Dataset Source: FLEMMS 2019

The Philippine Statistics Authority (PSA) organizes a survey known as the Functional Literacy, Education and Mass Media Survey (FLEMMS) as a regular check on the state of the country's education systems. FLEMMS collects general information about each respondent's educational background, demographic information, and household information, followed by questions to ask and assess their ability to read, write, and answer questions. These assessments are then interpreted to classify each respondent's educational skills through a metric known as the Functional Literacy [Level]. As of April 2024, the most updated results available for the FLEMMS survey are those from the FLEMMS 2019 survey, whose report may be found through this link .

Data Collection

The results of the FLEMMS 2019 survey are divided into several CSV (Comma Separated Value) files tabulating the information gathered from each respondent. Among the files, this study intends to investigate the data from the responses in the Household Name CSV file. The Household Name CSV contains the respondents' answers to the Individual Questionnaire, which contains questions that directly assess their educational background and overall literacy. The relevant files for the FLEMMS 2019 dataset can be downloaded from the Philippine Statistics Authority Data Archive (PSADA) website .

Data Preprocessing

Dataset Link

Here we have three box plots which plots the participant's highest educational attainment against their level of participation. The left-most, middle, and right-most plots correspond to the participants' participation in the Arts, Sciences, and Business respectively.

The specifics of each level of participation and educational attainment is discussed below.

The "boxes" in the plot corresponds to the frequency of answers for each category. This means that each column's width is only respective to itself and not globally related. It's important to note that most respondents are under the junior high school and bachelor categories, with the other categories having significantly less respondents compared to these two.

For the arts, we can see a general upward trend between higher educational attainment and level of participation. This suggests that education is seen to be closely related to the level of one's participation in the field of arts. This suggest that to feel more involved in art, a higher educational attainment would be better.

For sciences, there's also a general upward trend, similar to the arts, where a higher level of educational attainment allows one to feel more involved. It is interesting to note that those who only received an early childhood education has responses evenly spread out from 1–4. This either suggest that the participant is not aware of the full scale of their participation, or that more education does not necessarily equate to more participation

Finally, for business, it's interesting to see that after finishing highschool and your bachelor's degree, there is a general downward trend in level of participation. This may suggest that further education after your bachelor's degree will actually decrease your level of participation.

Here we have three violin plots which plots the participant's level of participation against their percieved usefulness of knowledge. The left-most, middle, and right-most plots correspond to the participants' participation in the Arts, Sciences, and Business respectively.

The specifics of each level of participation is discussed below. Meanwhile, 1 corresponds to yes (Knowledge IS useful) and 2 corresponds to no (Knowledge is NOT useful). There are no values in between despite the visuals suggesting so.

Similar to the box plot, each "swell" in the plot corresponds to the frequency of answers for each category. This means that each column's width is only respective to itself and not globally related.

For all three categories—the arts, sciences, and business—we can see a general larger swell in the perceived usefulness of knowledge with a score of 1 as level of participation increases. At a level of participation score of 1, we can see a roughly equal score of perceived usefulness of knowledge, with this knowledge generally becoming more useful with higher levels of participation.

For the arts, specifically, we see the highest frequency with a level of participation score of 4.

For science, we see the highest frequency with a level participation score of 4, with 3 not differing by much.

For business, we see the highest frequency with a level participation score of 5. It is also interesting to note that those who perceive their knowledge as not useful is thicker compared to the arts and sciences for each level. This may suggest that knowledge is perceived as not as useful for business as compared to art and science.

We primarily focused on addressing the first objective and hypothesis, which was achieved through statistical testing. We have achieved a p-value lower than 0.005 for all three fields—the arts, sciences, and business. This means that the results are all statistically significant for all three fields. This also means that we reject the null hypothesis. As such, the level of participation of Filipinos is correlated to their highest level of educational attainment across all three fields of expertise.

Moving forward with this project, data used to conduct this study is actually quite limited and skewed across different categories of highest educational attainment. More than half of the participants' highest educational attainment is junior highschool with bachelor or eqv. taking a majority of what remains. Although unlikely, more data across different aspects of education would be best to better understand the distribution of responses.

Furthermore, machine learning can also be utilized in order to interpret the data as well, through data classification. Each objective can be studied with a neural network that attempts to predict the dependent variable based on set feature variables. For instance, the first model can attempt to determine an individual's most probable level of participation in each of Arts, Science, and Business using a classification model, which can be implemented using libraries such as TensorFlow and Keras. These level of participation may be predicted by the model based on the individuals':

• highest level of educational attainment
• other demographic information
• personal values
• experiences with sports
• exposure to communication media
• etc.