Entering a design
Description of the workspace
After pressing "Use" in the welcome screen you will get the following workspace:
Figure 2: the workspace (click on a subject to get more info)
Description of the buttons
We will now describe the parts numbered in the previous paragraph (description of the workspace). The number of each paragraph refers to the number in figure 2.
1. Number of cases per condition
This is the number of subjects that are participating per condition. You should enter a numerical value here. Not entering any value will result in not simulating any data.
Example: If your experiment would consist of testing two groups of 50 persons, this part would look like the following:
Figure 3: Number of cases
2. Independent variable X and dependent variable Y
By dragging these icons (X or Y) into the workspace you can add dependent and independent variables. Both kinds of variables need some input:
2.1 The independent variable X
After dragging an X into the workspace, following window will appear:
Figure 4: the independent variable
When you drag an X into the workspace you will have to enter a variable
number and a condition number. The X will get an index using those number. For
example: variable 1, condition 3 would result in having X13 in the workspace.
Every variable has it's own colour. When the control option is on, the variable will become grey, no matter what colour is had before. The index for the condition number will turn into a C (for example: the control condition for variable 1 would be X1C).
2.2 The dependent variable Y
When you have dragged a dependent variable into the workspace, following dialog box will appear:
Figure 5: the dependent variable
You will need to enter a numerical value for the average
and the standard deviation. When you have a correlation with another
(dependent!) variable, you will need to enter the desired coefficient and the
column you want to correlate with. Read more about correlations in the
Note that the standard distribution of your data is normal. ELEDES will generate continuous data by default, but you can choose to generate discrete (rounded) data. There are no limits by default. Choosing an upper or lower limit will reduce the range of the data and can hereby affect the desired distribution.
3. Assignment of the subjects
Icons were added to indicate how subjects are assigned to the conditions. You
can drag one of these icons into the area indicated by number 4 in the workspace.
You can choose between four methods:
R: Random assignment
E: Existing groups
Note: the assignment method will not affect the data in contrary to reality! It will however notice you what the implications of the chosen method are.
4. Place where the assignment method is dragged
You can drag one of the icons here.
5. Clear all
This button will clear the design.
6. Quit program
When you click this, ELEDES will be closed.
7. Workspace where the dependent and independent variables are dragged into
The workspace exists of 19 columns en 12 rows. It is possible to make a combination of 19 dependent or independent variables. Per variable there is a maximum number of 12 conditions. Within one column there can be only one type of variable.
8. Change the variable name
You can change the variable names in two ways. You can either click the number above each column and rename each variable or you could click the button Variables where you can change the name of every variable.
variables can be correlated and this can be simulated
in ELEDES. As you can see in figure 5 you need to enter a correlation
coefficient. This coefficient may be between -1 and 1. Entering no correlation
means that two variables are calculated independently.
Only dependent variables can be correlated. If you would try to correlate with a dependent variable, no data will be simulated. Of course you can obtain a correlation with an independent variable by adapting the entered averages.
You can correlate a variable with one column. You have to select this column. If this column does not contain a dependent variable, no data will be simulated. Note also that ELEDES will correlate with the variable on the same line. Following figure contains a good and two faulty examples:
Figure 6: example of correlations
Correlations are indicated by a green (correct) or a red (faulty)
The first correlation is a good one.
The second correlation is faulty because the cell it correlates with is empty.
The third correlation is impossible: every variable in a column should be correlated with the same column. Secondly, if the correlation were possible it would be faulty because it's correlated with an independent variable.
Note that it is possible for this design to have a correlation even if for the two last rows there is no dependent variable to correlate with: every row can have a different correlation coefficient. As long as you leave this coefficient unchanged (zero) for the last two rows it will generate data for those two conditions. The first two correlations can be chosen as wanted.
Also note that a correlation can only be calculated with a variable that is in a previous column.