Chapter 8 Test Feedback Guidance

The written feedback report has a 1,500 word limit and should address all the elements outlined in the written feedback checklist. In addition, you should append to the written report the completed Test Data Record Sheet, with all calculations, and the sheet showing plotted 68% percentile Confidence Interval (68% percentile CI) for each of the four scores. This appended information does not contribute to the overall word-count. The main body of the report should not need to make reference to the appended information, but is included within the portfolio so that markers can assess the clarity and accuracy of the underlying information and calculations. The assessment of the written feedback report includes:

• The accuracy and clarity of Data Record Sheet and plotted 68% percentile CIs • The extent to which each element of the written feedback checklist has been addressed within the main report • The extent to which the main report provides a professional, informed and useful feedback

Please see the complete set of marking criteria within the unit handbook.

There are some challenges when attempting within the written report to convey effectively the statistical and technical information contained within the Data Record Sheet. For the feedback to be useful, information needs to be conveyed in ways that are accurate and full, but also in a non-technical way as appropriate for the target hearer/reader.

Below are some examples of how to convey some technical aspects within feedback. Within these examples, the phrase ‘fairly confident’ is used to express the meaning of the technical 68% Confidence Intervals. For your oral or written feedback you do not need to address the 95% Confidence Intervals (but these do need to be shown within the Test Data Record Sheet).

Please note that there may be many ways in which the following example extracts could be appropriately expressed. Please consider how these extracts may be improved. For the measure below, an adult general population comparison group is used rather the adult offenders when scoring your test score profile. Please also note that these are only extracts.

On the Dark Triad - Psychopathy, the individual scored 30 out of a possible 35. The score has been compared with the performance of 80 UK Employees who have previously taken this measure. The score of 30 is higher than 85% of norm group.

No tool is perfectly accurate in its measurement and we would always expect some margin of error around a score. This is the extent to which the score underestimates or overestimates the individual’s current in the area measured by this test. Taking this likely margin of error into account, we can still be fairly confident that the individual’s current score on similar psychopathy measures (taken under the same testing conditions) would lie somewhere between 27 [Raw score - SEM] and 33 [Raw score + SEM] out of 35. The lower limit of this range (a score of 27) would place the individual higher than 70% of UK adults, and the upper limit of this range (a score of 31) would place the individual higher than 95% of the UK adults.

Overall, the individual score on the Psychopathy scale can be described as lying within the highest range when compared to Adult Employees.

The individual was also administered the Narcissistic subscale of the Dark Triad.

Because these two measures use the same comparison group (i.e. UK adult employees), the individual’s score on these two measures can be directly compared. To do this the score on each measure has to be converted to a standard scale. The middle-point of this standard scale has a value of 50 and is equal to the middle-point performance of the UK adult employees. On this scale, most people score between 40 and 60. Using this way of scoring the measure, the individual’s standard score on Psychopathy is 65 and the individual’s standard score on Narcissism is 59. Taking into account measurement error on both tests, the 6 point difference between these two standard scores is large enough for us to be fairly confident that there is a reliable difference here. That is, the individual would be likely to attain a similar difference in scores between these two scales on repeated testing.

Please note that the above passage would of course be affected by the particular difference in standard score in comparison to the value of the standard score SEdiff (in the above extract, the scores differ by more than the value of the SEdiff, but the difference is not as large as 1.96*value of the SEdiff)