Correlate Groups Report
The Correlate Groups Report is used to view correlation values between total raw scores, proficiencies for each construct, and selected case-level variables for two groups of cases. The options are selected in the Correlation Report Wizard (see Figure 1). ConstructMap computes product-moment correlations to compare variables within cases and between groups from .
When a demographic field defines two groups, ConstructMap temporarily assigns the values of 0 and 1 to the grouping variable. The original values are retained in the database.
- To view the report, select Reports – Group Reports - Correlate Groups.
- Note: If no demographic fields are listed in the window, ConstructMap is indicating that it could not identify a binary user-defined demographic field. The Correlation Report requires binary grouping criteria. Use the Filter Cases option to reduce the number of categories to two.
- Once the desired demographic fields have been selected, click OK to view the correlation report. For this example, we selected "Class" as the demographic field to examine after excluding the class "02/031" through filtering.
- Since this report produces text file output as well as an output screen, you can print this report using Word or Notepad. The file will be located in the folder you specified (note the filename in the upper left-hand corner of the heading area).
- Close the map display by clicking on the close box, , in the upper right-hand corner.
This will bring up the Correlation Report Options dialog window, shown in Figure 1, where users select the title of the report, the filename to store the report, and the demographic fields to be included in the correlation calculations. Multiple demographic fields are selected by holding down the Ctrl key (or the key on a Mac) while selecting fields with the mouse.
As shown in Figure 2, the report heading indicates that a total of 54 cases were considered for the correlation analysis and that raw scores were calculated and proficiencies estimated from 5 items.
The body of the report is laid out in a matrix, with the total raw score, each variable, and any user-selected demographic fields listed as the column and row headings. Note that values above the diagonal are correlations while values below the diagonal are the numbers of cases used in computing the correlations.
The first row of data in the matrix show that the correlation of raw scores to the MLE proficiencies computed on the “DCI” dimension is 0.577, to the "ET" dimension is 0.548, and to the class the student was in (i.e., in class “02/034” rather than class “01/001”) of 0.007. Using the terminology chart below, we would say that the raw scores are moderately correlated with the MLE estimates, but are not correlated with class membership.
|Range of r||Magnitude Term|
|00 - .20||Negligible|
|.20 - .40||Low|
|.40 - .60||Moderate|
|.60 - .80||Substantial|
|.80 - 1.0||High|
The second row of data show that the MLE proficiencies on the DCI variable had low correlation with the profieicneis on the ET variable (r = 0.218). We also note that the profiencies on the DCI and ET variables had negligible correlation with class membership.
Looking at the bottom half of the matrix, we find that 46 cases had response data on items associated with the DCI variable, 48 cases had response data on items associated with the ET variable, and 45 cases had response data on both variables.
ConstructMap assumes that selected demographic fields are numeric. When a user-selected demographic field contains strings, ConstructMap assigns numeric values, beginning at 0, to unique strings. For example, in the case of Gender, with values “M” or “F” ConstructMap could assign the value 0 to either “M” or “F,” depending on which it encounters first in the data set. ConstructMap reports these automatic assignments below the correlation table. For the example displayed in Figure 2 we see that ConstructMap assigned the value 0 to the class identified as “01/001” and the value 1 to the class “02/034.”