1995 Vermonter Poll - Introduction and Methodology

January 1996

Introduction:

The 1995 Vermonter Poll is a unique opportunity for researchers, policy makers, social advocates, and citizens to examine changes in public opinion regarding contemporary, and often controversial, issues in the public arena. Since 1990, the Center for Rural Studies at the University of Vermont has conducted four, state-wide polls of registered Vermonters (1990, 1993, 1994, and 1995). Results of these studies help document evolving public attitudes toward a variety of concerns with implications for public policy, programming development, and the allocation of shrinking public resources. Specific issues addressed in the Vermonter Poll Series include: favorability toward political figures; agriculture and the environment; community growth and economic development; satisfaction with community services; property tax reform; and the use of computers and communications technology. Additional questions are added to, or deleted from, the Vermonter Poll as issues become more or less timely.

The first report of the 1995 polling results focuses specifically on Vermonters' support of the "Contract with America" and favorability toward a number of state and national political figures. Political figure favorability responses are first compared to Vermonter's responses to similar questions in the 1993 Vermonter Opinion Poll. Additional comparisons are made according to whether respondents live in Chittenden County or Vermont's Northeast Kingdom. Finally, responses are analyzed according to selected respondent demographics including age, education, income, and gender.

Additional reports will be produced as analysis of the Vermonter Poll data progresses and issues become more timely. Future reports will focus on bovine somatotropin, the quality of Vermont's natural environment, community growth and development, public transportation, and electronic communications--computer use and Internet access.

Methodology

The 1995 Vermonter Poll was conducted and sponsored by the Center for Rural Studies, College of Agriculture and Life Sciences, University of Vermont. This is the fourth in a series of state wide polls assessing the attitudes and orientations of voting Vermonters. This survey was conducted from December 4th to 7th, 1995, during the evening hours of 4 p.m. and 9 p.m.. Vermont households were contacted at random and survey respondents were selected randomly within participating households.

The 1995 survey was expanded to include specific survey samples for Chittenden County and the Northeast Kingdom (Caledonia, Essex, and Orleans Counties). While 400 surveys initially were needed to permit a statewide analysis of sufficient statistical confidence and precision (95%, +/-5%), both Chittenden County and Vermont's Northeast Kingdom (Essex, Orleans, and Caledonia Counties) were over-sampled to allow for separate and comparative analyses of these distinct regions. In total, 702 registered Vermont voters completed the survey. To compensate for over sampling in certain parts of the state, different weights were assigned to data depending on the analysis required (See Table A). Weights were determined using the proportion of Vermont's total households found in these regions during the 1990 U.S. Census of Population and Housing. The results of the weighting process yielded a representative statewide sample population of 705 cases.

Table A.	Survey weights.
						Assigned Weights
Region		Households 	Total 		Statewide 	Chittenden	NEK 
		in 1990*	Responses	Analysis	County		Comparison
				(N=702)				Comparison

Chittenden 	48,439		281		0.573016	1.000000	0.531515
County	

Northeast 	21,855		118		0.632668	0.488113	1.000000
Kingdom

Remaining 
10 Counties	140,356		303		1.549295	1.195300	1.195300

Confidence/Precision				95%, +/-5%	90%, +/-5%	90%, +/-5%

* Source: U.S. Census of Population and Housing, Summary Tape File 1A: Vermont.

A conservative estimate of the margin of error associated with the statewide analysis is +/- 5.0% with a confidence interval of 95.0%. This means that if the survey were to be redone, 95 percent of the time the results would be the same within +/-5.0 percentage points for each particular question. The margin of error for the Chittenden County and Northeast Kingdom analyses is +/- 5.0% with a confidence interval of 90.0%. The margin of error associated with any given item in the survey increases as the sample size for individual questions are examined in greater detail such as in cross-tabulations.