Relationship of Amount of Treatment and Outcome with Chronicity Modifier

Data Analysis and Program Evaluation – Canfield

There are ___ steps in the following procedure:

1. Subtests are scored and change scores are computed in the CMPSUB1.SPS syntax file.

2. Treatment and chronicity variables are added to the data file.

3. Fictitious treatment and chronicity variables are generated using the CREAT program.

4. Functional treatment variables are computed using the CMPTRT1.SPS syntax file.

5. The relationship between Functional treatment and outcome is computed use the PPOSREG1.SPS syntax file.

Subtests are scored and change are computed:

File Name = cmpsub1.sps

*get file = '\proeval\pashrt.sav'.

missing values jumpy to leisure (9).

compute worthr=8‑worth.

compute calmr=8‑calm.

compute outgor=8‑outgoing.

compute sdepress= mean(sad, hurtself, useless, worthr).

compute sanxious= mean(jumpy, fear, tense, shy, worried, calmr).

compute sborderl= mean(hurtself, angry, trouble).

compute sschiz= mean(halluci, confuse, outgor).

compute stotal= mean(sdepress, sanxious, sborderl, sschiz).

missing values jumpy2 to outgor2 (9).

missing values calmr, calmr2 (9).

compute pdepress= mean(sad2, hurtsel2, useless2, worthr2).

compute panxious= mean(jumpy2, fear2, tense2, shy2, worried2, calmr2).

compute pborderl= mean(hurtsel2, angry2, trouble2).

compute pschiz= mean(halluci2, confuse2, outgor2).

compute ptotal= mean(pdepress, panxious, pborderl, pschiz).

compute cdepress=sdepress‑pdepress.

compute canxious=sanxious‑panxious.

compute cborderl=sborderl‑pborderl.

compute cschiz=sschiz‑pschiz.

compute ctotal=stotal‑ptotal.

execute.

Add the variables TRTTOT, CHRONTOT, TRTDEP etc (see CMPTRT1.SPS syntax) to the data file.

File Name = cmptrt1.sps

compute trtdchnt=trttot/chrntot.

compute trtdchnd=trtdep/chrndep.

compute trtdchna=trtanx/chrnanx.

compute trtdchnb=trtbord/chrnbord.

compute trtdchns=trtschz/chrnschz.

execute.

Run the CREAT program.

Create a prototype regression syntax file using the following method:

1. Put the data file into SPSS.

2. Click Analyze.

3. Click Regression

4. Click Linear

5. Enter CTOTAL into the Dependent Variable box

6. Enter CHRNTOT, TRTTOT, and TRTDCHNT into the Independent Variable box

7. Click Statistics

8. Select the following:

R Squared Change

Descriptives

Part and Partial Correlations

9. Click Continue

10. Click Paste

That will generate the following:

REGRESSION

/DESCRIPTIVES MEAN STDDEV CORR SIG N

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA CHANGE ZPP

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT ctotal

/METHOD=ENTER trttot trtdchnt chrntot .

Fix that up to look like the following syntax file:

File Name = pposreg1.sps

*get file= '\proeval\pashrt1.sav'.

REGRESSION

/DESCRIPTIVES MEAN STDDEV CORR SIG N

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA CHANGE ZPP

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT ctotal

/METHOD=ENTER trtdchnt chrntot trttot .

REGRESSION

/DESCRIPTIVES MEAN STDDEV CORR SIG N

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA CHANGE ZPP

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT canxious

/METHOD=ENTER trtdchna chrnanx trtanx .

REGRESSION

/DESCRIPTIVES MEAN STDDEV CORR SIG N

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA CHANGE ZPP

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT cdepress

/METHOD=ENTER trtdchnd chrndep trtdep .

REGRESSION

/DESCRIPTIVES MEAN STDDEV CORR SIG N

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA CHANGE ZPP

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT cborderl

/METHOD=ENTER trtdchnb chrnbord trtbord .

REGRESSION

/DESCRIPTIVES MEAN STDDEV CORR SIG N

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA CHANGE ZPP

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT cschiz

/METHOD=ENTER trtdchns chrnschz trtschz .

Part of the last run is presented:

Coefficients

		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	Correlations
Model		B	Std. Err	Beta			Zero-order	Partial	Part
1	(Constant)	1.156	.244		4.732	.000
	TRTDCHNS	2.422E-02	.009	.668	2.554	.021	.801	.538	.356
	CHRNSCHZ	-1.592E-02	.016	-.240	-.996	.334	-.748	-.242	-.139
	TRTSCHZ	-1.799E-03	.002	-.138	-.861	.402	.185	-.211	-.120

a Dependent Variable: CSCHIZ

The latest version of SPSS does not seem to give R squared change. The variance of TRTCHNS can be obtained by squaring the Part Correlation (.356). .356 X .356 = .126736 or 13% of the variance of is accounted for in CTOTAL (the change score) by the Functional Amount of Treatment (TRTDCHNS).