Informative and great alternatives to the mean, but I would have liked to see the median highlighted to remain in the realm of simplicity. Sure, the median appears in the box plot but the entire plot is still too complex for high level summaries. Any reason you didn’t explicitly mention the median as an easy (one statistic summary) alternative to the arithmetic mean or did I miss it? Again, looking at full data should be the default, but given people will typically end up summarizing info with one or two numbers, it will be good to advocate the median if one has to settle for one statistic.
Median is good as a one single and simple alternative, but it still lacks the sufficiency. One other mean is preferred is that business folks can easily multiple the mean by the sample size or population size to get the total which is money in the bank (for example spend per customer is $50 so if I can drive 1000 customers, I would get $50,000 in sales). The median fails at that. So this discussion is focused on going beyond the one metric which businesses use (mean) and encouraging them to understand data distribution and spread, and the impact of outliers. But ultimately, the median provides a better alternative if the goal is to have just a one summary measure in some situations...like median income for example.
Informative and great alternatives to the mean, but I would have liked to see the median highlighted to remain in the realm of simplicity.
Sure, the median appears in the box plot but the entire plot is still too complex for high level summaries. Any reason you didn’t explicitly mention the median as an easy (one statistic summary) alternative to the arithmetic mean or did I miss it? Again, looking at full data should be the default, but given people will typically end up summarizing info with one or two numbers, it will be good to advocate the median if one has to settle for one statistic.
Median is good as a one single and simple alternative, but it still lacks the sufficiency. One other mean is preferred is that business folks can easily multiple the mean by the sample size or population size to get the total which is money in the bank (for example spend per customer is $50 so if I can drive 1000 customers, I would get $50,000 in sales). The median fails at that. So this discussion is focused on going beyond the one metric which businesses use (mean) and encouraging them to understand data distribution and spread, and the impact of outliers.
But ultimately, the median provides a better alternative if the goal is to have just a one summary measure in some situations...like median income for example.