The Assessment of Strategies In

Families - Effectiveness

1998 TESTING AND RESULTS

The instrument was revised after the testing in 1996 (sample 6). The instrument developed at that time had 17 items with acceptable reliability. Thirteen out of the original 15 item tool were included.

Weaknesses of the 17 item instrument were:
1.	System Maintenance and Coherence could not be separated in two factors
2.	Individuation had only two items with internal consistency of .65.

A revised 24-item version was constructed for testing by adding:
	4 individuation items; 1 coherence item; and 2 system maintenance items in the hope to separate out all four dimensions. The minor wording changes suggested from the qualitative study (Friedemann & Smith, 1997) were also included.

The final instrument has 20 items. A total score and sub-scores for the targets of Stability, Growth, Control and Spirituality are available with construct validity and acceptable reliability.

Demographics of the Sample

N = 282

Age:

Range 15-81

X = 36.8

SD = 13.53

Gender:
Males	98	34.75%
Females	180	63.83%
Missing	4	1.42%
	282	100.00%

Family status:
Heads of household	143	50.71%
Other members	135	47.87%
Missing	4	1.42%
	282	100.00%

Education:
Less than 8th grade	8	2.84%
Some high school	10	3.55%
High school grad. or GED	55	19.50%
Education beyond high school	106	37.59%
College graduate	101	35.82%
Other	1	0.35%
Missing	1	0.35%
	282	100.00%

National Origin:
Native Americans	18	6.38%
Far East Asians	25	8.87%
Near East Asians	13	4.61%
African Americans	93	32.98%
Hispanics	17	6.03%
European Americans	109	38.65%
Other	2	0.71%
Missing	5	1.77%
	282	100.00%

Religion:
Protestant	87	30.9%
Catholic	104	36.9%
Christians no denomination	38	13.5%
Jewish	8	2.8%
Muslim	10	3.5%
No religion	23	8.2%
Other (e.g. Hindu, Jehovah’s Witness)	10	3.5%
Missing	2	0.7%
	282	100.0%

Income:
less than $10,000	33	11.70%
$10-25,000	82	29.08%
$26-40,000	70	24.82%
$41-55,000	49	17.38%
$56-70,000	29	7.09%
$71,000+	23	8.16%
Missing	5	1.77%
	282	100.00%

Children in the Household:
With children	61.0%
No children	35.9%
Missing	3.1%

Range 1 to 8 children

X = 1.35 children

SD = 1.43

Adults in Household:

Range 1-7 people

X = 2.19 people

SD = 1.04

Family Structure:
Parents with children up to 12 years	75	26.6%
Parents with teens ages 13 and older (with or without younger children)	55	19.5%
Parents and other adults with children up to 12	19	6.7%
Parents and other adults with teens 13 and older	43	15.2%
Related adults	70	24.8%
Unrelated adults	8	2.8%
Related and unrelated adults	6	2.2%
Missing	6	2.2%
	282	100.0%

Family Type:
One-generation families	73	25.89%
Single parent with children under 18	44	15.60%
Two-parent families with children under 18	84	29.79%
Two-generation families other than single or two-parent families with kids	48	17.02%
Families with more than two generations	30	10.64%
Missing	3	1.06%
	282	100.00%

Tool Completion:
Independent	233	82.63%
With help	27	9.57%
Telephone	21	7.45%
Missing	1	0.35%
	282	100.00%

Test of Items

Two Items (12 and 22) with p-values > .80 were eliminated

Item-to-item correlation within dimension: Items were retained if r = .30 or statistically significant.

Factor Analyses

Initial Factor Analysis was done with 22 items. Result: 8 factors with Eigenvalues 1.0 and greater, explaining 60% or the variance. Items designated for the four dimensions grouped with like items with few exceptions.

Experimented with restriction of factor number to 6, 5 and 4. Four-factor solution had best theoretical congruence. Two additional items needed to be eliminated. They had factor loadings < .40 and their elimination raised the reliability within the subscales.

Final Factor Analysis (Principle Component extraction with Varimax rotation)

Restricted to 4 factors. Explained 45% of the variance.

Cross loadings were less than .40. One item (No. 17) loaded with .35 on its factor. It also failed to correlate significantly with another item on the same dimension. It was left in the scale, however, because its elimination would have decreased the internal consistency of the subscale.

Result of the final factor analysis is represented below. The content of each item is summarized on the left, followed by the dimension to which the item was originally assigned:

SM = System Maintenance

C = Coherence

SC = System Change

I = Individuation

Based on this analysis and the theoretical explanations listed below, some items were reassigned to the dimension expressed by their factor.

Item Description	Factor 1	Factor 2	Factor 3	Factor 4
	SM	C	SC	I
.	.	.	.	.
Agreement about division of labor (SM)	0.71	.	.	.
Celebration of special events (SM)	0.69	.	.	.
Solving problems together (C)	0.64	.	.	.
Caring about each other (C)	0.61	.	.	.
Understanding each other (C)	0.46	.	.	.
======================	========	========	======	======
Feeling about being with family(C)	.	0.63	.	.
Satisfaction with decisions (SM)	.	0.55	.	.
Feelings in the family (C)	.	0.55	.	.
Evaluating family support (C)	.	0.52	.	.
Being helpful/not doing enough (SM)	.	0.48	.	.
Respecting opinions (I)	.	0.47	.	.
Resolution of problems (C)	.	0.42	.	.
======================	=========	========	======	======
Asking for help (SC)	.	.	0.63	.
Independent decisions (I)	.	.	0.63	.
Satisfaction with neighborhood (SC)	.	.	0.61	.
Personal freedom (I)	.	.	0.48	.
Sharing problems with friends (SC)	.	.	0.35	.
======================	=========	========	======	======
Participation in community (I)	.	.	.	0.75
Help out in the community (I)	.	.	.	0.69
Dealing with organizations (SC)	.	.	.	0.48
======================	=========	========	======	======
Internal Consistency (Cronbach’s alpha)	0.7	0.67	0.55	0.58

Theoretical Explanation
Factor 1:	The strongest two items in the factor were meant to belong to system maintenance. Three coherence items also grouped with these system maintenance items. All these items point to some activity or effort needed to achieve coherence. It seemed that the respondents reacted to the activity - which infers system maintenance - rather than the outcome of coherence. The items, therefore, were in no theoretical conflict and were reassigned to system maintenance.
Factor 2:	The majority of the items were coded as coherence. Two items originally assigned to system maintenance items were viewed as coherence by the subjects. These two items alluded to satisfaction with a family activity and therefore could be understood as coherence in that the positive result of the activity fostered a sense of togetherness. One individuation item describing the freedom to express opinions also could be looked at as a family pattern that leads to coherence. All three items were reassigned.
Factor 3:	Three items were originally coded as system change. Two more items were originally assigned to individuation but could be system change: Both have to do with reaching outside the family to make decisions. Connecting with the environment signifies openness toward change in the family.
Factor 4:	Two items were coded individuation. The third item, system change, could be understood as individuation easily since the individuals have relationships with other systems such as organizations and the family promotes these activities.

Factors 3 and 4 are not independent as can be seen from the cross loading of three items (.32 to .38). Theoretically, the factors should be related. In fact pairs of two dimensions express the targets of Stability, Growth, Control and Spirituality. If theory holds true, one should be able to combine the two relevant subscales to arrive at scores for the targets.

Internal Consistency-Reliability (Cronbach’s Alpha)

Subscales:

Stability (System Maintenance and Coherence) .78 (12 items)

Growth (System Change and Individuation) .72 (8 items)

Control (System Maintenance and System Change .69 (12 items)

Spirituality (Coherence and Individuation) .72 (8 items)

Total Scale: .85 (20 items)