Calculating norm-based T-scores from your FACT-G data
What are norm-based T-scores?
T-scores are standardised scores. A score of 50 represents the mean. A difference of 10 indicates a difference of one standard deviation. In the calculator below, these means and standard deviations refer to scores reported for the Queensland general population. This means that a cancer sample with a mean T-score of 40 is scoring one standard deviation below what might be expected for someone from the general population of Queensland.
Why are norm-based T-scores useful?
Compared with raw scores, norm-based T-scores enable a more direct comparison between different subscales. Raw score distributions vary between subscales, while norm-based T-scores on the various subscales share a common referent in the form of both location (mean) and spread (SD). The mean of 50 represents the mean score of the general population. The mean T-score of your sample for each subscale can therefore be interpreted in the same way, that is, terms of how many standard deviation units it is away from 50. This enables you to tell at a glance which scales have been most impacted by disease or treatment.
Where does the normative data come from?
The normative data used in this calculator comes from the Queensland Cancer Risk Study, which was funded and conducted by the Cancer Council Queensland and collected FACT-G data from 2,727 Queenslanders in 2004. A summary of the sample characteristics can be viewed
HERE. These data have been provided by the Cancer Council Queensland and have been published by Janda M
et al. Psycho-Oncology 2009;18:606-14.
To make these normative data more representative of the Australian general population, PoCoG hopes to supplement the current dataset with FACT-G data from other states and territories. If you own FACT-G data from the Australian general population and would like to have it contribute to the calculator, please contact us at
pocogqoloffice@psych.usyd.edu.au
How do I use the calculator?
The calculator below is designed to provide norm-based T-scores based on group mean raw scores on the FACT-G. To use the calculator, simply:
- Select the gender/age characteristics that best match those of your sample.
- Enter your sample mean scores for each of the FACT-G’s subscales and its overall score. When calculating mean raw scores, please remember that scores should only be included from participants for whom at least 50% data is available for the subscale in question. Where missing data amounts to less than 50%, subscales can be prorated by multiplying the sum of available items by the number of items in the subscale, then dividing by the number of items actually answered. The FACT-G score is the sum of all items and should only be calculated where more than 80% of items are available.
- Click the ‘Calculate’ button to view norm-based T-scores for each subscale and overall FACT-G scale. A ‘View Graph’ option will then become available.
Syntax
Using mean scores rather than whole datasets inevitably limits the functionality of the norm-based T-score calculator. Click
HERE for a text file containing SPSS syntax that will compute a 95% confidence interval around each T-score estimate and run a one-sample t-test comparing the sample mean to 50. To use the syntax, open SPSS, go to the
File Menu, go to "
New", and select "
Syntax". Copy and paste the contents of the text file and select "
Run".
Acknowledgements
PoCoG would like to thank the following people for allowing us to use data from the Queensland Cancer Risk Study: Dr Monika Janda; principal investigator Assoc Prof Joanne Aitken; chief investigators Prof Beth Newman, Prof David Whiteman, and Assoc Prof Liz Eakin; and steering committee members Prof Ian Frazer, Prof Ross Young, Dr Brian Cole and Prof Jeff Dunn.