Checklist of data problems:

Here is a list of issues which can cause problems with the running of the program, or give unexpected results. We suggest that you run through the list to eliminate simple causes like these, before contacting us for support.

  1. Many software problems are transient, so first, try closing the program and restarting it.
  2. If you are opening an Excel spreadsheet, have any recent changes to the data set been saved in Excel, and was your data set on the active worksheet (the visible one) when the file was saved?
  3. Are all the columns labelled, with no duplicates? They should be.
  4. Is cell A1 (the top left-hand corner of the data set) blank? It should be.
  5. Have you got any columns or rows in the data set which you had not intended? For instance, was a row of column totals, means or standard deviations added in the original spreadsheet?
  6. Are there any blank cells, or cells containing non-numerical data, in the data set? There should not be, apart from the column and row headers.
  7. Are there any values (or even a non-numerical character, or a space) entered into cells outside the main block of data? The easiest way to do this is to open the data set in a spreadsheet, select the first 10 or so columns to the right of the data, and the first 10 rows below it, and press the Delete key, to clear any unwanted cell contents. Then save the data set and try again.
  8. Do your data have no variability - i.e. are all the numbers the same?
  9. Are the data sets perfectly correlated - either positively or negatively? This will occur if one variable is a simple factor of another.
  10. Is your data set in the right format for the analysis you want to perform? Do you have the correct number of columns and rows? If in doubt, use the Data Entry Wizard.
  11. Have you got real numbers (i.e. 2.35, 1.796), when the analysis requires integer data (1, 2, 3)?
  12. If the analysis depends upon having fixed categorical or covariate variables, are you sure that the range of data is consistent with this? For instance, a fixed categorical variable must run from 1 to n.
  13. Is the data set excessively skewed, or in some other way out of the ordinary? Check Skewness, Kurtosis and the Histogram plot.
  14. If you are comparing two or more samples, have you considered the possibility that their variances might be equal?
  15. If you are performing a Multiple Linear Regression, have you checked for multicollinearity? This is when two (or more) variables are not truly independent, but can be expressed as a function of each other. One or more redundant variables from a set of directly-related variables should be removed.

For a full explanation of how to input data, and many other aspects of the program, please see the demonstrations.

If you still have problems using the program or entering data which you cannot solve, then contact Pisces Conservation by e-mail or phone, during office hours (09.00 to 17.00 UK time).