that is not the case, then understanding why will give you an indication of which model to prefer.

---

Sociology of education lecture starts here:

With that setup you are looking at the perception of equal educational opportunities. In general people from a privileged background perceive opportunities to be more equal as people from less privileged backgrounds . One explanation is that that way the privileged can fool themselves into thinking that they actually earned their own position (I am blanking out on the technical term for that mechanism), another explanation is that they never experienced the barriers other people face. It is probably a combination of the two. Regardless, what people say about the equality of opportunity and what it actually is can be wildly different, and the extend of this disconnect between perception and reality is dependent on the social background.

Equality of opportunity is a property of an educational system, not a property of an individual. So if your data comes from a single educational system (country or federal state at one point in time, depend on how education is organized in that country), then the equality of educational opportunities is a constant, and your question can not be answered with the data you have.

[Ajay Joshi]

Regarding your first point, yes, that is what I am actually looking at. My word choice was poor, so will change it so that it makes my research question clearer.

So far, I have only learnt about the LPM, logit and probit models. I believe a logit could be suitable due to the smaller sample size? My knowledge isn't that extensive regarding these models, if I'm completely honest.

Sorry again, I've used poor wording. My variables don't look at the opportunities, but instead they look at the individuals perception on equal educational outcomes. I understand that the equality of education is a property of the educational system and not the individual. But that doesn't mean their perception on equal educational outcomes won't have value?

[Maarten Buis]

No need to be sorry about imperfect wording. A lot of the effort in my intial write-ups of research projects goes into finding right phrases and words to exactly say what I want to say. My comments were only meant to help you speed up that (perfectly normal) part of the write-up process. If you found them helpful, then I would consider this a success, and something we should feel happy about and not sorry.

The association between parental occupation and education of the children is typically used to measure inequality of edcuational opportunity. So your research question is how does the perception of something influence that something. That gets very complicated, even more so while the perception itself is influenced by the background, and probably also by the achieved level of education... Disentangling all that, could be a somewhat ambitious 3 or 4 year PhD thesis. But only if you can find a convincing story why your intended audience should care. My intuition says that it may not be impossible for such a story to exist, but it is not going to be an easy story. Making sure that that story really exists and is solid is something I would definately do before spending lots of effort and time on solving technical issues. This topic is definately not something I would recommend for a BA or MA thesis.

[Ajay Joshi]

Hi Maarten,

You previously said "With that setup you are looking at the perception of equal educational opportunities. In general people from a privileged background perceive opportunities to be more equal as people from less privileged backgrounds . One explanation is that that way the privileged can fool themselves into thinking that they actually earned their own position (I am blanking out on the technical term for that mechanism), another explanation is that they never experienced the barriers other people face. It is probably a combination of the two. Regardless, what people say about the equality of opportunity and what it actually is can be wildly different, and the extend of this disconnect between perception and reality is dependent on the social background. "

Do you know of any papers that highlight the point of individuals from more privileged background are more likely perceive opportunities as more equal compared to people from less privileged backgrounds?

Thanks.

[Barc]

Remember, in statistics "significance" is not a synonym for "importance". So I suspect that you are interested in how the extend to which educational opportunities are equal influences the size of the effect of parental occupation on the probability of attaining tertiary education.

You only mentioned the Linear Probability Model (LPM). What other two are you considering? There are many more, as the number of ways of analyzing data is only bounded by your creativity.

Like any other model a LPM is a simplification of reality. That is after all the definition of a model. Simplifications will work in some situations and not in others. So that is the answer to your question: It may or may not work. Just carefully diagnose your model, and see if works. Look at alternative methods and see if the results are equivalent. The parameters often, but not always, have very different numerical values, but the substantive conclusion your draw from these models should be the same. When that is not the case, then understanding why will give you an indication of which model to prefer.

---

Sociology of education lecture starts here:

With that setup you are looking at the perception of equal educational opportunities. In general people from a privileged background perceive opportunities to be more equal as people from less privileged backgrounds . One explanation is that that way the privileged can fool themselves into thinking that they actually earned their own position (I am blanking out on the technical term for that mechanism), another explanation is that they never experienced the barriers other people face. It is probably a combination of the two. Regardless, what people say about the equality of opportunity and what it actually is can be wildly different, and the extend of this disconnect between perception and reality is dependent on the social background.
In addition, if you have difficulty writing a writing assignment that deals with a topic, including sociology, you can turn to https://writix.co.uk/write-my-assignment where you will find professional and very qualified help from cool writers.
Equality of opportunity is a property of an educational system, not a property of an individual. So if your data comes from a single educational system (country or federal state at one point in time, depend on how education is organized in that country), then the equality of educational opportunities is a constant, and your question can not be answered with the data you have.

Thanks a lot for the explanation.

[Barc]

Remember, in statistics "significance" is not a synonym for "importance". So I suspect that you are interested in how the extend to which educational opportunities are equal influences the size of the effect of parental occupation on the probability of attaining tertiary education.

You only mentioned the Linear Probability Model (LPM). What other two are you considering? There are many more, as the number of ways of analyzing data is only bounded by your creativity.

Like any other model a LPM is a simplification of reality. That is after all the definition of a model. Simplifications will work in some situations and not in others. So that is the answer to your question: It may or may not work. Just carefully diagnose your model, and see if works. Look at alternative methods and see if the results are equivalent. The parameters often, but not always, have very different numerical values, but the substantive conclusion your draw from these models should be the same. When that is not the case, then understanding why will give you an indication of which model to prefer.

---

Sociology of education lecture starts here:

With that setup you are looking at the perception of equal educational opportunities. In general people from a privileged background perceive opportunities to be more equal as people from less privileged backgrounds . One explanation is that that way the privileged can fool themselves into thinking that they actually earned their own position (I am blanking out on the technical term for that mechanism), another explanation is that they never experienced the barriers other people face. It is probably a combination of the two. Regardless, what people say about the equality of opportunity and what it actually is can be wildly different, and the extend of this disconnect between perception and reality is dependent on the social background.

Equality of opportunity is a property of an educational system, not a property of an individual. So if your data comes from a single educational system (country or federal state at one point in time, depend on how education is organized in that country), then the equality of educational opportunities is a constant, and your question can not be answered with the data you have.

I absolutely agree with you. Very deep thoughts that respond within me. In many ways, education plays a major role in this process, and the more quality education is received by the vulnerable segments of society, the better it is for society itself in the long run. I recently wrote my writing assignment about this and sought help from https://writix.co.uk/write-my-assignment who helped me with it. I came to the same conclusion you did - Equality of opportunity is a property of the educational system, not the individual.