Analysis of THC Urine Drug Test Submission Data
|Weight of Body Fat (Lbs) vs. Days||Smoking Frequency vs. Days|
|User's Metabolism vs. Days||Marijuana Potency vs. Days|
These charts and passing percentile trendlines were quantitatively derived from the urine drug test results of more than 5,000 marijuana users. They represent the four common factors that determine how long marijuana THC metabolites can remain detectable in a user's body:
Urine was tested at a threshold sensitivity level of 50 nanograms of THC metabolites per milliliter, the industry standard.
For explanation, if all else were equal, and your odds of passing a drug test depended on only one factor, like Smoking Frequency, a point on the 75th percentile trendline means that you would have a 75% chance of passing a drug test after that many days. Unfortunately, that is not the case, as there are several variables that impact the calculation. The following is a breif overview of how data was analyzed and produced for our Marijuana Drug Test Calculator.
Before linear regression analysis of data took place, a Wald–Wolfowitz Runs Test was performed. The results of this test failed to reject the null hypothesis that our drug test data was submitted by random chance (2 sided p-value = 0.43). In other words, the order in which data was submitted to us was independent of one another and no pattern was identified. This points to data quality, integrity, and allows us to move forward with analysis.
Datasets were separated into two categories, Drug Test Pass and Drug Test Fail, and then further separated into 12 subcategories: Gender, Age, Body Weight (Lbs), Waist (Inches), Body Fat Percent, Body Fat Weight (Lbs), Smoking Frequency, Marijuana Potency, Amount Smoked Each Day, Estimated Metabolic Speed, Cardio Activity Level, and Muscular Composition Level. Outliers within each subcategory were then identified and removed (when necessary).
Linear regression analysis was then performed on every single Drug Test Pass/Fail subcategory (96+ variables) in regards to time (in Days). Calculations gathered included Regression Statistics (R, Adjusted R^2, etc.) and ANOVA data (degree of freedom, t Stat, p-level, etc.). The following are the results, based on subcategory, in terms of the null hypothesis.
REJECTED the null hypothesis that the following variables had no effect on time (p-value < 0.05):
The six subcategories that rejected the null hypothesis were then re-analyzed together to determine the levels of correlation and covariance. It was identified that Body Weight, Waist, and Body Fat Weight were highly related, as expected. After further investigation into these three variables, the weight of a person's body fat (not body weight) was determined to be the best predictor for how long THC metabolites can remain detectable by a standard urine test.
The percent weight of each variable is also calculated, and is essential to generating the final equations for our calculator. This information will be published at a later time.
As stated before, these are the four common factors. When using our calculator to determine your odds of passing a lab test, seven of the variables are actually used.