Appendix C
Methodology and Regression Output
To compare the number of licensed or registered hair braiders per 10,000 black immigrant or African-Americans among the states, this analysis measures differences using a time-series GLS regression with robust standard errors. The number of braiders per 10,000 black immigrant or African-Americans (γ) was regressed on state (βx) and year (θ) dummy variables, which control for state- and year-based effects on the number of licensed braiders. The following is the model used:
γ = β0 + βx (States) + θ + ε
Table C1 provides the results of the equation. Oklahoma (600 hours) is the control state and 2006 the control year.
Table C1: Regression Output of Braider Population Analysis
| Coefficient | Robust SE | p | |
|---|---|---|---|
| State (hours) | |||
| Louisiana (500) | 0.119 | 0.000 | 0.00 |
| Ohio (450) | 0.174 | 0.000 | 0.00 |
| Illinois (300) | -0.189 | 0.543 | 0.73 |
| New York (300) | 0.101 | 0.000 | 0.00 |
| Pennsylvania (300) | -0.739 | 0.391 | 0.06 |
| Tennessee (300) | 0.801 | 0.000 | 0.00 |
| Nevada (250) | -0.607 | 0.543 | 0.26 |
| District of Columbia (100) | 1.935 | 0.000 | 0.00 |
| Texas (35)* | 2.890 | 0.000 | 0.00 |
| Florida (16) | 7.825 | 0.000 | 0.00 |
| South Carolina (6) | 12.313 | 0.000 | 0.00 |
| Mississippi (0) | 5.614 | 0.000 | 0.00 |
| Year | |||
| 2007 | 0.786 | 0.516 | 0.13 |
| 2008 | 1.370 | 0.947 | 0.15 |
| 2009 | 1.724 | 1.076 | 0.11 |
| 2010 | 2.355 | 1.455 | 0.11 |
| 2011 | 2.417 | 1.320 | 0.07 |
| 2012 | 3.091 | 1.644 | 0.06 |
| Intercept | -1.614 | 0.979 | 0.10 |
| sigma_u | 0.000 | ||
| sigma_e | 1.650 | ||
| rho | 0.000 |
* Texas stopped licensing hair braiders in 2015.