Comments on: You Do Not Really Know Statistical Significance, Do You?

By: Shock Marketer

Shock Marketer — Tue, 17 Aug 2010 19:56:59 +0000

I don’t know of any books on stats that are explicitly for marketers. However, I’ll send you an email in a few days.

By: Slave Rat

Slave Rat — Tue, 17 Aug 2010 15:33:06 +0000

Yes, that’s where I was headed.

I’m trying for the love of me to remember all the Chi-Square tests for statistical significance and the calculations for sample size – It was pretty darn hard on me when I did this for my dissertation, and I had to do it some two dozen times with US Census data against my own survey data. It wasn’t that long ago – a little over two years, since I barely understood how I was doing what I was doing I’ve forgotten a lot of it. Would you have any good book to recommend? All I have here to help me is what I used back then, “SPSS survival manual” (I do still have SPSS 16), a marketing research book (McDaniel and Gates) and my “Dissertation Skills” (Brian White) book.

None of them address the Internet Marketer’s issues head-on… I KNOW half the battle is spending less on what doesn’t work and quickly figuring out what does work – For that I’m thinking being able to figure out sample sizes and statistical significance are essential.

Any good reads?

By: Shock Marketer

Shock Marketer — Mon, 16 Aug 2010 20:53:32 +0000

@Slave Rat: If you mean by alternating each visit to A, A’, then B (instead of A gets first 1,000, A’ gets second 1,000, then B gets next set), yes, that’s what you’d do — until you reached percent significance you were anticipating.

By: Slave Rat

Slave Rat — Mon, 16 Aug 2010 20:10:50 +0000

Couldn’t we accomplish the same as an A/A test by simply splitting the results of “A” into half (alternating their placement to A or A’ instead of grabbing the first 1000 and sending them to A and the last 1000 and assigning them to A’)?

By: Ed

Ed — Thu, 29 Jul 2010 12:31:52 +0000

You used to ‘run’ fairies there Tom?

I smiled when I saw the Hathaway man. Great article, thanks!

By: Tom

Tom — Wed, 30 Jun 2010 17:26:01 +0000

And the guy who was the Hathaway Shirt man was John Whyte. He ended up running the most popular gay bar on Fire Island and invented the Tea Dance.

I used to run ferries there. Not that there is anything wrong with that.

By: ewoww

ewoww — Wed, 09 Jun 2010 08:02:07 +0000

I’m so sad I’m late tot his posting partay… David Ogilvy = my hero, direct response was his dirty secret, and and that faux eyepatch propelled Hathaway into advertising greatness, not to mention a compelling and lasting brand. While admittedly confidence intervals weren’t my favorite part of school, it’s fantabulous to see someone putting it to relevant use. Great post – thanks for writing!

By: Justin Dupre

Justin Dupre — Wed, 02 Jun 2010 13:55:16 +0000

Very informative! By being able to determine of what you are doing right or wrong on the project you’re working on is huge. I’ve also done some split testing as well.. great results! Good post.

By: Shock Marketer

Shock Marketer — Tue, 04 May 2010 22:09:47 +0000

@Ben: Yes!

By: Ben

Ben — Tue, 04 May 2010 20:53:01 +0000

i believe the man who created that ad is david ogilvy