The Bass-Ackward procedure explores hierarchical relationships among factor solutions with different numbers of factors. WebPower runs the analysis on the numeric variables in the uploaded data, determines the largest factor solution that meets its automatic stopping rule, and displays a diagram and the factor loadings.
The automatic stopping rule requires every factor in a solution to have at least one variable for which it has the largest loading. The search proceeds from one factor up to one fewer than the number of numeric variables and stops when this condition is no longer met or a solution cannot be estimated. The resulting number of factors is a suggested solution, not a substitute for evaluating the factors and their substantive meaning.
Arrange the data so that each row represents an observation and each column represents a measured variable. The analysis uses numeric columns only; nonnumeric columns are dropped, even if they are included in the selected columns.
The following input types are available:
For CSV and TXT files, select the option that matches whether the first row contains variable names. Make sure the data are arranged using the delimiter and format appropriate to the selected file type.
Upload the data file to analyze.
Choose the format of the uploaded file. For CSV and TXT files, also indicate whether the file contains variable names.
Leave this field blank to use all columns. To select columns, enter their column numbers separated by commas. Ranges can be written with a hyphen or a colon. For example, 1, 2-5, 7-9, 11 selects columns 1, 2, 3, 4, 5, 7, 8, 9, and 11.
Column numbers refer to the columns in the imported data. After selection, nonnumeric columns are removed automatically.
Enter values in the data that should be treated as missing. Separate multiple values with commas; for example, -999, -888, NA. Leave this field blank if the data do not use additional values to represent missing data.
The analysis may take a short time to finish. The results page refreshes while it is running.
The results include:
Use the diagram and loadings together when interpreting the hierarchy. Factor labels and apparent groupings are descriptive; substantive interpretation depends on the measured variables and the research context.
If the analysis fails, check that the file matches the selected format, the selected column numbers exist, and the data contain numeric variables suitable for factor analysis.