Open the data in the folder pf10_dich. What is the estimated theta for the person with research id: 716405B (look in Name column).
- Compare this person’s theta (person 716405B) with all of the item deltas. Which items would you expect this person to get a 1 score at least 50% of the time?
- Compare this person’s theta (person 9413932) with all of the item deltas. Which items would you expect this person to get a 1 score at least 50% of the time?
- Which item (or items) is the most difficult and which item (or items) is the easiest.
- Which item seems to be problematic in terms of not fitting the Rasch Model very well. (Use the outfit statistic). How would you describe this item using variable screening methods for skewness and kurtosis. Any evidence for floor or ceiling effects.
- For the item identified in #7, what does the infit value suggest with respect to this item. Is the item contributing less to the overall estimate of the thetas in the sample than the other items that are not problematic in terms of fit to the Rasch Model. Does this value suggest that the outfit value for this item was perhaps not so bad afterall.
- Please use dataset “for HW 3 and 4.xls”. This is an excel file. Enter this data into the ConstructMap software. Cutting and pasting saves time. Create names for the variables. Pretend this dataset was obtained from a measure in your research area of interest:
- List each of the 10 variables.
- What construct does this hypothetical instrument measure.
- Run a Rasch Model analysis for these data.
- Report the average theta value (you may need to copy the thetas into SPSS).
- Using the Wright Map, which item would you expect only the highest theta
individuals to obtain a 1 over 50% of the time? What else can you do to see if this item is indeed a very hard item?
