home *** CD-ROM | disk | FTP | other *** search
- STATS 2 MENU ITEMS
-
- DISTRIBUTIONS
- This set of items calculates probabilities for the given
- distribution based upon values which you input in response to
- questions. No data variables are used.
-
- -T-Tests
- There are three T-tests available. Test 1 calculates the chance
- that a given variable has a certain mean. You are asked to choose
- a variable and to specify a mean to test.
- The second test checks the chance that two variables have the same
- mean. You have to select two variables. The two variables should
- be randomly selected and unrelated.
- The third T test is designed for two variables which are related.
- Once again you simply select two variables.
-
- -Multivariate
- There are three somewhat similar procedures: Factor analysis,
- Discriminant analysis and Canonical Correlation analyses.
-
- For factor analysis, you first select the variables which you
- want in the study. You must then select the type of rotation
- desired. If you choose an orthoblique rotation, then an
- orthoblique factor must be entered. The orthoblique factor must
- be from 0 to 0.5.
-
- In discriminant analysis, you must first select the independent
- variables. You next choose the dependent variable.
-
- In both procedures the data entry screen is used to display
- results. No text input can be made, but all menu functions
- and arrows work.
-
- You are given an opportunity to save the results to the data
- matrix so that you may use them in other work. Be carefull however
- as this is a destructive process. Your original data will be lost.
-
- The third choice is for Canonical Correlations. In this case you
- must pick two sets of variables. The number of variables in the
- second set must be less than or equal to those in the first set.
- You will next be asked if you want to save the canonical variates.
- This is a destructive process and all existing data will be wiped
- out. The output is similar to that for discriminant analysis.
-
- There are two other tests which are not similar to the previous
- three.
-
- -Generalised Distance. This selection is used for testing the
- normality of multivariate data sets. You must select the
- variables for the test plus one for the distance and one for the
- chi square critical values. If the data set is normal two things
- should be true. One the percentage of distances inside the 50%
- contour should be about 50%. Equally, if you graph the distances
- against the chi square values as an XY graph the resultant should
- be a straight line and at a 45 degree angle. If you regress the
- chi square values against the distances the regression should
- have a very high R squared and the coefficient should be near 1.
-
- As well there is a procedure for determining the real eigen
- values and associated eigen vectors of square matrices. You will be
- asked to specify a range of data for the matrix. This is done
- in the same way as for specifying a range to read from a 123
- spreadsheet. The program will return the answers or an error
- message if the area selected is not square.
-
- -ANOVA
- The Analysis of Variance section allows for flexibility in data
- structure. The one-way studies all require that the data be set up
- in a traditional ANOVA matrix structure. Questions will not be
- asked for the random, blocked, or latin square designs. For
- nested, you will be asked to specify the number of treatments in a
- block. For the 2- and 3-way ANOVA designs, either the data can be
- set up in a matrix (the default), or separate variables can be
- used to represent the levels of the variables. In a 3-way design
- not using matrices, you would be asked for the variable that held
- the results and then, in turn, for the variable containing the level
- information for factors A, B, and C. For data in matrix form, you
- would need to specify the number of levels in A and B. For the 2-
- factor case the input required is similar.
- The non parametric two factor method can be used when the data is
- not normally distributed. It relies on ranking data. The input is
- the same as for regular 2 factor except that no residual
- variaable is allowed. It uses a chi square statisic. It is not as
- powerfull a test as regular 2 way ANOVA and should be used only
- if necessary.
-
- -Variance Tests
- The 1-factor test asks for a variable name and a value to test.
- The value tested is the standard deviation and the program will
- give back the chance of the variable was chosen from a population
- with the given standard deviation. The 2-factor test compares two
- variables to determine the chance that they are both drawn from a
- population with the same standard deviation.
-