The Impact That Steroids Has on Major League Baseball

Presentation by:
Kyle Kim-E
Andy Bisch
Lei Hookano
Joshua Hurst

Research Article: http://faculty.haas.berkeley.edu/rjmorgan/mba211/s...

Completed by faculty members and students at the University of California (Berkeley)

Research study conducted in order to understand the effects of steroid use in baseball, specifically the MLB

Step 1: Research Study: Summary

i. The Problem
This research study addresses the relationship between performance enhancing drugs in the MLB and their impacts on players' performance. It also discusses the relationship between performance enhancing drugs and the effect they have on revenue gains and losses for a MLB franchise

Do performance enhancing drugs play a significant role on MLB players' performances and affect MLB franchises' revenue?

ii. The Data Increases in Salary of Steroid Use:

Researchers examined the relationship of a players OPS (On Base Percentage and Slugging Percentage) vs. salary in 2004

A regressions scatter plot was conducted Researchers determined that an increase of OPS by .100 led to an increase in salary of $2 Million +

Average Team Revenue Increases based on Steroid Use:

Researchers conducted a study comparing the relationship of the average team revenue from 1985-1993 (Pre-Steroid Era) to 1994-2004 (Steroid Era)

They discovered that the average MLB franchise has increased by $52.2 Million since the Pre-Steroid era

The average MLB franchise value rose from $140 Million in 1994 (Pre-Steroid era) to $332 Million in 2004 (Steroid Era)

Average Team Revenue Losses due to Scandal:

Researchers analyzed the expected impact of the Steroid scandal on each MLB franchise by looking at their five main individual revenue streams in the Post-Steroid Era: game receipts, local media, post-season, other local media, and national broadcasts

They used these revenue streams to forecast the expected decrease in revenue for teams and estimated there to be a loss of $111.9 Million in 2005 (Post-Steroid Era)

Fans cannot directly impact the salaries that players make, but they do have a major effect on the revenue of a team because they can express their displeasure by not purchasing tickets to games or not buying and team merchandise

iii. The Data Collected
Increase in Salary of Steroid Use:

Researchers created two methods to determine the effect of steroid use on the increase of OPS and its relationship with a players salary

The “All Player” Method

Researchers compared the OPS for players in the Pre-Steroids Era to the OPS for players in the Steroids Era they collected the OPS for all players in the major leagues for those years they omitted players with fewer than 100 at bats in a season so as not to include pitchers or players with limited effect on the game in those years

The "Steroid Seven" Method

Researchers looked at seven players who have either admitted to or been accused of using steroids, or the general perception of them by the general public is of a steroid user

Rather than compare these players’ performance in the two eras, they chose to look at their performance during the last few years of the Steroids Era (2001-2004) and compare that to 2005, the first year of the Post-Steroids Era

Average Team Revenue Losses and Gains based on Steroid Use:

Researchers analyzed the revenue for all 30 Major League Baseball Franchises based upon the five main individual revenue streams and compared the economic gain from the Pre-Steroid Era to the Steroid Era

iv. The Data Interpreted

The interpretation made by the article is that the punishment for those that choose to use performance enhancing drugs is not severe enough to detract from the possibility and likelihood of economic gain and a significantly increased net present value (NPV)

On the other hand, a chart (pg. 3) that displays the evolution of punishment for testing positive for performance enhancing drugs shows that MLB is moving in the right direction to deter participants from gaining money or fame by any means other than old fashioned hard work and effort

Step 2: Research Study: Limitations

Their method for measuring OPS is limited because its analysis is focused on offensive production

Much of the attention around steroids in baseball has focused on the production of more and longer distance home runs

It lacks analysis of the defensive side and therefore doesn’t accurately represent the whole population of baseball players

Given that:

the dependent variable in this data set is player production/performance

independent variable is use of steroids

the following threats to validity of the dependent variable are made:

random sampling

normal distribution

The study compares players in their definition of a Pre-Steroid Era (1985-1993) to the players of their definition of a Steroid Era (1994-2004)

there is no real marking point for when steroids were brought into the game and also because there could be player cross over from one era to the other - it is hard to say that random sampling actually took place

Additionally, because some players get more playing time than others, it is also hard to say that normal distribution of the dependent variable is true

Because this study was conducted using observance methods of collecting data, it is hard to pinpoint which players were actually measured, as it does not indicate that in the research study

In regards to systematic errors:

the Hawthorne Effect (participants know they are being watched) could be a limitation

Though players may not know they are being watched for this particular study, the subject of steroids is a major problem in Major League Baseball, so players are usually aware that people are watching them in regards to their steroid use
- Whether this affects the findings of the study is hard to tell

Maturation (participants may change over time) is another limitation
- As younger players get older, they usually get stronger and reach a peak, as middle aged players get older after reaching their peak, they tend to weaken
- It is not indicated in the article whether this variable is taken into account, which could affect the findings

Experimental mortality (participants dropping out of study over time) is also a limitation

as the data mentions, the data compares two 8 or more year time frames

according to a research team at the University of Colorado at Boulder (http://www.sciencedaily.com/releases/2007/07/070709131254.htm), the typical career span of an MLB player is 5.6 years, so players will drop out and new players will be drafted within those time frames

In light of these threats, the findings of this study and claims in the supporting article do generalize to other contexts (other sports) and encourages multiple studies to be done not only in baseball, but to expand to all other sports where steroids are seen to be an institutional issue.
- Therefore, it supports external validity.

Step 3: Do We "buy" the Interpretation?

Despite that fact that there are some threats to validity and critiques to be made, we do “buy” the interpretation made in this article.

There was significant empirical data to support the claims of the article and given that they used regression to test the data, they did adhere to the following steps which increases our confidence in the interpretation of this article:

Check Assumptions
Examine Scatterplot
Run Analysis
Write Regression Equation
Make Statistical Decision
Interpret Results

*APA results aren’t shown, but sufficient interpretation is represented

Additionally, with the results of the study, the article describes the relevance to the acts of congress and the effect on baseball as a whole

The article does this by including House Representative excerpts from the Associated and multiple recorded acts of congress in reaction to publicly known steroid usage.

Finally, given its findings, the article outlines possible ways to ban steroid use and suggests further measures that should be taken and how their model is an accurate measure of doing so

We agree with their concluding thoughts and believe the claims of the article are sufficiently supported by the data collected

Step 4: Ideal Way to Conduct Research Study

i. Ideal situations versus Problems

Ideal situations:

Test every single MLB player

do a carbon isotope test on each sample

Allow MLB players that test positive to continue playing

Problems:

expensive to conduct a test for each MLB player

adds on a $1.5 million tag for every test add-on and also extremely time-consuming (3-4 days per test)

the MLB wouldn't allow positive-testing players to play a season out
- wouldn't be able to get performance results of a player using performance-enhancing drugs

ii. Ideal situation: research study
In order to conduct a research study that allowed us to research the following: salary increases, performance-enhancing drug results, and actual performance improvements - we will have to assume that we have our ideal situations

1. Spend $300,000 - $480,000 to test every MLB player (750 to 1200 total)
- $400 per test
- there are 30 teams with 25 players
- after regular season, each team may expand to 40 players
2. Spend $1.5 million to add on the carbon isotope test to each sample received
- Depending on how large of a test lab we had, the tests could take nearly a month to complete
3. After results are received, keep track of which players will be playing throughout the season using performance enhancing drugs (PEDs) and which players will not be playing using PEDs

4. Record each players' performance through stats and run a scatter plot to determine if correlation exists between players taking PEDs versus those that do not take PEDs
- Based on the results from our article, we can hypothesize that there is a significance
5. Record the salaries for players taking PEDs versus those that are not taking PEDs and determine whether there is a significant difference between the two variables
- Based on the results from our article, we can hypothesize there will be a significance

6. Conduct the same research study with the tests and statical information for a number of years, following certain players (if not all of them), and determine whether the data is normally distributed or skewed

7. After 5 years of conducting the research on several different baseball seasons, you establish more valid and reliable data (2014-2019)

8. Run an analysis on the data to understand if there is a significance for the performance and salaries of players between players taking PEDs versus those that are not taking PEDs