Little's MCAR test: mismatch in DF

#1
Hi,
I have a question concerning degree of freedom in little ’s MCAR test.
There seems to be a mismatch in the degree of freedom in the results of MCAR test.
->Little 's MCAR test: Chi-Square = 237,694, DF = 168, Sig. = ,000

However I suspected DF to be 245 because the data consists of 146 individual cases.
Does someone had the same problem or an explanation for this?

Looking forward to your insights.
Cheers,
P

background MCAR test:
GET
FILE='C:\Users\S\Desktop\data analyse\data\FINAL 25-5-13\NEW\BL SELF FINAL (only BL) [Clean]v1.sav'.
DATASET NAME DataSet1 WINDOW=FRONT.
MVA VARIABLES=ID# SELF_BRAND BRAND_AGREE BL_INT1 BL_INT2 BL_INT3 BL_INT4 BL_INT5 BL_INT6 BL_INT7 BL_INT8 BL_PAS1 BL_PAS2 BL_PAS3 BL_PAS4 BL_PAS5 BL_PAS6 BL_PAS7 BL_PAS8 BL_COM1 BL_COM2 BL_COM3 BL_COM4 BL_COM5 BL_COM6 BL_COM7 BL_COM8 SEB1 SEB2 SEB3 SEB4 SEB5_n SEB6 SEB7 SEB8 SEB9 SEB10_n sCOM1 sCOM2 sCOM3 sCOM4_n B_ID HEDv1 HEDv2_n HEDv3 HEDv4_n HEDv5 HEDv6 Gender Age Edu Edu_sty Edu_folow Incom Shop_freq Shop$ Eth
/EM(TOLERANCE=0.001 CONVERGENCE=0.0001 ITERATIONS=25).

Background research:
During data cleaning missing value where found: From the 247 respondents, 10 respondents (1 case: 11 items & 9 cases: 1 item). 1st case with 11 missing items was deleted because it exceeded the acceptable rate of 10 % missing per individual case (Hair, 2006:55), -> N=246.
The remaining 9 cases with missing value’s are ignorable, because the amount of missing per individual case (<10%) and per variable (<15%) are acceptable (Hair,2006:56).
Nevertheless, I wanted some insight in the pattern of the missing value’s in terms of randomness and therefore conducted Little’ s MCAR test.