This slide displays the View Results page for the following query: Grade 8, Reading, all available years, jurisdictions National Public and Virginia, and variable Gender. First, the panel "Are differences statistically significant?" and the "Find Out" button are highlighted. The data table containing the results of the query is highlighted next with an outline box.
Statistical significance indicates whether data is worth paying attention to, rather than a statistical fluke. The first step in setting up a significance test is deciding what data you want to compare.
This slide displays the View Results page with the same query as before. Outline boxes appear first over the Average Scale Scores column, then the Years column, Jurisdictions column, and variable Gender column. The table fades except thode data for National Public, highlighting those data. Outlines highlight the isolated data to be used in the example significance test to follow.
All NAEP data depend on several criteria, but a significance test can test only one of those criteria -years, jurisdictions or variables. Choosing one criterion to vary limits options for the others. This isolates a cross-section of data for the test. The following example will compare trend data over time.
This slide displays an animation of the View Results page. The cursor moves to click the "Find Out" button. The Check Statistical Significance popup window appears. In the popup window, no selections have been made and catagories are collapsed.
To test for significance, click the "Find out" button on the View Results page in either mode.
The cursor selects Jurisdiction to expand that category and then selects National Public. The cursor then selects Year to expand that category and then clicks the Select All choice. The Table Preview updates with a sample table and a red arrow appears next to the Variables category. The cursor selects Variables to open that category and then Gender to open that category. Then the cursor clicks Select All to select both Male and Female. The Table Preview updates once again to a sample table.
Make selections where indicated by the red arrow. Multiple selections are possible in some cases. Variables become available after Jurisdictions and Years are selected. Display Options lets you add extra information in your table.
This slide displays the Check Statistical Significance table from before. The cursor moves to point out features in the Table Preview panel including the sample table and the description built from selected options.
As variables are selected, the preview panel builds the table visually. It reflects all changes made in criteria selections.
The cursor moves to click the Compute button at the top of the Table Preview panel. The sample table changes to an actual statistical significance table. A legend appears at the bottom of the window and is outlined with a box. One row is then outlined with a box, the row comparing 2002 to other years. It shows that 2002 was significantly higher than all other years, including 2003.
When finished, click the "Compute" button. The table will appear, along with legends to explain their contents. If multiple tables are created, scroll vertically to see them. Read across a row to compare a chosen criterionin this case a yearwith other criteria.
A cutout view of the table on the Go To Results page overlays the Statistical Significance table. The cursor moves to point out year 2002 average scale score of 258. The cursor then moves to point out the "significantly higher" symbols in the table indicating that the score for 2002 is significantly higher than 1992, 1994, 1998, and 2003.
In this case, the average scale score for males nationwide in 2002258is significantly higher than for males in 1992, 1994, 1998 and 2003. This finding raises confidence that 2002 marked a high point for 8th grade male reading scores.
Return to Lesson Menu