Statiscope

Help for Statiscope 1.0beta8 Applet and Application

Getting Started

At the text boxes near by the Simulate button, enter 170 in the Number box, 90 in the Mean box, and 30 in the Std. dev. box. Click the Simulate button. Click through the radio buttons Distribute, Probability mass, etc. Click the button Sort to sort and then on Clear data to empty the list. The Stop button is used to interrupt lengthy operations.

Data Entry

Enter a number in the text box Data/Location and press Enter. Continue to enter a few numbers. Decimal numbers require a point as a decimal separator. You can enter several numbers at a time by separating them with commas (or space).
Try e.g. this list: 5, 3.5, 9, 3.5
You can mark it and copy it with Ctrl+C in Windows (Cmd+C in Macintosh, only marking in X-windows), click the Data/Location box and paste the clip with Ctrl+V in Windows (Cmd+V in Macintosh, the middle mousebutton in X-windows) and press Enter. This is regrettably the only method you can use to copy and paste in a Java applet.

It is possible to delete numbers in the list by clicking them and pressing the Delete or Backspace key. You can also insert numbers into the list by clicking where you want the new number to appear and pressing the I key. (Insert would have been more appropriate but it regrettably won't work in Java.) When you are in the list you can move up and down by pressing the up and down arrow keys, but this also works when you are in the text box Data/Location. If you edit a value and press Enter, the value is updated. To add new values you must be at the end of the list.

In the application you can also use the functions of the File menu. Open lets you choose a file to load. Save is used to store the data to a file.

Another method of getting data into the program is to write an URL in the Data/Location box. You can open a new browser and bring up the page:
http://www.maths.lth.se/matstat/datalib/
(Click the link, and a new browser window is opened automatically.) Click to bring up foton.dat, copy the URL from the browser and paste it into Statiscope's Data/Location box and press Enter. Now the first column will be read into the program. If you want to load another column, enter the URL comma the column number.
E.g. http://www.maths.lth.se/matstat/datalib/foton.dat,2 will read the second column.

A special link can be used to get Statiscope to load your data automatically. This is explained on the Statiscope URL Encoder page.

The last method, simulation, has already been described at "Getting Started". The simulation yield normally distributed numbers.

Statistical Charts

By clicking the radio buttons at the left, the corresponding statistical chart is showed. There are real charts and there are text diagrams. The scales are automatically calculated, but the ticks on the x-axes can be manipulated by pressing the < and > buttons. The distance between the ticks is always 1, 2, or 5 times a power of ten. If the check box Rel. is checked the corresponding relative function is showed instead of the absolute. If the check box Theo. is checked the corresponding theoretical function based on the normal distribution with the same mean and standard deviation as the measured data is showed.

Distribution

This shows the empirical distribution function, that is the height of the graph for a certain x-value is the probability that a value will be less than or equal to that value.

Probability mass

This shows the probability mass function, that is the number of data points with a certain x-value. If Rel. is checked it is instead the probability of the value. This is a bar diagram.

Density

This shows the density function, that is the number of data points that fall within the different intervals which the x-axis is divided into. If Rel. is checked it is instead the probability of falling within the interval is showed. This is a histogram.

Box plot

This shows a box plot over the data. The vertical lines in the box plot correspond to - from left - minimum, 1st quartile, median (or 2nd quartile), 3rd quartile, and maximum. This gives a summary of the data.

Stem and leaf

This shows a stem and leaf diagram over the data. It shows a kind of lying histogram. It works best for positive integers when the difference between maximum and minimum is below a few hundreds. It's easiest explained by an example.

Data listing

This is the cat among the ermines. It simply shows the same data as in the list. The purpose of this is that you can copy data from this print-out and paste it into an editor, to perhaps finally publish your data on the net.

Number

The number of data points.

Sum

The sum of all data points.

Mean

The mean, or empirical expectation value, that is the sum of the data points divided by the number of data points.

Sam. std. dev.

The sample standard deviation, this is the measure of the spread in the data which is used most often. Especially when the data is a sample from a larger or infinite population. It is the root of the sum of the squares of the differences between the data points and the mean divided by the number of data points less one.

Pop. std. dev.

The population standard deviation, is used if the data is collected from the whole population. It's seldomly used in practice. It is root of the sum of the squares of the differences between the data points and the mean divided by the number of data points.

Minimum

The least of the data points.

Quartile 1

The 1st quartile, that is the median of the data points between the least value and the median of all data points.

Median

The median, that is if you sort the data it is the middle value or the mean of the two middle values, depending on if the number is odd or even.

Quartile 3

The 3rd quartile, that is the median of the data points between the median of all data points and the largest value.

Maximum

The largest value.

Confidence Interval and Hypothesis Testing

The percentage of the confidence intervals does only affect the confidence intervals and not the hypothesis testings. But the radio buttons Lower, Middle, and Upper affects both the confidence intervals and the hypothesis testings.
[t- and Chi-2-distributions are used. The algorithms for Chi-2 will probably need some improvement.]

Confidence Interval

Confidence interval in general means that the probability that the real value should be within the interval is equal to the Confidence percentage (e.g. 95%).
Mean is the confidence interval for the standard deviation. You can determine lower, two-sided, and upper confidence intervals, by selecting Lower, Middle, and Upper, respectively.

Hypothesis Testing

By hypothesis testing you can examine if a certain hypothesis should be accepted or rejected. If the alternative hypothesis is used as a new hypothesis and is rejected, the former hypothesis is statistically proved.
The hypothesis is H₀ and the alternative hypothesis is H₁.

If
  H0: real value <= hypothesis,  H1: real value > hypothesis,
use Lower.
If
  H0: real value = hypothesis,   H1: real value <> hypothesis,
use Middle.
If
  H0: real value >= hypothesis,  H1: real value < hypothesis,
use Upper.
   

The hypothesis is rejected if the p-value is less than a certain fixed level e.g. 0.05, called the alpha-level. The higher the level the stricter the test. The hypotheses is entered in the boxes below Hypothesis and the p-value is showed in the boxes below P-value.

Right Browser

For the applet Netscape Navigator 3.01 is recommended, earlier versions have had problems with the behaviour of the list box. If certain controls should not be painted you can try to minimize the browser and restore it again. An alternative browser is Microsoft Internet Explorer 3.0.

Comments Welcome

Bug reports and suggestions for improvements on the program Statiscope and this documentation is gratefully received by Mikael Bonnier.

DISCLAIMER

THIS PROGRAM IS USED AT YOUR OWN RISK. IT SHOULD SPECIFICALLY NOT BE USED IN HIGH RISK ACTIVITY SUCH AS DIAGNOSIS AND TREATMENT OF PATIENTS, OR THE OPERATION OF NUCLEAR FACILITIES OR WEAPONS SYSTEMS.

The program Statiscope and this documentation is Copyright © 1996-1997 by Mikael Bonnier, Lund, Sweden. All rights reserved.