Main Features

The Statistics Pac presents a set of menu driven options
from which it is easy to set up the specific problem to be solved. You can prepare your
raw data set ahead either on the calculator or on a PC and load the data into the
statistics application from the calculator command line, or generate the data in the
application input environment. There is a friendly text editor, and a case sensitive
Search mode. The final set of computations is accomplished with a few menu key strokes.

**
**Matrix Manager mean

This is the core area of the software, serving several
purposes. It allows the user to add new matrices, create new directories, edit existing
matrices, group matrices to create frequency distributions, copy matrices to the stack,
sort matrices by rows or columns, transpose matrices, and copy, transfer, and translate
data between the HP 48 and a computer.

**
**Basic Statistics

Features an interactive user interface prompting for types
of central tendency, dispersion measures and standard errors.

- Parameters of interest in Central Tendency include: number
of data points, total, mean, median, mode, geometric mean, harmonic, first decile, first
quartile, third quartile, and ninth decile.

- Dispersion measures of interest include: minimum, maximum,
range, mean deviation, population standard deviation, sample standard deviation,
population variance, sample variance, coefficient of variation (relative variation as a
percentage), moment of coefficient of skewness (third moment about the mean), and moment
coefficient of kurtosis (fourth moment about the mean).

- Standard error computation includes standard error of the
mean, standard error of the median, standard error of the sample standard deviation and
standard error of sample variance.

**
**Probability Distributions

The software covers b , binomial, negative binomial,
chi-square, exponential, F, gamma, geometric, hypergeometric, normal, Poisson’s,
student-t, uniform, and Weibull’s distributions. The user input form will request
known distribution parameters, and from there the software can compute the probability
function value, the upper tail value, the mean and the variance of the probability
distribution function. Systematic analysis of probability distribution functions is made
easy.

**
**Estimation

Provides interactive prompts for eight hypothesis tests,
seven confidence intervals, and two estimations for sample size for normal hypothetical
tests and confidence intervals.

**
****Hypothesis testing** - This menu covers:

Normal for m Proportion for p

t for m Chi-squared variance for s ^{2}

Normal for m 1-m 2 (s 1-s 2) t for m 1-m 2 (s 1=s 2)

t for m 1-m 2 (s 1¹ s 2) F for s 12 /s 22

**
****Confidence Intervals - **This menu computes lower and
upper confidence limits parameters. Menu items include:

Normal for m Proportion for p

t for m Chi-squared variance for s ^{2}

Normal for m 1-m 2 t for m 1-m 2 (s 1=s 2)

F for s 12 /s 22

**
****Sample Size - **Estimation of sample size handles
population parameters, level of significance a , (also called type I error or
producer’s risk), and level of significance b (also called type II error or
consumer’s risk).

**
**Plotting Manager

Provides an interactive means for plotting data in a
variety of formats, including histograms, bar plots, scatter plots, relative frequencies,
cumulatives frequencies, and ogives.

**
**Regression Analysis

Through an interactive prompt, the software performs linear
regression of the data entered according to a number of models including linear,
exponential, logarithmic, and power fits. The calculation of correlation coefficients is
done using Pearson’s approach, wherein the uncertainty of data entered as abscissa is
considered as a key component. This is in sharp contrast with the traditional method of
computing correlation coefficents where there is no provision made for uncertainty in data
enetered as abscissa.