Related Posts
Additional Posts in Excel Genius
Anyone know what these badges are?

New to Fishbowl?
Download the Fishbowl app to
unlock all discussions on Fishbowl.
unlock all discussions on Fishbowl.
Anyone know what these badges are?

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Download the Fishbowl app to unlock all discussions on Fishbowl.
Copy and paste embed code on your site

Scan your QR code to download
Fishbowl app on your mobile

Subject Expert
Ooohhh this is so good
So, first things first, you can't prove a negative in statistics (in life as well), so whenever you're trying to ascertain a relationship between two phenomena, you need to analyse it by trying to prove that there is NO RELATIONSHIP, and by failing to prove that, comes to the conclusion that there is indeed a relationship.
There are TONS of tools to do with, going from simple Pearson's correlation (the common, linear one) to multi-variable anslysis; and some of those (luckily) are built-in on excel (or can be built somewhat easily).
So, in short, if you want to BE ABLE TO RUN "BASIC" analysis, I'd recommend looking at the basics of those tools who are applied "in the industry" - A/B tests, basic correlation/covariance; and if you need, you can move to F or Z-statistic tests (ANOVA/MANOVA, z-tests).
If you really wanna learn about it (take to the next level, but reaaally the next level) you need to study inferential statistics - it's a wonderful tool, but there are soooo many ways to fuck it up (the clever Hans effect - the Horse who could perform arithmetic calculations - being the funniest)
Subject Expert
P.s.: sorry if it was too long, but there's so much to this topic
If you need statistics/data science then an excel forum isn't your best bet
Subject Expert
Depends on what you mean by large, and why would statistics/data science be applied only for 1 MM+ databases?
Also knowing excel doesn't mean you don't know the other tools
Coach
Look up A/B testing calculators, so many online for basic A/B tests with 1 metric (response rate for example). To your point, if they are inconclusive, even if a statistical test says that the impact is statistically significant, if you can’t easily observe an impact, I would question if it’s strategically significant and worth investing time and resources into.
Coach
Things like t tests can be easily built in excel so that might be the next step if you want to do something more complicated or more flexible.
What an amazing answer victor. Thank you so much! I will look into all of these things. I’ve never even heard of inferential stats but it sounds exactly what I was looking for
An analytical crowd, though!
Alteryx
When you know, you know