Liars, damned liars, and statisticians: Why peer review is good and necessary

This was meant to be posted in the comments of Ko's recent post, "The Misleading Data Found in "The Split of the June 2026 Presidential Election", but I spoke for too long and hit the character limit. Ko essentially did a volunteer peer review of the article I posted regarding the stats of #campaign, critiquing my methodology and analysis, and it should be required reading to read his post if you read my post. More critique!!! Call me out for lying to you!!!!! Anyways, here's the comment:

Thank you so much for reading my article and responding, you bring up some very good points!!! I absolutely do need to be more careful when I am presenting my data. I'd like to note that I'm coming from more of a place of "hmmm there's something interesting, let me make some predictions and see if my dataset supports them", but I didn't make that very clear in the article. You're thinking like a journalist and you're rightfully calling out how the data I'm presenting doesn't show the whole story, and I really should be putting more disclaimers on my stuff. Statistics are lies made up by big propaganda to give more credence to an idea than it deserves.

I honestly was not thinking too hard about the impact of this article and just really wanted to get some interesting numbers out there for other people to chew on, but I also like hearing myself talk, so I talked a bit about the numbers and what I thought they meant. This was not meant to be an epic investigative journalism win, but instead a total nerd infodump. This is why I cannot be a journalist and must instead remain an unreliable narrator niche Gnomestack blogger.

Re: "This shows that all presidential candidates actually underutilized the new post-Jaron players.", I absolutely agree. Thank you for grabbing all of that data, it just wasn't something I wanted to focus on from the data set I did grab :) I do still think there's value in comparing the candidates, though. Both "other candidates" and "the expected number from the population statistics" are really valuable reference points, and can give us different perspectives on a situation :) "All presidential candidates underperformed with new players compared to the population statistics" and "These two presidential candidates had over 50% of their campaigning base come from new players, while these three had 0%" are different ideas that can coexist.

Re: "This presumes that it is possible to have a highly established new player, which I don't believe is realistic with how this is being measured." I agree with you, but I needed some kind of number to investigate the question being asked ("are these players being exploited?"), and with the dataset I was collecting & the method of collection (scrolling thru discord and clicking on each individual profile), "connectedness to juniperfig" was a much simpler set of data to collect than "actual quantitative data on how established you are". Even if something is obvious, like "new players have shaky connections", it's important to show evidence supporting that "obvious" thing that before speaking about it, else you're just pulling stuff out of your ass. I'd be interested in your ideas for a quantifiable measure of "political connection" that is relatively simple to collect and won't break any laws LOL

Re: "All she did was take the average spent on other campaigns and multiply that by the number of messages for multiman.", I totally understand how you got that assumption, but it's incorrect. I just didn't explain my data at all, so I will explain it now!

Since I didn't get a figure from Multiman155, I went off the publicly available data I had: His campaign was offering $150 per ad sent in #campaign. I decided to run with this, and say his cost for discord campaigning (and therefore everyone else's cost for discord campaigning, to keep things standardized) was $150 per ad.

"Maximum commission" here is "How much did the highest paid campaigner for this candidate make?". If the highest paid campaigner sent 10 ads, the "maximum commission" is $1500 ($150/ad * 10 ads).

"Average commission" here is the average total compensation of ALL campaigners at $150 an ad. If the average campaigner sent 3 ads, the "average commission" is $450.

"Total commission" is how much the campaign would have spent if they paid for every single one of their #campaign ads, so it's "total # of campaign ads * $150/ad".

Re: "While it is interesting to see that a potentially high amount was spent, it would actually be more interesting if the opposite were true.". I would argue (and I did!) that $30k was a low price tag for 227 ads in #campaign. I vehemently disagree with this line of criticism as well. "It would actually be more interesting if the opposite were true" is not a perspective I can take. I'm interested in irresponsibly fucking around with data and then SHOWING EVERYONE EVERYTHING I DID. I am NOT interested in cherry picking data points to make my stories more interesting. If my data disproves my assumptions, and makes my articles "less interesting", good! I don't need entertainment, I need to infodump about my shitty google sheets on my silly little blog.

In conclusion, you think me an investigative journalist, when I'm actually just a nerd who makes graphs for fun and then speculates on them. Thank you for calling me out on my bad data presentation practices, never let people get away with showing you a graph.