• http://www.medium.com/@patrickbaek Patrick Baek

    #mean vs #median

  • http://www.capitalcitystats.com/ George Richardson

    sampling method. Prim is sampling prospective clients and only receiving voluntary responses from interested clients. The other two samples involve customers showing the initiative to visit a website first, so they expressing interest already to some degree, and from those prospects you should have a higher conversion. Not a bad attempt at explaining sampling distribution in your article, especially from a non-statistician

  • http://conversionscientist.com Brian Massey

    The audio can be found here: http://conversci.com/statsig

  • http://conversionscientist.com Brian Massey

    Is this a correction? A question?

  • http://conversionscientist.com Brian Massey

    While your observation is correct, it is not what I am looking for. Hint: Look at the function choices for Standard Deviation in Excel.

  • http://www.capitalcitystats.com/ George Richardson

    Excel is gross, SAS is King. The T-distribution and sampling distribution are all based on sample size, reference William Gosset and Central Limit Theorem, and as sample size increases the standard error decreases…its in the formula

  • http://www.capitalcitystats.com/ George Richardson

    but the major issue still remains the “sampling” method Prim used. Sample size, even in the 100′s of thousands makes no difference if the samples are not properly collected. That was the glaring mistake I saw before getting carried away with asymptotics

  • http://conversionscientist.com Brian Massey

    But is Prim’s “sample” really a sample?

  • http://www.captivetouch.com/ Sherry Nouraini

    I love this article! As you mentioned in the comments, Prim’s sample is not a sample, it is the whole population.

  • http://conversionscientist.com Brian Massey

    Sherry, you are correct! Thanks for the kind words. Look for the second part on the 21st.

  • http://conversionscientist.com Brian Massey