So I was recently reading my textbook for my auditing class when I came across the section on statistical and nonstatistical sampling. I became extremely distracted and started reading everything I could from a basketball perspective. I found a lot of parallels. And I'm not trying to start Stat Wars V: Revenge of the WARP, I just thought it was an interesting correlation and wanted to share. Argue if you really want, but just know that this isn't my objective.
Because I will be copying most of the work on auditing almost word for word, I better cite: Rittenberg, Johnstone, and Gramling's Auditing: A Business Risk Approach
Nonstatistical Sampling in auditing:
- Evaluation is based off of the auditor's judgment (I should note that there is nothing wrong with this, according to the textbook or myself)
- Requires audit judgment to determine an appropriate sample size and evaluate the results
- Does not provide an objective way to control and measure sampling risk
- Does not require additional software
- Can be based on auditor's prior expectations about errors in the account
- May take less time to plan, select, and evaluate the sample
Nonstatistical Sampling from auditing to basketball (A lot are pretty obvious, skip them if you want to save some time):
- Evaluation is based off of the viewer's judgment
- Requires the viewer to determine if they are seeing a play that is representative of the player and to evaluate the results accordingly
- It's nearly impossible to look at every aspect of the game completely objectively
- Goes back to the whole "don't take the fun out of the game" argument. Without stats, you really don't need to go online to look at all of the numbers just to prove what you already saw
- In auditing, this is really just a good way to add extra skepticism into the process. In basketball, this can be good or bad. Imagine that you know a player has been historically bad at a specific thing, but you saw them do something well. The good is that you can say to yourself, "Hey, that could very well have been a fluke. Or! Or he could be improving, so I'll withhold judgment until further proof." The bad is that you could say, "Hey would you look at that! He fixed his problems." Doesn't happen very often, though (McHale suffered from it...), as I know some eyes-guys are thinking right now. But the worse? Thinking to yourself, "No, I've seen three seasons of him being bad at that. That was a fluke." And being wrong. Probably happens a little more often. The worst? Prior expectations with no base in reality. As much as I know everyone wants to claim they don't do this, everyone does it. Even the stats guys such as myself. It's why Anthony Randolph continues to be so lucrative. Why Darko will always be a bust even if he is average.
- It takes a heckuva lot longer to input the data and calculate PER than it does to say, "Hey that Lebron guy, he's pretty good." On the other hand, watching a guy shoot well from behind the arc takes a lot of time to watch games, whereas looking up his shooting percentage on basketball-reference takes 20 seconds. Obviously someone still had to enter the data, but it saves everyone who wants to know from watching and finding out.
Statistical Sampling in auditing:
- Requires extra knowledge of methods or a special computer system
- Requires definitions of acceptable risk and objectives.
- Evaluates the results by providing an objective measure of sampling risk
- Gain efficiencies through computerized selection and statistical evaluation
Statistical Sampling from auditing to basketball:
- Requires the user to know what the numbers mean and how to use them
- There must be set values. What goes into the formula and what does it mean when it comes out?
- Basically that the number is objective. Doesn't always mean it's correct, obviously.
- Once the work is done, it's pretty easy to look at a guy's stats and determine what kind of player he is. i.e. High or low volume/usage, scorer, rebounder, passer, defender, etc.
Now it should be noted that this entire textbook emphasizes the need for an auditor to have knowledge of the company and industry they are auditing. In basketball, that's the equivalent of saying that only seasoned scouts should be using their eyes. Obviously that's a little silly. I'm not trying to say you can't analyze a play if you aren't a pro. But it does imply that your conclusion means less than someone trained in the area, no matter if you are right or wrong. And to be fair and because I'm not trying to start anything, this is just as true for statistical analysis in that you should know what the numbers mean and the strengths and weaknesses behind them before asserting anything. Once again, I'm not trying to tell people not to do this anyways. Have fun, make as many assertions as you want, eyes or stats. But don't expect people not to call you out on them if they disagree with similarly defended assertions.
And because I'm a stats guy and this is usually the basis for my argument, I absolutely have to include the last sentence in the section of the book:
"Combining statistical sampling with
audit judgment generally produces a higher-quality audit conclusion than using audit judgment alone." (which obviously goes for statistical sampling alone as well)