Data Source

The Common Core State Standards (CCSS) were developed for Mathematics and English Language Arts with the goal to provide consistency in education in the US. They have been widely debated through many platforms since their introduction, including social media. We will explore the sentiment of tweets about CCSS between January and May of 2020.

Does public sentiment towards Math Common Core State Standards differ from public sentiment towards English Common Core State Standards based on tweets from January to May 2020?

Our data set includes tweets from Jan - May 2020 that include #ccss and “common core”.

To create two sets of data for comparison, the following words were filtered:

• Math - math, mathematics

• English - ela, english, eng, reading

To tidy the two data sets, the following was completed:

• Filtering for possibly sensitive language

• Tokenizing text

• Removal of stop words (including custom stop words such as “https” and numbers)

To measure sentiment, AFINN, Bing, and NRC were used for comparison

There were many more daily tweets about Math (11,148 total) compared to daily tweets English (1,548) from Jan-May 2020.

AFINN assigns values to words on a negative to positive scale that ranges from -5 to 5. By taking the sum of each data set, we get an idea of overall positive or negative sentiment. From this analysis, Math is overwhelmingly negative and English is positive.

## # A tibble: 2 × 3
##   lexicon subject sentiment
##   <chr>   <chr>       <dbl>
## 1 AFINN   English       296
## 2 AFINN   Math        -5481

Let’s look at a few more sentiment measures to see if these findings are consistent.

Bing categorizes words as positive or negative. To measure overall sentiment, I took the count of positive and negative words, and found the difference (pos - neg). Using this measure, both were negative but Math was much more negative than English.

## # A tibble: 2 × 4
##   subject negative positive sentiment
##   <chr>      <dbl>    <dbl>     <dbl>
## 1 English     1096     1069       -27
## 2 Math        6331     3609     -2722

Since Math has more tweets than English tweets, this shows percentages instead of counts. English is more balanced, while Math has a larger percentage of negative words.

NRC expands categories to include emotions. I was curious if any other emotions would be revealed. Surprisingly, the highest count and percentage for both data sets is positive.

• Strengths: Using multiple lexicons gave a broader picture of sentiment for Math vs English Common Core Standards and revealed inconsistencies that would be worth investigating further.

• Limitations: The data sets may need further investigation based on the words used for filtering. For example, some tweets included references to both Math and English, and could be examined for accuracy. There might be other words that could generate more accurate data sets.

• Audience Considerations: This information could be used as a “lessons learned” to consider how Math vs English educators were prepared through training and resources to implement these standards, and how the implementation was communicated to parents and the larger community.

• Future Analysis: NRC categorized both Math and Eng as being largely positive, potentially due to the inability to identify sarcasm or negations. It would be worth it to investigate NRC further to consider how words are classified.