The two preceding topics have discussed how the MDO and did-not-see options affect Statsf and Statsb reports. Now: Lertap's third report for cognitive items, Statsul -- what happens when the MDO and did-not-see options are used?
First up, above, the standard format for Statsul, what's seen before the MDO and did-not-see options are put to use.
Item Q50 is from the Test Pilot system mentioned in the previous topics. Over 400 students took the "EE 101 T3" exam over the internet; 25.2% of the students did not see Q50 as Test Pilot did not present it to them, while just under 1% of the students who saw Q50 didn't answer it.
The table above gives Q50 results after the MDO option has been turned on. Its statistics are based only on those students who answered the item. This means that all the did-not-see people, plus the we-saw-it-but-did-not-answer-it people, have been excluded from the proportions and from the calculations underlying U-L diff. and disc. The did-not-sees are excluded at this point as the did-not-see option has not yet been used; the did-not-see code has been processed as missing data.
In this table (above), the did-not-see option has been activated, but MDO has not. The "other" column is now indicating the proportion of students in each group, from upper down to lower, who were presented with Q50 by Test Pilot, but did not answer it. Those not presented with Q50, the did-not-sees, have been excluded.
For a refresher on how the U-L diff. and U-L disc. values are calculated, pay a visit to your local ice cream shoppe, then have a look at Chapter 10 of the manual.