Background
A detailed introduction to the positional distinctiveness model can be found in Nairne, Neath, Serra, & Byun (1997).
The basic idea was to see whether Estes' (1972) perturbation model could serve as the cause of distinctiveness. An item's position is encoded accurately, but then perturbs (or drifts) in psychological space. The expected values are used as the basis for determining the distinctiveness of each item, relative to the other items, using Murdock's (1960) model.
Instructions
Click once on the "Run the Program" button. You will see a screen that looks like the following (the version number and date that your copy of the program was compiled may be different):
Positional Distinctiveness 0.5 30 March 2004 11:53 EST ineath@purdue.edu Enter the list length [3-19, 5] ?
Your first choice is to select the number of items in the list. For most parameters, you are given a range of permissable values followed by a suggested value. For list length, there must be at least 3 items.
Enter the IPI factor, Si [1-?, 10] ? Enter the RI factor, Sr [1-?, 10] ? Enter the IPI duration [1-?, 1] ? Enter the RI duration [1-?, 2] ? Enter theta [0.0-1.0, 0.05] ? Enter constant, k [0-?, 10] ?
With the suggested settings, you will see a new window with the following:
Positional Distinctiveness 0.5 30 March 2004 11:53 EST Caution: Simulation is accurate enough for demonstration purposes but is *** NOT *** sufficiently accurate for scientific research. List length 5 IPI 1 IPI factor 10.0 RI 2 RI factor 5.0 theta 0.05 konstant 10 Recall Probabilities by Position: 0.4741 0.3127 0.1453 0.0511 0.0168 0.3107 0.3335 0.2230 0.0970 0.0358 0.0945 0.2242 0.3626 0.2242 0.0945 0.0087 0.0507 0.2126 0.4657 0.2623 0.0001 0.0014 0.0202 0.1804 0.7979 The diagonal: 0.4741 0.3335 0.3626 0.4657 0.7979 The expected values: 1.8239 2.2138 3.0000 3.9223 4.7748 The distinctiveness values: 6.6152 5.4456 4.6594 5.5816 8.1391 The predicted values: 0.6615 0.5446 0.4659 0.5582 0.8139 Slope = 0.1740
The slope is the for the last 3 items.
Suggested Simulations
What follows are suggestions for simulations to run and how to run them.
Simulation 1
Plot out predictions for the ratio rule (see Figure 3 of Nairne et al.) going from a 12:1 ratio of IPI to RI to 1:12.
Simulation 2
Plot out predictions for the ratio rule (see Figure 4 of Nairne et al.) when the ratio is constant (e.g., go from a 1:1 ratio of IPI to RI to 12:12.
