There were several factors and ways that could have made the index more representative of each senator.

1- Senators without key issues

18/40 senators did not have a key issues page. To estimate these senators' scores, the score of each senator after calculating the amount of legislation introduced, voting records, and public statements were added up, weighted, and ordered.  The scores of the senators who had a key issues score were then added. The mean (out of one) of the three scores of senators with a key issues score directly above a senator with no key issues and the three scores directly below a senator with no key issues was calculated and added as the key issues score to the senator with no key issues score. This number was then multiplied by 15 in the final calculation. Additionally, senators with no key issues score towards very top of the order, before adding all the key issue score estimates, did not have three senators scores above theirs so the three below and as many as were above (whether 0, 1, or 2 scores) were compiled into the mean. The opposite was done for senators at the very bottom. The mean of the three scores from above and all the ones below (whether 0, 1, or 2 scores) were compiled to calculate the means. Some senators who did not have a key issues score also did not have scores directly above them in the order, so the three closest that were above were selected and contributed to the mean. The calculation of scores started at the bottom of the ranking so there was not the problem of not having senators directly below a senator with no key issues score that did not have a key issues score themselves. This calculation made many of the scores of the senators an estimate as opposed to based on data.  A future way to make the study more empirical is to broaden the scope of where key issue information is gathered. This study only considered key issues as what is directly stated on a senator's campaign website. Reaching out to each state senator's (with no key issues on their website) team individually about key issues of that senator would a way to decrease the number of senators with no key issues score.

2- Skew towards introducing legislation

Because of the nature of the method, the senators who introduced a high amount of legislation received significantly higher scores than those who did not. This was justified by the belief that senators who introduce legislation show initiative regarding environmental action and should thus receive higher scores.  Whether this is the most rational methodology is debatable but is certainly something that should be noted when evaluating scores. 

3- Every "green" aspect is weighted the same

Each "green bill," "green public statement," or "green issue" was not adjusted for how impactful they were.  No bill, key issue, or public statement was weighted more significantly than the others. This was to eliminate personal bias in determining the significance of any bill.  However, this is not the most effective in determining how much environmental initiative a senator shows. For example, a senator who introduces the Virginia Green New Deal should receive a higher score for introducing that legislation than someone who introduces legislation that promotes solar energy in a single locality because the Virginia Green New Deal is a piece of legislation with a far greater impact.  This also applies to voting records, public statements, and key issues. A future study that weighs the significance of each bill and assigns scores accordingly would contribute to a more effective analysis and more representative index score. 

4- Harmful legislation

No senator's score was reduced for supporting policies that did harm to the environment or were counterproductive towards environmental initiatives. Another beneficial study could consist of evaluating senators that propose environmentally harmful legislation, voting "nay" on "green bills," or having environmentally counterproductive public statements or key issues as decreasing a senator's score as opposed to leaving their score unimpacted. 

5- Bias

Opinion and bias are very present in this index. What constitutes a "green bill" or a "green public statement" is often left to subjectivity. As emphasized prior, even the development of the method had very present bias such as possibly favoring senators whose campaigns have more money as they are more likely to introduce more legislation. Many other factors contained personal bias and opinions that decrease the empiricism of the index.

6- Human error

Human error is most likely present in this analysis. There were thousands of bills, votes, public statements, and key issues that were analyzed. It is most reasonable to assume that given the mass volume of data analyzed, there is present some amount of human error such as typing "nay" when in fact a senator voted "yay" or skipping over a "green bill" by mistake.  A check of the data collection by a reputable third-party source would prove beneficial in decreasing the amount of human error.