MedicalStat

## Tables

Decimals:

The Standard Normal Distribution: $$N(0,1)$$

### The z-value and the p-value to be entered in the table below are defined to satisfy the following:

$$p = 2 \ \cdot P(Z > |z|) = 2 \ \cdot \int_{|z|}^{\infty} \phi(z) \: dz$$

### Where $$\phi(z)$$ is the standard normal distribution with mean 0 and standard deviation 1:

$$\phi(z) = \frac{1}{\sqrt{2\pi}} \ \cdot \text{e}^{-\frac{1}{2} z^2}$$

### For example:

$$0.05 = 2 \ \cdot 0.025 = 2 \ \cdot P(Z > |1.96|) = 2 \ \cdot \int_{1.96}^{\infty} \phi(z) \: dz$$

### The p-value is the two-sided (or two-tailed) area under the bell curve; one area from each "tail". The two areas are the same, since the function is symmetric. If you only need (or only have) half of the p-value, the table can still be used; just enter the |z| value and get the p and ½p values. Or enter either p or ½p to get the other two values.

Enter either |z|, p or $$\frac{1}{2}$$p to get the other two
Standard Normal distribution X ~ N(0, 1) table
|z| p (two-sided) $$\frac{1}{2}$$p (one-sided)

The Normal Distribution: $$X \text{~} N(μ , σ)$$

### If you have a general normal distribution with mean μ and standard deviation σ, and need the area under the bell shaped curve from x1 to x2 then the p value is defined as follows:

$$p = P(x_1 < X < x_2) = P(X < x_2) - P(X < x_1) = \int_{x_1}^{x_2} f(x) \: dx$$

### Where $$f(x)$$ is (in this case) the normal distribution with mean μ and standard deviation σ :

$$f(x) = \frac{1}{\sqrt{2\pi \sigma^2}} \ \cdot \text{e}^{-\frac{(x - \mu)^2}{2 \sigma^2}}$$

### If you need the lower bound to be minus infinity ( $$-\infty$$ ) or the upper bound to be infinity ($$\infty$$), in other words if you need one of the following areas;

$$p = \int_{-\infty}^{x_2} f(x) \: dx \:\: \color{blue}{\text{or}} \:\: \color{black}{p = \int_{x_1}^{\infty} f(x) \: dx}$$

### then you can type either "-INF" as the x1 value or "INF" as the x2 value.

Two of the four inputs must be μ and σ in order to get the fifth value
Normal distribution X ~ N(μ, σ) table
Mean (μ) SD (σ) x1 x2 p-value

The F Distribution

### An f-test is performed to test whether the variances of two samples can be assumed equal. In other words whether the null hypothesis is rejected or not rejected. The null hypothesis states that H0: Var1 = Var2 , where Var1 = (SD1)2 is the variance of the first sample and Var2 = (SD2)2 is the variance of the second sample. Consequently, $$SD_1 = \sqrt{Var_1} = \sqrt{(SD_1)^2}$$ is the standard deviation of the first sample. And $$SD_2 = \sqrt{Var_2} = \sqrt{(SD_2)^2}$$ is the standard deviation of the second sample involved in the f-test. Another way of stating the H0 hypothesis is that the ratio of the variances is 1:

$$H_0 : \frac{(SD_1)^2}{(SD_2)^2} = 1 \:\: \text{(or)} \:\: H_0 : \frac{(SD_2)^2}{(SD_1)^2} = 1$$

### The f1-value, the f2 value and the corresponding p1 and p2 values are defined as follows:

$$f_1 = \frac{(SD_1)^2}{(SD_2)^2} , f_2 = \frac{(SD_2)^2}{(SD_1)^2} = \frac{1}{f_1}$$

### The number p1 is the smallest of the two following numbers:

$$p = \int_{f_1}^\infty{f(x)} \: dx \: \: \: \: \: \: \text{and} \:\:\:\:\:\: 1-p$$

### The number p2 is the smallest of the two following numbers:

$$p = \int_{f_2}^\infty f(x) \: dx \:\:\:\:\:\: \text{and} \:\:\:\:\:\: 1-p$$

### The overall p-value in an F-test, taking into consideration both f1 and f2, is therefore the following number:

$$p_{total} = p_1 + p_2$$

### The function $$f(x)$$ is the probability density function of the F-distribution defined as:

$$f(x) = \frac{df_1^{\frac{df_1}{2}} \cdot df_2 ^\frac{df_2}{2} \cdot \Gamma\left(\frac{df_1 + df_2}{2}\right) \cdot x^{\frac{df_1}{2}-1}}{\Gamma\left(\frac{df_1}{2}\right) \cdot \Gamma\left(\frac{df_2}{2}\right) \cdot \left(df_2 + df_1 \cdot x\right)^{\frac{df_1 + df_2}{2}}} ,$$

### Use the following table if you have the standard deviations SD1 and SD2 from the two samples:

Enter the four inputs SD1, SD2, DF1 and DF2 to get the remaining five
F-distribution table in the case of known standard deviations
SD1 SD2 DF1 DF2     f1         f2         p1         p2     ptotal

### Use the following table if you have the variances Var1 = (SD1)2 and Var2 = (SD2)2 from the two samples:

Enter the four inputs (SD1)2, (SD2)2, DF1 and DF2 to get the remaining five
F-distribution table in the case of known variances
(SD1)2 (SD2)2 DF1 DF2     f1         f2         p1         p2     ptotal

### If you only need to find a specific p-value corresponding to a particular f-value (or the other way around), in other words an f and a p value satisfying that

$$p = P(X > f) = \int_{f}^\infty{f(x)} \: dx ,$$

### you can then instead use the table below over f-distribution values given DF1 and DF2 and either p or f:

Two of the three inputs MUST be DF1 and DF2 in order to get the fourth.
F-distribution table
DF1 DF2     f         p

The T-Distribution

### The |t|-value, the DF-number and the corresponding p-value are defined as follows:

$$p = 2 \ \cdot P(T > |t|) = 2 \ \cdot \int_{|t|}^{\infty} f(x) \: dx ,$$

### where $$f(x)$$ is the probability density function of the T-distribution with DF degrees of freedom:

$$f(x) = \frac{\Gamma{\left( \frac{df+1}{2}\right)}}{\sqrt{\pi \cdot df} \cdot \Gamma{\left(\frac{df}{2}\right)} \cdot \left( 1 + \frac{x^2}{df}\right)^{(df+1)/2}}$$

### |t| is the numerical (positive) value of the t-value. You can input either the two-sided (two-tailed) p-value in order to get the |t| value and the one-sided ½p value. Or you can input the one-sided ½p value and get the |t| value and the p-value. Or input the |t| value and get p as well as ½p. For more details, go to the page "Formulas" in the menu.

One of the two inputs MUST be DF in order to get the remaining two values
T-distribution table
t-value DF p (two-sided) $$\frac{1}{2}$$ p (one-sided)

### The p-value will then be the integral:

$$p = P(t_1 < T < t_2) = \int_{t_1}^{t_2} f(x) \: dx ,$$

### The p-values would then in those cases be:

$$p = P(-\infty < T < t_2) = \int_{-\infty}^{t_2} f(x) \: dx \:\: \color{blue}{\text{or}} \:\: \color{black}{p = P( t_1 < T < \infty) = \int_{t_1}^{\infty} f(x) \: dx}$$

One of the three inputs MUST be DF in order to get the fourth
T-Distribution table
DF t1 t2 p-value

The Chi-Square-Distribution ($$\chi^2$$)

### The χ2 - value, the DF-number and the corresponding p-value are defined as follows:

$$p = P(X > \chi^2) = \int_{ \chi^2 }^{\infty} f(x) \: dx ,$$

### where $$f(x)$$ is the probability density function of the chi-square-distribution with DF degrees of freedom:

$$f(x) = \frac{x^{\frac{1}{2} \cdot df - 1} \cdot \text{e}^{-\frac{1}{2} \cdot x}}{2^{\frac{1}{2} \cdot df} \cdot \Gamma\left( \frac{df}{2}\right)}$$

### In other words the p-value is the area under the curve from χ2 to infinity. For more details, go to the page "Formulas" in the menu.

One of the two inputs MUST be DF to get the third
Chi-Square-Distribution table
$$\chi^2$$ value DF p-value