The city of Corvallis continues to have the same budget problems that all other levels of government are having and there is no sign that this will change any time soon. I have addressed priorities and choices in previous write-ups and will not do it again here because my positions have not changed. What I will do here is see what I can glean from the latest Corvallis Attitude Survey (CAS) that may be of use to the powers-that-be in establishing priorities.
In recent years, Corvallis used the data collected in the annual National Citizen Surveys, Corvallis (NCS) -- a survey, including Corvallis households, measuring citizens' attitudes and behaviors -- as the CAS. This year it was decided to use another contractor to conduct the survey, the Oregon State University Survey Research Center (OSU-SRC). Their product is entitled: City of Corvallis 2012 Citizen Survey Report, January 2013, Survey Research Center Oregon State University. The reader is encouraged to find the report on the Corvallis web site.
OSU-SRC collected data from 778 people who are registered voters and live in Corvallis. However, all respondents did not answer every question. Therefore, all sets of statistics do not all have the same number of data points. The data, as collected by OSU-SRC is not weighted to reflect individuals' -- with specific gender, racial, and age characteristics - probability of inclusion which left me wondering if weights would make any difference. Therefore I created weights and found that there was very little difference in the results. The creation of the weights and the results computed with the weights is discussed in note one.
The Corvallis powers-that-be need to know how the voters, that usually vote on city issues, think and who they are. The voters - who usually vote on city issues - appear to be more informed about and engaged in city matters than the non-voter. Presumably, the non-voter does not care enough to become informed, let alone vote on city issues. The powers-that-be should give more weight to the positions of the people who care enough to vote on city issues. Moreover, the powers-that-be should tend to fund the services that the people who vote on city issues use.
Older people tend to vote on city issues. They are more likely to use ambulance services more than younger people. Moreover, the people over 61 are more likely to use the senior center more than younger people, which does not support the notion that the senior center is a community center. People over 61 are less likely to use bicycle and multi-use paths than other age groups indicating that spending money on these paths might not be an optimal use of city funds.
People over 35 are likely to use the bus system less than the younger people leading me to ask if the younger people are mostly students. If so, perhaps OSU should be asked for a greater contribution to running the buses instead of taxing the general population through their water bills.
The "old" - over 61 - and the "young" - less than 35 -- are less likely to use the Osborne Aquatic Center than the "middle aged" - 35 to 44. The people in the middle who use the Center the most could probably pay more real user fees; not taxes cleverly disguised as user fees.
One of the questions asked was: Q12. Would you say that you usually vote, or usually do not vote, on city issues?
I computed likelihood ratio chi-squared statistics for voter-nonvoter -- on city issues -- versus the answers to questions one through seventeen printing out only the results that had a p-value of 0.05 or less. The null hypothesis is that the two dimensions are independent of each other. The alternative hypothesis is that the two dimensions are not independent of each other. That printout can be found here. I will discuss the results that I think are the more important to the management of Corvallis where the opinions of the voters tend to be different than the opinions of the non-voters. Readers may wish to peruse the printout themselves and determine which results they think are the more important.
Q1: People who have lived longer in Corvallis tend to vote on city issues while the people living in Corvallis for shorter periods don't - which is not a surprise.
Q3: The voters - who vote on city issues -- tend to think Corvallis is growing too quickly or at about the right pace while the non-voters tend to think Corvallis is growing at about the right pace. Are the voters saying that they don't want change?
Q4-c: The voters tend to think citizen involvement is very important while the non-voters tend to think it is somewhat important.
Q5-n: Voters tend to use the library while non-voters don't.
Q8-d: Voters tend to be more supportive of Sunday library hours than non-voters.
Q8-f: Voters tend to be more supportive of police enforcement services than non-voters.
Q9: All tend to favor a combination of tax increases and budget cuts to balance the budget. However, the non-voters tend to not know what was the best course of action.
Q10-f: Voters tend to think library services were more valuable than non-voters.
Q11-a: One-half of the voters agree that things in Corvallis are headed in the right direction. One-quarter of the voters disagree and the last quarter don't know. Almost one-half of the non-voters don't know and another almost one-half agree that things are headed in the right direction. The voters seem to have definite opinions, the non-voters, not so much.
Q11-b: Almost one-half of both voters and non-voters don't know if the city spends its revenues wisely. A little over two-fifths of the voters and two-thirds of the non-voters don't know. Less than one-third of the voters and a little over one-fifth of the non-voters agree that the city spends its revenues wisely. Another one-fifth of the voters and one-tenth of the non-voters disagree that the city spends its revenues wisely.
Q11-d: Over one-half of the non-voters don't know if the city does a good job informing citizens about city council decisions while over one-half of the voters agree that the city does do a good job of informing citizens about city council decisions. One-fifth of the voters disagree and about one-seventh of the non-voters disagree. More of the voters seem to be aware of what is going on in the city and, in my opinion, the non-voters not being aware is not the fault of the city staff. It could be that the non-voter does not care.
Q16: The older people tend to vote more on city issues than the younger people. Younger people not voting has been addressed elsewhere on this website.
Q17: Ninety percent of the sample is white. The other ten percent is split amongst six categories. However, it does appear that non-whites tend not to vote on Corvallis issues with the exception of Native Americans - who, in this sample, all vote on city issues.
Question 5 asks if the respondent, or someone in their household, has used a particular city service within the past twelve months2. And, if used, the respondent is asked to rate the quality. We have several personal descriptors, gender, age, and race. I thought it would be interesting to see who, by personal descriptor, uses what services and what the users thought of those services. To that end, I ran binary logit regressions of the use/did-not-use decision and multinomial logit regressions of the users quality rating. A printout of the more significant results, where there were not fatal numerical problems, can be found here. The null hypothesis being tested is that the effect of a personal descriptor is zero. The alternate hypothesis being that the effect of a personal descriptor is not zero. I rejected the null hypothesis at 5(plus a little more in some cases)%. Again, I will discuss the results that I think are the more important to the management of Corvallis. And again, readers may wish to peruse the printout themselves.
Q5-a-Use: Native Americans were more likely to use 911 emergency dispatch more than other races.
Q5-a-Qual: People 61 and over were more likely to judge the quality of 911 emergency dispatch excellent rather than good than other age groups.
Q5-b-Use: People 61 and over were more likely to use the ambulance service than other age groups - hardly a surprise.
Q5-c-Use: Females are less likely to use the bicycle and multiuse paths than males. People 61 and over were less likely to use bicycle and multiuse paths than other age groups.
Q5-c-Qual: People 35 to 60 were more likely to judge the bicycle and multiuse paths excellent rather than good.
Q5-d-Use: People 35 and over were more likely to use building inspection services than younger people - which I suspect just means that older people tend to own real estate. Asians are more likely than other races to use building inspection services.
Q5-e-Use: People 61 and over were more likely to use the Chintimini Senior Center than younger people - which is not a surprise. Asians are more likely to use the Senior Center than other races.
Q5-f-Use: People 35 and older were less likely to use the city bus service than younger people. People 61 and older were even less likely to use the city bus than people younger than 60. This surprised me. I would have thought that senior citizens would be more likely to use public transportation to avoid driving.
Q5-f-Qual: People 61 and over were more likely to judge the quality of bus service excellent than good and even more likely to judge the quality excellent than fair.
Q5-g-Use: People 61 and over are less likely to use the parks/trails/open spaces than younger people.
Q5-h-Use: People 35 to 44 are more likely to use city recreation programs than other age groups. Asians are more likely to use city recreation programs than other age groups.
Q5-I-Use: People 45 and over are more likely to use code enforcement services than younger people.
Q5-j-Use: People 35 and over are more likely to use fall leaf collection services more than the younger people.
Q5-j-Qual: People 35 and over are more likely to judge leaf collection excellent rather than fair.
Q5-k-Use: The "other" gender is more likely to use housing assistance programs than males or females. Using weighted data, the "other" race is more likely to use housing assistance compared to all other races. The "other" gender and "other" race is less than 1% of the sample; but these results may indicate there is a problem with discrimination.
Q5-l-Use: People 35 to 44 are more likely to use the aquatic center than those that are 18 to 34. And, people 61 and above are less likely to use the aquatic center than those that are 18 to 34. Asians are more likely to use the aquatic center than other races.
Q5-l-Qual: People 35 to 44 are more likely to judge the aquatic center excellent than fair.
Q5-m-Use: Females are less likely to use police services than the other genders. People 61 and over are less likely to use police services than people under 35. People of mixed race are more likely to use police services than other races.
Q5-n-Use: People 45+ are more likely to use the library than younger people. Asians are more likely to use library services than other races.
Q5-p-Use: People 45 and over are more likely to use utility billing customer service than younger people - which I suspect reflects home ownership. Asians are more likely to use utility customer billing service than other races.
I created weights from the population data from the Population Research Center, March 2012, population estimates by age and sex for Oregon counties as of July 1, 2011 and the Benton County voter file dated September 6, 2012. I computed the probability of gender given age for each age -- eliminating the "other" gender -- and the probability of voter age for voters in Corvallis. I then computed the probability of inclusion for data point for each gender/age category specified in the survey. There were eight different gender/age categories. I used the reciprocal of the probability of inclusion as the weights and calibrated the weights so the sum of all weights equaled the number of data points with methods contained in Fuller (2009) and Thompson (1997).
The likelihood ratio chi-squared statistics for voter-nonvoter -- on city issues -- versus the answers to questions one through seventeen -- printing out only the results that had a p-value of 0.05 or less -- using the calibrated weights can be found here. As the reader can see, weighting did not make much difference.
I also ran binary logit regressions of the use/did-not-use decision and multinomial logit regressions of the users quality rating using the calibrated weights.There was only one result that was useful and it can be found here.
It would have been better if the respondent had asked how many times that particular service had been used within the past twelve month. More could have been done with the data than could be done with the present data.
All computations were done with Gauss V11, a product of Aptech Systems Inc.
City of Corvallis 2012 Citizen Survey Report, January 2013, Survey Research Center Oregon State University
Fuller, Wayne A. Sampling Statistics. John Wiley & Sons. 2009.
Thompson, M.E. Theory of Sample Surveys. Chapman & Hall. 1997.