Data Collection
SOC Questionnaire
  Survey Data
   Physical Observations
  Crime Data
   Census Data
   Voting Data
   Recycling Data
   Assessor's Data (coming soon)

Data Collection

The data collection method consisted of personal interviews which were conducted by trained students. Interviewer training described interviewing procedures in detail. For example, who is eligible to participate was of critical importance. Only one member of each household could participate and this person must be eighteen or older.  The trained student interviewers were identified with name tags as being part of the Sense of Community Project from MSU Urban Affairs Program. Each student had a format which they followed for introductions.

The data collection process could be divided into two distinct periods. The SOC staff called the first period of data collection Wave 1. The second period is Wave 2. Wave 1 included the original 37 blocks.

Interviews first began in May of 1995 and data collection for the first 37 blocks continued until January of 1996. Data was first collected on 37 blocks which were geographically dispersed throughout the City of Lansing. Due to Michigan's adverse weather conditions, interviewing did not occur between the months of February and May of 1996. This concluded Wave 1. During the Summer of 1996, Wave 2 began and seven more blocks were added to the database with a total of 44 blocks.

The major distinguishing feature between the two waves is the streamlining of the questionnaire. Before continuing, it should be noted that although data collection on block number 38 occurred in Wave 2 interviewing, the questionnaire from Wave 1 was used. The refined questionnaire was used after the 38th block.

During the break in data collection (February - May, 1996), the SOC staff did some preliminary analyses of the questions that were being asked. What was assessed was the relevance of the questions (i.e., were they capturing anything), and the sequencing of the questions. Some question, for example, influenced other questions based upon where they fell in the sequence of questions being asked.

The data collected from the 44 blocks was not designed for an individual study. The data collection itself was for the creation of a database on sense of community indicators in Lansing neighborhoods. This database includes the data from the residents as well as other data from the police department (crime data), the Census, recycling data, physical observations of the block, and open ended questions. Each is discussed in the following sections.

Data Base Menu

SOC Questionnaire

Click Here to View the HTML Version of the Sense of Community Questionnaire

The primary data collection instrument was a two-section questionnaire. The first section asked block particular questions. Most items were on a five point likert-type scale which measured the degree of agreement from strongly agree (or definitely true) to strongly disagree (or definitely not true). It also asked questions which had a yes/no/don't know format as well as some open-ended questions.

The second section of the questionnaire asked question particular to the individual respondent or household. All respondents were over the age of eighteen. This section focused on demographic questions such as ethnicity/race, religion, age, etc. Also, this section contained questions like how much they like living on their block, where they get community information, and the extent to which neighbors interact with each other.
 

Wave 1 versus Wave 2 questionnaires

There are two sets of SOC questionnaires - Wave 1 and Wave 2. On the upper right hand corner, the interviewer's name, the block name and the household number are recorded. The first part of the questions in two questionnaires are basically the same except that there are 33 questions in the first wave but only 24 in the second wave. Five statements in the 2nd wave questionnaires are negative statements. For example, people on this block don't trust each other; residents don't care about the block's future.

Six codes are used to measure SOC and are put at the end of each question. These codes are:
 

Number of SOC Questions by survey wave:
 

Five scales were used as the answers for each question from the first section of the survey.
     The scales were reverse for wave 2:
 

There are 445 cases and 37 streets in the 1st wave; there are 81 cases and 7 streets in the 2nd    wave. One Wave 2 street with 19 respondents used the Wave 1 questionnaire: Kipling Street.

In addition to SOC questions, there are questions about respondents' personal information: age, race, gender, income, educational level, religious affiliation, house ownership, ownership of cars and telephones, number of people in the household within different age ranges (under 5, 5-12, 13-17, 18 or older, 65 or older), ownership of a library card, and if any of the adults in a household regularly attend church, stay at home or work at home during the day.

Other parts of the questionnaire include closed-ended questions and open-ended questions. Close-ended questions ask about information on the block (if there are neighborhood organization, neighborhood watch, community newsletter, a community police officer, and a clear leader on the block, the amount of criminal activity, if people get together to hold meetings and socialize, the number of social activities and community improvement activities in the last year, where people get information about the community), respondents' feelings about the block (whether they are interested in assuming a leadership role and in working with neighbors to improve conditions on the block, whether they like living on the block, how long they plan/like to stay on the block) and respondents' interaction with neighbors (the percentage of households on the block they know by sight/name, whom they socialize with, they consider friends; how often they talk with neighbors, if there are anyone on the block they particularly like or dislike, if they volunteered to help someone on the block without getting compensated, and if they donated blood within the last year).

The following are additional questions added to the 2nd-wave questionnaires because the streets surveyed were racially diverse:

• 1. the racial/ethnic group of most of the people on the block,
• 2. the degree of racial diversity on the block,
• 3. respondents' feelings about the level of ethnic diversity,
• 4. race relations on the block, if they affect the SOC on the block, and how it is affected,
• 5. how the increase of ethnic diversity would impact the SOC on the block,
• 6. if respondents voted in any political elections within the past year and how they rate the job done by their City Council representatives, and
• 7. how frequently each household recycles.

Survey Data

There are two sets of SOC survey data: individual data and aggregate data.

The individual data consists of 526 cases/respondents. At the beginning of each case is the street number, the household number and the code for the interviewer. The codes of five negative statements in the first 24 questions of the 2nd-wave questionnaire were re-coded so that they are consistent with those of the other 19 statements.

For the aggregated data, there are 44 cases (44 streets). The aggregated data were obtained by aggregating individual data by street number. For the first 33 questions, the mean and the percentage of "strongly agree" and "agree" were aggregated at the block level.

For other close-ended questions, the percentage of answering "YES" is represented as __#1 and that of answering "NO" (or negative) as __#2. Sometimes the percentages of all answers to the same question were aggregated; so there are such variables as __#3, __#4, __#5 and so on. (Note: Gender#1 stands for female). For Q35 in the new questionnaire, the racial group with the highest frequency is selected for each block (1= white, 2= black, 3= Hispanic, 4= Asian, 5= Native American, 6= mixed races).

Open-ended questions were included as a part of the survey instrument in the Sense of Community (SOC) Project for more in-depth explanation into each block's sense of community, history, helping activities, reasons for living/buying on the block, social activities, perceptions of other block residents and neighborhood projects. There were 18 open-ended questions included on the survey. Examples of the questions include:

• What is it that you feel contributes most to your sense of community on this block?
• Within the past month, have you provided a service to someone on your block without getting compensated? If yes, what did you do?
• In the last year, what social activities have occurred on this block? (e.g. pot lucks, group garage sales, etc.) [not including private events]
• Why did you buy on this block? or Why would (or wouldn't) you buy a house on this block?
• In the last year, what community improvement activities have occurred on this block? (e.g., community clean-ups, flower plantings, etc.)
• Is there anyone on the block you particularly like? Why?

The total number of codes for all of the open-ended questions was seventy-eight (78). Codes were created to be used for certain questions:

• 38 codes were available for all open-ended questions,
• Question 5 had four codes specific to it,
• Question 6 and 7 had six codes specific to it,
• Question 8 had eight codes specific to it,
• Question 9 had four codes specific to it, and
• Questions 14 and 15 had twelve codes specific to them.

• [physic_p] is the Physical conditions/Appearance of Block, and
• [physic_n] is the Negative Physical Appearance.

After each questions were double coded and a master copy of each coded question was made, a matrix was designed to enter the data into Statistical Package for Social Sciences (SPSS). (In order to simplify data entry, Questions 9 and 12 were combined because of their similar nature and measurements.) All the data from Wave 1 has been entered in SPSS and SOC is beginning to run statistics to examine what the open-ended questions tell SOC about each blocks' sense of community. The SOC staff intends to use this information to compare SOC variables across blocks.

Response Rate

The response rate for each street was obtained using the following formula: DONE / (TOTAL - VACANT). DONE is the number of residents on the block who agreed to participate in the survey. TOTAL is the number of households on the block. VACANT (not shown in the table) is the number of unoccupied houses on the block.

This table demonstrates the response rates per block
 

 Data Base Menu
 

Physical Observations

Physical observations were taken on every one of our blocks. Observations were collected using an "Observational Rating Form." SOC staff hypothesized that physical objects such as fences or hedges could act as a barrier for the interaction of neighbors. This would have a negative affect on the overall sense of community. Other things, such as porches or mailboxes on the road, were believed to help facilitate a higher sense of community because it would be a good place for neighbors to meet/interact with each other.

Other things which were looked at included the physical surroundings around the neighborhood. For example, were there neighborhood watch stickers posted? This could act as a deterrent to would-be thieves. It's also a measure of the sense of safety on the block and a sense that people could be taking ownership for their block by collectively establishing a neighborhood watch. These two issues - safety and empowerment - are both variables which were looked at in the 33 sense of community variables.

The list of physical observations which were used reflected various possible indicators that could increase the interaction between residents on the blocks. The last set of observations included the categories of structure (i.e., residential, business, or mixed) or housing (single, duplexes, etc.). The following is a list of all the physical categories which were observed.

This list is broken into three areas. The first area rates the individual house. The second area rates the entire block and its surroundings. The third area consisted of two variables describing the type of structure (residential, commercial, or a mixture of the two), and the type of housing (single family, duplex, multiple family or mixed housing) on each block. Most of our blocks consisted of single family homes.

The following are the different ratings found on the observation form.

Rating For Each House:

• Porches
• Windows that face the street
• Toys/bikes
• Sidewalks
• Shared driveways
• Recycling bins
• Blinds drawn
• Fences or hedges
• Garage doors open
• Mailboxes on the road

Two SOC persons went out to collect the physical data on each of the block and recorded their observations on the rating form. The coding was very simple, if there was the existence of one of the categorical variables (porches, for example) then it would be recorded as a "yes." If the category was not present, it would be recorded a "no." When the results were brought back to the SOC Project Office, the percentage of "yes" was calculated on every block. These were then entered into an observational data base to be used in the future.
 

Data Base Menu
 
 

The crime data for 37 streets in Lansing were collected in the late Fall of 1995. The data were retrieved from the main frame of the Police Department in Lansing, using a departmental computer program. In order to retrieve the data, specific address levels for each block were used. Two sets of data were obtained. The time span of the first one was from October 1st, 1989 until September 30th, 1995.   The second set started from October 1st, 1994 and ended on September 30th, 1995.

Since the data were in D-Base format, they were imported to Excel and saved as 37 Excel files. Later, the total of each crime for each street was computed and entered in SPSS format.

There are about 100 codes for the crimes in the list provided by the Police Department. However, only 25 crimes were entered because others have zero totals. The variable names begin with "C." At the end of each case is the total of the crimes for each block.

In the late fall of 1996, the crime data for seven new streets were collected. After entering the total of crime for new streets in SPSS format, the old streets and new streets were merged into one file.

Larceny is the most common crime (350 cases in six years on 44 streets) , followed by assault (283 cases) and malicious destruction of property (236 cases). Burglary (153 cases) and riot situations (114 cases) are more frequent than juvenile delinquency (71 cases), violation of controlled substance (54 cases), and motor vehicle theft (52 cases).
 

Data Base Menu
 

Census Data

The socio-economic data for the face blocks were secured from the latest census, the 1990 U. S. Census of Population and Housing.  The census data was used at the block level. The Census Bureau (1991) defines a block as follows:

        Census blocks are small areas bounded on all sides by visible features such as
        streets,  roads, streams, and railroad tracks, and by invisible boundaries such
        as city, town, township, and county limits, property lines, and short, imaginary
        extensions of streets and roads (A-3).

As one can see from the definition of a census block, the Census Bureau does not measure the face block as a unit of analysis. Due to this limitation, the census statistics for the two census blocks belonging to the face block were collected.

The sums or averages of the two facing census blocks were used to estimate the socio-economic characteristics of each face block. Means were calculated for the mean contract rent and mean housing values. Other items are the sums of the two statistics from the facing census blocks.

Variable items available for the census block level

• Total Number of persons in the block
• Number of White
• Number of Black
• Number of American Indian, Eskimo, or Aleut
• Number of Asian or Pacific Islander
• Number of persons under 18 years
• Number of persons 65 years and over
• Number of total housing units
• Number of 1 unit detached or attached
• Number of 10 or more units
• Mean number of rooms
• Number of owner occupied housing units
• Mean value of owner occupied housing units
• Mean contract rent of renter occupied housing units
• Number of total occupied housing (housing units with 1.01 or more persons per room)
• Number of rental occupied housing units
• Number of one person households
• Number of householder, no spouse present with 1 or more persons under 18 years

Data Base Menu
 

Recycling Data

The increasing solid waste problem has become one of the major issues that confront communities, policy makers, and environmental groups in the last few decades.

One of the potential solutions to the waste problem is recycling (Porter et al., 1995). Recycling has good potential for success because it is endorsed by a large majority of people (Belsie, 1990). A body of scholarly literature on the development of recycling and citizen participation exists. Most of the studies attempt to address ways and strategies to increase recycling and develop pro-environment behavior. Researchers have explored extrinsic (i.e., behavior-change / antecedent-consequences) and intrinsic (i.e., attitude, values) approaches in motivating and increasing recycling.

An area in recycling studies that is not adequately documented so far, is the role of the "sense of community" that exists among residents in a particular locale. Sense of community has been related to positive social outcomes, such as neighboring and community participation (Chavis & Wandersman, 1990; Unger & Wandersman, 1986). It has also been related to crime prevention, drug abuse prevention, and mental health (Chavis & Wandersman, 1990). The implications of sense of community in promoting collective action provided the impetus for undertaking this study. This study explores the relationship between sense of community and recycling participation in the context of 36 street neighborhoods in Lansing, Michigan. Three hypotheses were tested in this study.

• The stronger the sense of community on the street neighborhood, the higher is the recycling participation;
• Demographic variables (income, home ownership, and education) are positively associated with recycling participation; and
• Sense of Community predicts recycling participation better than the demographic variables

• To understand the relationship between sense of community and recycling participation;
• To examine the relationship between demographic variables (income, education, and home ownership) and recycling participation; and
• To discuss implications of this study in increasing recycling and promoting pro-environment behavior.

In the survey, residents were asked whether people on their street neighborhood participated in the recycling program, thereby generating the data on self-reported recycling participation at the street neighborhood level. It should be noted that residents did not provide individual recycling information, but only block level perceptions.

A six week data collection was also conducted during the months of April and May 1996 to determine the street neighborhood level actual recycling participation. Houses on each neighborhood that have recycling bins filled with recyclable materials were recorded during each of the scheduled curbside recycling days.

Using the street neighborhood as the unit of analysis, self-reported and actual recycling participation rates were correlated with the sense of community measures and the demographic variables. And, multiple regression analysis (stepwise method) was used in testing the hypotheses derived for this study.

A total of 441 interviews from the 36 street neighborhoods in Lansing, Michigan, were conducted, resulting in a participation rate of 62%. About two thirds or 65% of the respondents were female. Most (72%) of the respondents were home owners; the average age of respondents was 43; average income or financial condition of the block that the respondents specifically identified with was "middle income on the lower side;" and the average level of educational attainment reported was 1-2 years in college or completion of some sort of skill training.

Table 1 shows that sense of community, particularly people's feeling of belonging to the street neighborhood is associated with both self-reported and actual recycling participation. Also, income and home ownership significantly correlated with both self-reported and actual recycling participation. Education did not indicate statistical significance for either self-reported or actual recycling participation.

Table 2 shows the results of the multiple regression analyses showing the percent of variance in recycling participation accounted for by belonging and income. Belonging and income significantly predicted both self-reported and actual recycling participation. Using the stepwise procedure, it was determined that connection and support did not account for any significant variance beyond the variable of belonging. Among the demographic variables, income was the only variable that accounted for any significant variance in recycling participation.
 
 

In order to compare the relative strengths of the contributions of belonging and income, the multiple regression analyses were performed twice. In one regression, we entered belonging first, and the second analysis, we entered income first. When belonging was entered first, income did not add any significant proportion of variance in predicting either self-reported or actual recycling participation. When income was entered first, belonging added a significant amount of variance above the variance accounted for by income in predicting self-reported recycling participation. Belonging did not contribute significant additional variance to predicting actual recycling participation.

This study showed that sense of community and income relate to and predict recycling participation at the street neighborhood level. The implications of this study rest on its practical use in promoting recycling and other forms of civic behavior. Further, policy makers, program planners, and implementers will find this study useful in any community development efforts that they are undertaking.
 

Data Base Menu
 

Voting Data

The voting data were provided by a political consultant in East Lansing. There are two sets of voting data: One was purchased in the fall of 1995 and the other was purchased in the spring of 1997. Both data set consist of voting information on general election, primary election, presidential primary and city election , school election and tax proposal.

The years of election in the two data sets are as follows:
 

The two data sets are in D-Base format. To retrieve the streets needed, the data were transformed to SPSS format and string variables were changed to numeric variables. However, the values for D (democrat), R (republican), and I (independent voters) were entered as D, R, and I. So, they had to be changed to "1" first. Otherwise, all the letters would disappear when the variable type was changed from "string" to "numeric."

Because of the huge size of the data (45000 cases), census tracts were first used to reduce the cases. Then plus-4s of each street (from two to six) were used to select the cases for 44 streets. It is very common that there are more than one voter in one household. After processing the data, two data sets consisting of individual voters were completed. Some variables not to be used for statistical analysis were dropped.

In order to get the information of voting behaviors at the block level, the data were aggregated by the street number. In the first aggregated data, g94 stands for the percentage of voters on each block voting in the general election of the year 1994. In the second aggregated data, sch95 stands for the percentage of voters on each block voting in the school election of the year 1995.

The values of the following variables were computed and put at the end of each case:

• vote6all = (vote6reg * n_indiv) / adul_blk . "Vote6all" stands for the average % of all adults on the block who voted in 6 elections.
• vote6reg = (g.94 + p.94 + pp.92 + city.93 + sch.any + tax9495)/6. "Votereg" stands for the average percent of 6 voting on each block.
• adultreg = n_indiv / adul_blk . "Adultreg" stands for the mean of adults on each block voting in 6 elections .

Data Base Menu