Can the coefficients of dummy variables be more than 1 or less than 0? Can coefficients of dummy variables be more than $1$ or less than $0$? I am getting coefficients ranging from $-6$ to $-16$. 
I am specifically asking about the coefficients of dummy variables, not the values of dummy variables.
 A: Yes, coefficients of dummy variables can be more than one or less than zero.
Remember that you can interpret that coefficient as the mean change in your response (dependent) variable when the dummy changes from 0 to 1, holding all other variables constant (i.e. ceteris paribus).
The mean height of people in the United States is around 176 cm for males and 162 cm for females. If we regressed our dependent variable $\text{Height}$ against a dummy variable $\text{Male}$ (which is one for males and zero for females), then in the model
$$\text{Height}_i = \beta_0 + \beta_1 \text{Male}_i + \varepsilon_i$$
we would estimate $\hat \beta_0 = 162$ and $\hat \beta_1 = 176 - 162 = 14$, meaning that the mean height is 162 cm when the dummy is zero (i.e. at the baseline or reference level, which is female in our case) and mean height increases by 14 cm when the dummy variable changes from 0 to 1 (in other words males are, relative to females, 14 cm taller).
If instead we used a dummy variable $\text{Female}$, coded one for females and zero for males, then in the model
$$\text{Height}_i = \beta_0 + \beta_1 \text{Female}_i + \varepsilon_i$$
we would estimate $\hat \beta_0 = 176$ and $\hat \beta_1 = 162 - 176 = -14$, meaning that the mean height is 176 cm for males (baseline) and mean height is 14 cm lower for females relative to males.
(If you are surprised that coefficients of dummies do not have to lie between zero and one, I wonder whether you are incorrectly thinking of the coefficient as the effect on, or value of, the dummy itself?)
