score2adjust {test2norm}

Convert neuropsychological test scores to demographically adjusted norms.

Description

Convert neuropsychological test scores to demographically adjusted norms.

Usage

score2adjust(
  data = NULL,
  test.score = NULL,
  group.id = NULL,
  control.id = NULL,
  all.controls = FALSE,
  demographics = NULL,
  mfp.alpha = 1,
  rnd.a = TRUE,
  mean.a = 50,
  sd.a = 10
)

Arguments

data

a data frame containing the variables needed for the norming process. The current version of the function does not accomodate missing data. For best results, exclude cases with missing test scores or missing demographics before applying this function.

test.score

a character string specifying the name of the test to be normed, usually the output of the raw2scaled() function.

group.id

a character string specifying the name of the variable containing group identification (i.e. control vs exposed/test/risk). Ignored, if all.controls = TRUE.

control.id

a character string specifying the label of the control group within group.id variable. Ignored, if all.controls = TRUE.

all.controls

a logical indicating whether all observations should be treated as controls. Overwrites group.id and control.id.

demographics

a single or multiple character strings (concatenated by c() function) specifying the names of demographic predictors to be included into normative formulas. Demographic variables should be numeric or binary (0/1).

mfp.alpha

a numeric value between 0 and 1 that sets significance level for inclusion of demographic predictors into normative formula. Passed to the mfp2() function (its select argument). Default value is 1 for inclusion of all predictors regardless of their significance.

rnd.a

a logical indicating whether the adjusted scores (T-scores) should be rounded. Default is TRUE.

mean.a

numeric value for the mean of adjusted score (T-score) distribution. Default is 50.

sd.a

numeric value for the standard deviation of adjusted score (T-score) distribution. Default is 10.

Details

The score2adjust() function can be used by neuropsychologists, who wish to construct normative formulas for cognitive tests that adjust for expected effects of demographic characteristics (e.g., age), using methods described in Heaton et al. (2003 & 2009). The adjusted scores are sometimes referred to as T-scores in the literature. The norming procedure makes use of the mfp2() function from the mfp2 package to explore nonlinear associations between cognition and demographic variables. Detailed description of the procedure are found in Umlauf et al. (2024). (Previous versions of the function depended on mfp package.)

Value

A list consisting of 3 objects. The first two are vectors containing the non-adjusted test scores and the calculated demographically adjusted scores. The last item in the output list is also a list called MFP.formulas. It contains the information for calculation of adjusted scores, including variable transformations (if any), multiple fractional polynomial (MFP) model coefficients, the standard deviation of residuals resulting from the MFP modeling, and a matrix with number of rows equal to the number of predictors and 2 columns containing powers (in numeric form) selected for variable transformations.

Author(s)

Anya Umlauf

References

Umlauf A et al. (2024) Automated procedure for demographic adjustments on cognitive test scores. <doi:10.1080/23279095.2023.2288231>

Heaton RK, Taylor MJ, & Manly J (2003) Demographic effects and use of demographically corrected norms with the WAIS-III and WMS-III. In: Tulsky D et al. (Eds.) Clinical Interpretation of the WAIS-III and WMS-III. San Diego, CA: Academic Press, 183-210.

Heaton RK, Ryan L, & Grant I (2009) Demographic influences and use of demographically corrected norms in neuropsychological assessment. In Grant I & Adams KM (Eds.) Neuropsychological Assessment of Neuropsychiatric and Neuromedical Disorders. New York, NY: Oxford University Press, 127-155.

Benner A (2005) mfp: Multivariable fractional polynomials. R News 5(2): 20–23.

Examples

data(PsychTestData)
PsychTestData$scaledscore <- raw2scaled(data=PsychTestData, test="rawscore",
                                        test.min=0, test.max=36,
                                        test.better="High", group.id="group",
                                        control.id="control")[[2]]
score2adjust(data = PsychTestData, test.score = "scaledscore",
             group.id = "group", control.id = "control",
             demographics = c("age", "male"))