Findings

Table 1 shows the total number of recommendations that were implemented. Barriers counted in Phase One were completed prior to the priorities interview and were considered high priority barriers by default. Phase Two data includes the additional actions that were completed after the priorities interview and prior to the follow up.

The individuals in Group Two implemented almost twice as many recommendations as those in Group One by the end of Phase Two. This would seem to indicate that technical assistance and referral services were helpful for resolving problems. However, roughly the same ratio held for the number of barriers resolved by the end of Phase One, before any additional services were actually provided. This indicates that the people in Group Two were generally more inclined or more able to change their environments than those in Group One.

phase	1			2			Total
group	1	2	total	1	2	total
# barriers	321	426	747	321	426	747	747
# actions	30	54	84	19	33	52	136
%	4.0	7.2	11.2	2.5	4.4	7.0	18.2
# hipribar	101	188	289	101	188	289	289
hipri %	10.4	18.7	29.1	6.6	11.4	18.0	47.1

Table 1. Frequency of Actions

In addition, the number of recommendations implemented decreased over time. This is contrary to what one would expect if providing technical assistance and referral services are effective as a sole intervention. However, this finding does suggest that assessments on their own have an impact in increasing the rate of modifications. The initial assessment apparently focused participants' attention to problems and encouraged them to act.

Overall, only 18.2% of the barriers identified were resolved. However, of the 747 barriers encountered, 289 were considered a high priority and 47.1% of these were resolved. Thus, older people are willing to devote resources toward improving their home environment if they perceive barriers to be serious but they will not devote resources to low priority problems. What are the reasons for the lack of resolution for over half of the high priority barriers? Answering this question can help us discover ways to increase the rate of problem resolution.

An unanticipated reason for inaction was discovered during the interviews, perceived lack of control over the circumstances. This reflects a model of decision making based on autonomy. The individual wishes to resolve the problem but is unable to do so because others have control. This can be attributed in part to respondents living in rental properties. They were reluctant to approach a landlord or had already had a request refused and were hesitant to press their case. Roughly the same number of homeowners gave autonomy as a reason for not taking action. These people all cited family members who were unresponsive to their requests for assistance. Despite owning the home, they were dependent on others to make improvements. However, these homes had fewer barriers.

model	n	%	revised %
economic constraint	25	31.6	31.6
stress management	19	24.1	24.1
cognitive dissonance	11	13.9
social construction	5	6.3
self concept	16	20.3
autonomy	10	12.7	12.7
risk assessment	321	426	747
cost benefit	2	2.5
deferred priority	9	11.4
Total	79	100	100

Table 2. Frequency of Reasons for Inaction

Table 2 shows the frequency distribution of reasons given for inaction on high priority recommendations. Overall, the barriers that were identified and the proposed solutions were perceived as relevant and important, as indicated by the extremely low number identified as costbenefit, or "not making enough of a difference." While economic constraint was the most frequent reason given for inaction on a recommendation, 68.4% of the barriers were left unresolved for other reasons. Cost is clearly not the only reason why individuals do not make modifications. Stress management was the next most cited reason reflecting another dimension of resource constraints. The frequency of the other reasons were considerably lower than these two. It is possible that our original categories masked the relative importance of different reasons. Upon reflection, we noticed a close affinity between the risk assessment and cost benefit models of decision making. In both cases, the level of perceived benefit can trigger action. Until that benefit is perceived to be significant, either in terms of value or reduced threat, action will be deferred. Likewise, the cognitive dissonance and social construction models are both concerned with selfconcept. In the former, individuals do not act because they deny a change in status and in the latter, because they blame their limitations as the cause of the problem. Inaction, in both cases, can be attributed to an unrealistic assessment of self. Thus, we revised the categories to collapse these four models to two, "deferred priority" and "self concept." Conceived this way economic constraints and stress management are still the top two, but self concept is not far behind.