Evaluation of the internal construct validity of the Personal Care Participation Assessment and Resource Tool (PC-PART) using Rasch analysis

Darzins, Susan; Imms, Christine; Di Stefano, Marilyn; Taylor, Nicholas F; Pallant, Julie F

doi:10.1186/s12913-014-0543-z

Table 2 Model fit for the three-staged Rasch analysis (n = 958 in each analysis)

From: Evaluation of the internal construct validity of the Personal Care Participation Assessment and Resource Tool (PC-PART) using Rasch analysis

Subscale	No. of items	Item response categories	Overall model fit^a	Overall Item fit residual (SD)^b	Overall Person fit residual (SD)^b	No. of items with disordered thresholds	No of misfitting items^c	No of misfitting persons^c	PSI^d	Items with uniform/non-uniform DIF	No. of item-pairs with response dependency^e	Dimensionality
										(a) by age		(% of significant t-tests)^f
										(b) by sex		(% of significant t-tests)^f
Stage 1. All PC-PART items as one scale
(a) All original items	41	012	X² = 2003.85, df (369), p < .001, (α = .001)	−0.07 (2.84)	−0.46 (1.23)	27	9 (D6 C1 E6 C5 E1 F4 E2 D7 C7)	21	0.93	(a) B2 D6 E2/ –	48	11.5%
(a) All original items	41	012	X² = 2003.85, df (369), p < .001, (α = .001)	−0.07 (2.84)	−0.46 (1.23)	27	9 (D6 C1 E6 C5 E1 F4 E2 D7 C7)	21	0.93	(b) A5 B6 B7 C3 D9/ –	48	(95% CI 10.1-12.9%)
(b) All original items	41	001	X² = 1743.19, df (369), p < .001, (α = .001)	-0.47 (2.14)	-0.54 (1.03)	NA^g	3 (D6 D7 C1)	11	0.91	(a) B2 D6 E2/ –	39	7.1%
(b) All original items	41	001	X² = 1743.19, df (369), p < .001, (α = .001)	-0.47 (2.14)	-0.54 (1.03)	NA^g	3 (D6 D7 C1)	11	0.91	(b) A5 B6 B7 C4/ E3	39	(95% CI 5.7-8.5%)
(c) Removal of 6 items	35	001	X² = 693.50, df (234), p < .001, (α = .001)	-0.46 (1.80)	-0.52 (0.87)	NA	0	6	0.88	(a) B2 E2/ –	5	7.52%
(c) Removal of 6 items	35	001	X² = 693.50, df (234), p < .001, (α = .001)	-0.46 (1.80)	-0.52 (0.87)	NA	0	6	0.88	(b) C4/ –	5	(95% CI 6.1-8.9%)
Stage 2. Original PC-PART domains
Clothing	5	001	X² = 316.31, df (15), p < .001, (α = .01)	-0.81, (4.23)	-0.34, (0.60)	NA	1 (A5)	0	0.54	(a) –/ –	1	–^h
Clothing	5	001	X² = 316.31, df (15), p < .001, (α = .01)	-0.81, (4.23)	-0.34, (0.60)	NA	1 (A5)	0	0.54	(b) A5/ –	1
Hygiene	8	001	X² = 384.92, df (39), p < .001, (α = .006)	-1.03 (1.99)	-0.75 (0.86)	NA	0	0	0.68	(a) B2/ –	1	1.98%
Hygiene	8	001	X² = 384.92, df (39), p < .001, (α = .006)	-1.03 (1.99)	-0.75 (0.86)	NA	0	0	0.68	(b) –/ B2 B6	1	1.98%
Mobility	9	001	X² = 539.17, df (54), p < .001, (α = .006)	-1.51 (3.38)	-0.42 (0.48)	NA	2 (D6 D7)	0	0.68	(a) D3 D6/ –	2	1.67%
Mobility	9	001	X² = 539.17, df (54), p < .001, (α = .006)	-1.51 (3.38)	-0.42 (0.48)	NA	2 (D6 D7)	0	0.68	(b) –/ –	2	1.67%
Nutrition	8	001	X² = 413.10, df (48), p < .001, (α = .006)	-1.04 (3.05)	-0.53 (0.71)	NA	2 (C1 C5)	0	0.49	(a) –/ C6	0	4.92%
Nutrition	8	001	X² = 413.10, df (48), p < .001, (α = .006)	-1.04 (3.05)	-0.53 (0.71)	NA	2 (C1 C5)	0	0.49	(b) C3/ –	0	4.92%
Residence	5	001	X² = 189.26, df (5), p < .001, (α = .01)	-3.18 (2.9)	-0.59 (0.50)	NA	0	0	-0.31	(a) –/ –	0	–^h
Residence	5	001	X² = 189.26, df (5), p < .001, (α = .01)	-3.18 (2.9)	-0.59 (0.50)	NA	0	0	-0.31	(b) –/ F3	0	–^h
Safety	6	001	X² = 109.40, df (23), p < .001, (α = .008)	-0.44 (2.62)	-0.25 (0.48)	NA	0	0	-0.46	(a) E2/ –	0	0.00%
Safety	6	001	X² = 109.40, df (23), p < .001, (α = .008)	-0.44 (2.62)	-0.25 (0.48)	NA	0	0	-0.46	(b) –/ –	0	0.00%
Stage 3. PC-PART items separated into ‘Self Care’ and ‘Domestic Life’ scales
(a) All self care items	23	001	X² = 987.96, df (184), p < .001, (α = .002)	-0.64 (2.33)	-0.51 (0.77)	NA	3 (C1 C2 B2)	2	0.87	(a) B2/ –	11	14.61%
(a) All self care items	23	001	X² = 987.96, df (184), p < .001, (α = .002)	-0.64 (2.33)	-0.51 (0.77)	NA	3 (C1 C2 B2)	2	0.87	(b) B6/ –	11	(95% CI 13.2%-16.0%)
All domestic life items	18	001	X² = 1058.02, df (162), p < .001, (α = .002)	-0.77 (2.48)	-0.50 (0.81)	NA	2 (D6 E6)	0	0.79	(a) E2/ –	7	8.56%
All domestic life items	18	001	X² = 1058.02, df (162), p < .001, (α = .002)	-0.77 (2.48)	-0.50 (0.81)	NA	2 (D6 E6)	0	0.79	(b) A5 C4/ –	7	(95% CI 7.2%-9.9%)
(b) Refined Self Care scale	16	001	X² = 360.64, df (91), p < .001, (α = .003)	-0.86 (1.87)	-0.46 (0.62)	NA	0	0	0.85	(a) –/ –	0	4.18%
(b) Refined Self Care scale	16	001	X² = 360.64, df (91), p < .001, (α = .003)	-0.86 (1.87)	-0.46 (0.62)	NA	0	0	0.85	(b) –/ –	0	4.18%
Refined domestic life scale	14	001	X² = 515.48, df (91), p < .001, (α = .004)	-0.57 (2.02)	-0.45 (0.63)	NA	0	0	0.76	(a) E2/ –	0	6.16%
Refined domestic life scale	14	001	X² = 515.48, df (91), p < .001, (α = .004)	-0.57 (2.02)	-0.45 (0.63)	NA	0	0	0.76	(b) A5 C4/ –	0	(95% CI 4.8-7.5%)

SD = Standard Deviation, PSI = Person Separation Index, DIF = Differential Item Functioning, X² = chi square, df = degrees of freedom, p = probability value, α = Bonferroni adjusted alpha level, CI = Confidence Interval.
^anon significant chi-square item-trait interaction statistic is evidence of overall model fit.
^bItem-Person Fit Residual SD ≤1.5 is evidence of overall item/person fit.
^cIndividual item or person Fit Residual values of ≤2.5 are evidence of item/person fit.
^dPSI values ≥0.7 acceptable for use with grouped data, values ≥0.8 acceptable for use with individual data.
^eItem pairs with Residual correlation values of r ≥0.2 deemed to have local item dependency.
^fValues ≤5% considered evidence of a unidimensional scale (95% CI only presented when the proportion of significant t-tests exceeded 5%).
^gNot Applicable - only one threshold for items with two response categories.
^hInsufficient items and response categories to produce minimum scores for meaningful results.

Back to article page

ISSN: 1472-6963

Contact us

General enquiries: journalsubmissions@springernature.com

BMC Health Services Research

Contact us