2014-2016 National Survey on Drug Use and Health:
Comparison of Population Percentages from the United States, Census Regions,
States, the District of Columbia, and Substate Regions
(Documentation for CSV and Excel Files)

Documentation for CSV and Excel Files

Description of the CSV File Type

Files with a comma separated value (*.csv) extension are in plain text. They contain characters stored in a flat, nonproprietary format and can be opened by most computer programs. Each *.csv file contains a set of tabular data, with each record delineated by a line break and each field within a record delineated by a comma. A field that contains commas as part of its content has the additional delineation of a quote mark character before and after the field's contents. When a quote mark character is part of a field's content, it is included as two consecutive ""quote mark"" characters.

Computers with Microsoft Excel installed open *.csv files in Excel by default, with the fields automatically arranged appropriately in columns. Other database programs also open *.csv files with the fields appropriately arranged.

The 780 CSV files (i.e., "NSDUHsubstatePValueTab#-#_2016.csv") reflect the 780 Excel tables,1 and they contain the table title, table notes, column headings, and data. The webpage at https://www.samhsa.gov/data/ for the 2014-2016 NSDUH state p value tables includes a hyperlinked table of contents on the first sheet of the Excel file that combines all of the Excel tables, as well as a ZIP file containing all of the individual CSV files. Additionally, the ZIP file includes a text file with a list of the table numbers and titles.

How to Use the P Value Tables

The p values contained in these tables for each outcome and age group can be used to test the null hypothesis of no difference between population percentages for the following types of comparisons:

For example, within any given table, by scrolling across Alabama's state row to the South's census region column, the p value found will determine whether Alabama's state population percentage and the South's census region population percentage are significantly different for a particular outcome of interest. By scrolling across the row for Alabama's Region 2 to the column for the state of Alabama, the p value found will determine whether Alabama's state population percentage and Alabama's Region 2 population percentage are significantly different for a particular outcome of interest (in any Alabama table). Similarly, by scrolling across the row for Colorado's Region 7 to the column for Region 1 of Colorado, the p value found will determine whether Region 7's population percentage and Region 1's population percentage are significantly different for a particular outcome of interest (in any Colorado table). Note that the tests included here are for a given outcome and age group.2

The following example describes how to test the null hypothesis of no difference between population percentages in the national file. Table 6 in the national file contains p values for past month alcohol use among individuals aged 12 or older for all of the substate regions in the United States and for each state and census region. To find the p value for testing the null hypothesis of no difference between population percentages for past month alcohol use between two large metropolitan areas, such as Region 11 (Los Angeles) in California and Region 2: New York City in New York, scroll to the row for Region 2: New York City and navigate to the column for Region 11 (Los Angeles). That p value is 0.004. Thus, the hypothesis of no difference—that is, the Region 11 (Los Angeles), California, population percentage for past month alcohol use is the same as the Region 2: New York City, New York, population percentage for past month alcohol use—is rejected at the 5 percent level of significance, meaning that the two prevalence rates are statistically different. Note that the small area estimates for Region 11 (Los Angeles), California, and Region 2: New York City, New York, are 46.96 and 50.74 percent, respectively.3 Using the same procedure, if users want to test differences in past month alcohol use small area estimates between two adjacent substate regions that are in two different states (e.g., the Badlands region in North Dakota and Region 1 in South Dakota), they can use the national file to find the p value (i.e., 0.834).

The following example describes how to test the null hypothesis of no difference between population percentages in a state file. Table 2.2 in Alaska's file contains p values for past month marijuana use among individuals aged 12 or older for the four substate regions in Alaska and for the West census region. The p value for testing the null hypothesis of no difference between Anchorage, Alaska, and the West census region population percentages for past month marijuana use is 0.002, which is found by scrolling to the row for Anchorage and navigating to the column for the West census region. Thus, the hypothesis of no difference (i.e., the Anchorage population percentage for past month marijuana use is the same as the West census region population percentage for past month marijuana use) is rejected at the 5 percent level of significance, meaning that the two prevalence rates are statistically different. Note that the Anchorage, Alaska, and the West census region estimates are 13.51 and 10.48 percent, respectively.4

Comparison between Two Small Area Population Percentages

To produce substate region, state, census region, and national small area estimates, the 2014-2016 NSDUH data were modeled using the method discussed in Section B.1 of the "2014-2016 NSDUH: Guide to Substate Tables and Summary of Small Area Estimation Methodology" document at https://www.samhsa.gov/data/. This modeling results in 1,250 Markov Chain Monte Carlo (MCMC) samples that are used here to calculate p values for testing the null hypothesis of no difference between two small area population percentages.

Let pi 1 sub a and pi 2 sub a denote the 2014-2016 population percentages of two areas (e.g., state 1 vs. state 2 or state 1 vs. national, or substate region 1 vs. substate region 2) for age group-a. The difference between pi 1 sub a and pi 2 sub a is defined in terms of the log-odds ratio, The log-odds ratio, lor sub a, is defined as the natural logarithm of the ratio of two quantities. The numerator of the ratio is pi 2 sub a divided by 1 minus pi 2 sub a. The denominator of the ratio is pi 1 sub a divided by 1 minus pi 1 sub a, where ln denotes the natural logarithm, as opposed to the simple difference (Pi 2 sub a minus pi 1 sub a) because the posterior distribution of the log-odds ratio is closer to Gaussian than the posterior distribution of the simple difference.

An estimate, lor hat sub a, of lor sub a is given by the average of the 1,250 MCMC sample-based log odds ratios. Let lor i sub a denote the log-odds ratio for the i-th MCMC sample. That is,

The log-odds ratio, lor i sub a, is defined as the natural logarithm of the ratio of two quantities. The numerator of the ratio is pi 2 i sub a divided by 1 minus pi 2 i sub a. The denominator of the ratio is pi 1 i sub a, divided by 1 minus pi 1 i sub a..

Then lor hat sub a is defined as the ratio of two quantities. The numerator is the sum over 1,250 values of lor i sub a. The denominator of the ratio is 1,250., and the variance of lor hat sub a is given by The variance of lor hat sub a is defined as the ratio of two quantities. The numerator is the sum over 1,250 values of the square of the difference between lor i sub a and lor hat sub a. The denominator is 1,250..

To calculate the p value for testing the null hypothesis of no difference, (Log-odds ratio, lor sub a, is equal to zero), it is assumed that the posterior distribution of lor sub a is normal with Mean is equal to the estimate of the log-odds ratio, lor hat sub a. and Variance is equal to the variance v of the estimate of the log-odds ratio, lor hat sub a.. With (Log-odds ratio, lor sub a, is equal to zero), the Bayes p value or significance level for the null hypothesis of no difference is The p value is equal to 2 times the probability of realizing a standard normal variate greater than or equal to the absolute value of a quantity z., where quantity Z is a standard normal random variate, Quantity z is the estimate of the log-odds ratio, lor hat sub a, divided by the square root of the variance of lor hat sub a., and absolute value of quantity z denotes the absolute value of quantity Z. This Bayesian significance level (or p value) for the null value of log-odds ratio lor, say log-odds ratio lor sub zero, is defined following Rubin5 as the posterior probability for the collection of the log-odds ratio lor values that are less likely or have smaller posterior density d of the log-odds ratio lor than the null (no change) value log-odds ratio lor sub zero. That is, The p value of log-odds ratio lor sub zero is equal to the probability of d of the log-odds ratio lor when it is less than or equal to d of the log-odds ratio lor sub zero.. With the posterior distribution of log-odds ratio lor approximately normal, the p value of log-odds ratio lor sub zero is given by the above expression. If the p value is less than 0.05, then it can be stated that the estimates for the two areas are statistically different from each other.

End Notes

1 These include 15 tables for each of the 50 states and the District of Columbia, presenting p values for the substate regions in each state (see Tables 1.1 to 51.15), and 15 national tables that include estimates for all of the substate regions in the United States (see Tables 1 to 15). Estimates for the four census regions also are included in each state and national table.

2 The outcomes in these tables focus on illicit drug use, alcohol use, tobacco use, alcohol use disorder, serious mental illness, any mental illness, mental health services, suicidal thoughts and behavior, and major depressive episode. The age groups include individuals aged 12 or older, or adults aged 18 or older for the mental health outcomes, and individuals aged 12 to 20 for underage alcohol use.

3 See Table 6 of the "2014-2016 National Survey on Drug Use and Health Substate Tables, Percentages (Excel)" at https://www.samhsa.gov/data/.

4 See Table 2 of the "2014-2016 National Survey on Drug Use and Health Substate Tables, Percentages (Excel)" at https://www.samhsa.gov/data/.

5 Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys (Wiley Series in Probability and Mathematical Statistics: Applied Probability and Statistics). New York, NY: John Wiley & Sons.

Go to Top of Page