Volume 2, Issue 2, June 2000
ISSN 1096-4886 http://www.westerncriminology.org/Western_Criminology_Review.htm
© 2000, The Western Criminology Review. All Rights Reserved.
|
Volume 2, Issue 2, June 2000 ISSN 1096-4886 http://www.westerncriminology.org/Western_Criminology_Review.htm |
Citation: Williams, Frank P., Marilyn D. McShane, and Ronald L. Akers. 2000. "Worry About Victimization: An Alternative and Reliable Measure for Fear of Crime." Western Criminology Review 2(2). [Online]. Available: http://www.westerncriminology.org/documents/WCR/v02n2/williams/williams.html
The concept of fear of crime has generated substantial interest within criminology. While discussions of what fear of crime means, and what dimensions are represented by fear of crime, have become commonplace, actual measurement issues have only recently gained attention. This article questions the utility of fear as a concept and argues for "worry" about crime as a more reasonable substitute. We convert various commonly-used measures of fear to worry and compare their dimensions and reliability. A victimization-based worry scale is developed as part of this comparison and generates the most crime-focused measurement, as well as the highest reliability measures. The scale may have greater utility for measuring specific concerns about crime than existing approaches using fear of crime.
Keywords: Fear of crime, worry about crime, victimization, measurement
In just the last twenty years, criminologists have come to recognize that victimization is a major dependent variable in their field. As elementary relationships between victims and assorted independent variables became articulated, the interest of some researchers turned toward the victim proper. Soon it became clear there are two kinds of victims. First there are the traditional victims of crime--those who suffer from an assault, robbery, theft, and so on. Second, there are those victimized by the fear of crime regardless of whether or not they have personally experienced a crime. This form of crime victimization demonstrates a potential for greater harm than traditional victimization (grievous bodily harm aside) because of the effect of long-term stress coupled with changes in behavior that affect quality of life.
The discovery of fear of crime resulted in the addition of another potential dependent variable for criminology. In addition to its obvious interest for criminologists, the study of fear of crime has significant implications for political, economic, and social policy. As Reiss (1983:56) has noted, crime and fear of crime "coerces resource allocation" in all areas of society. Fear of crime has also become a mainstay of political campaigns, as evidenced by the infamous Willie Horton television advertisements during the 1988 presidential election campaign.
As soon as researchers became aware of fear of crime, they focused on a search for related independent variables. Several important relationships were located, including the usual sociological staples of sex, race, age, and social class. Yet on the whole, and as with examinations of crime and delinquency, concern over the measurement of fear of crime itself took a back seat to more interesting independent variables. Several approaches to measuring fear of crime were used without any real determination of exactly what was being measured and with little concern for the methodological adequacy of the measuring instruments. Recent work on fear of crime has been more sensitive to these issues and some researchers are now focusing on the measurement issues. In the face of this, however, most fear of crime research continues to use measures allied to the traditional National Crime Survey "walking in the neighborhood" questions.
RESEARCH ON FEAR OF CRIME
In general, much of the early fear of crime research is descriptive, atheoretical (Clarke and Lewis 1982; van der Wurff, Van Staalduinen, and Stringer 1989), and commonly relies on a posteriori constructions of personal characteristics. In trying to become explanatory, the search for "causes" of fear of crime correlated an increasingly complex set of variables. These range from traditional demographics to intricate combinations of psychological, social, economic, and geographic factors.
What has evolved after hundreds of studies is an overemphasis on independent variables as specific and fully developed concepts, while the dependent variable remains less sophisticated and much more ambiguous. It has even been the case, as van der Wurff et al. (1989) complain, that actual components of fear of crime have been misused as predictors of fear of crime and independent and dependent variables have been arbitrarily transposed. For instance, Garofalo (1979) uses perceptions of dangerousness of one's residential environment and Garofalo and Laub (1978) use "crime as a serious problem" as predictors of fear of crime.
The research on fear of crime has primarily dealt with attitudes and behavioral responses (Krahn and Kennedy 1985), although environmental concerns are attracting attention as of late. Studies show great discrepancies in the factors associated with fear of crime and the extent to which they shape perceptions. Much of this research deals not with general populations but specific groups--blacks, women, elderly, the homeless--and their fear of crime. Gender (Gilchrist et al. 1998; Clemente and Kleiman 1976; Markson and Hess 1980), age (Lee, 1982a, 1982b,1983; Lebowitz 1975; Warr 1984), neighborhood racial composition (Vander Ven 1998; Chiricos, Hogan, and Gertz 1997b; Moeller 1989), amount of contact with strangers (Kennedy and Krahn 1984), previous victimizations (Garofalo 1979; Gubrium 1974; Hindelang, Gottfredson, and Garafalo 1978; Parker and Ray 1990; Maguire 1980; Fitzpatrick, LaGory, and Ritchey 1993) and capacity to resist victimization (Riger, Gordon, and Le Bailly 1978), as well as an assortment of other variables, have all been associated with differing rates of fear. Other studies link rapidly growing communities or "boom towns" with heightened levels of crime fear (Freudenburg 1986; Krannich, Grider, and Little 1985, 1989) as well as the role of the media (Kuttschreuter and Wiegman 1998; Chiricos et al. 1997a; Williams and Dickinson 1993; Liska and Baccaglini 1990; Winkel and Vrij 1990). Research also looks at how police crime prevention strategies impact citizen fears (Vogel and Torres 1998; Bennett 1989).
The literature also produces information on factors that do not appear to be related to fear of crime, such as police force size (Krahn and Kennedy 1985), education, income, and social status (Toseland 1982). Even these general findings, however, are not without dispute as some of the basic relationships have not been consistently replicated across studies. Further, work has found variables that interact to produce different relationships with fear of crime than those previously attributed to the individual variables themselves (gender and age--Normoyle and Lavrakas [1984]; Lee [1982a]; age and economics--Baldassare [1986]; race and neighborhood racial composition--Moeller [1989]; age, race, and gender--Ortega and Myles [1987]).
Theoretical Problems
A major problem in fear of crime research is the number of different operationalizations of fear of crime (Warr 1984; Dubow et al. 1979; Ferraro and LaGrange 1987). Earlier research identifies three "forms" of fear of crime (Garofalo 1981; Maxfield 1984; Garofalo and Laub 1978; Taylor and Hale 1986). These are characterized as a fear of violent crime, a fear of property crime, and a "free-floating" form of anxiety (Hough 1995). More recently, definitions include more general environmental concerns such as personal safety (Gibbs et al. 1998). Taylor (1988:252) simply views fear of crime as "an emotional or affective concern for one's safety." Moreover, fear of crime is usually viewed as being synonymous with the fear of victimization. However, victimization fear may derive from violent or property crimes that may be either general or specific.1 A more recent thrust is to discuss a "formless" fear of crime, as opposed to a fear tied directly to a specific type of crime (Ferraro 1995; Ferraro and LaGrange 1987; LaGrange and Ferraro 1987; Weinrath and Gartell 1996). The combination of these different approaches clearly makes the identification and standardization of crime fear quite difficult. Hough (1995) has perhaps summed up the problem with an acknowledgment that fear is not merely a single, unitary concept but instead a very complex (and probably interactive) set of attitudes and feelings.
The specification problem is also important. It seems plausible that some forms of crime generate more fear than others. Therefore, more abstract measures of fear may not tap the immediacy of concern over some specific crime. Previous research implies that personal crimes--assault, rape, and robbery--are most frightening to people. Attempts to clarify these relationships suggest that some subgroups, such as the elderly or men and women, have a proportionately higher fear of a few specific crimes rather than crime in general (Gubrium 1974; Banks, Maloney, and Willcock 1975; Antunes et al. 1977; Ortega and Myles 1987; Newburn and Stanko 1994; Gilchrist et al. 1998; McCoy et al. 1996). This potential slant in survey outcome is best offset, according to Ferraro and LaGrange (1985), if one specifies the exact type of crime when attempting to measure fear.
Methodological Problems
One of the major methodological problems in any form of research is the ability to control extraneous variance in presumed relationships. Where fear of crime research is concerned, it seems that researchers have barely noticed this problem. In fact, it is almost as if error in measurement is considered acceptable as long as the measures being used are equally erroneous across studies. A primary example of this is the lack of concern with potentially confounding variables (Ortega and Myles 1987). Perhaps because of the preoccupation with surveying or reporting variation in levels of fear in different groups, little attention has been paid to thorough comparisons of these groups.
Some of the discrepant results noted above may be explained by this failure. For example, in two classic studies by Gerbner et al. (1977, 1978), relationships are reported between television viewing and fear of crime. However, a replication of the study, controlling for the actual amount of crime in the area, finds no significant relationship (Doob and MacDonald 1979). Controlling for objective crime conditions at the neighborhood level, though difficult, may be an important consideration since one study (Tseng 1986) finds that crime rates at a more inclusive county level are not a significant predictor of fear of crime. Taylor and Hale (1986) point out that it is this failure of fear levels to covary spatially with crime levels that has turned debate to the meaning or construct validity of fear of crime survey items.
Validity. How one operationalizes fear of crime may be the key to avoiding some problems of construct validity.2 Perhaps the most damning commentary of the operationalization problem is a recent assessment by Hale (1996:80): "One of the principal reasons for conflicting findings concerning fear of crime lies in the confusion and lack of agreement in the construction of empirical instruments." Polls often equate "concern," "awareness," or "risk" with fear (Furstenberg 1971; van der Wurff et al. 1989:142; Williams and Dickinson 1993). While the classic factoring of fear into violent, property, and anxiety components (a la Taylor and Hale [1986]) is relevant here, the greater problem is the specific use of terminology in the measuring items themselves. For instance, authors have argued for a quarter-century (from Furstenberg in 1971 to Ferraro in 1995) that careful operationalization will filter out a broader "state of the world" worry from the more narrow fear of crime. However, little has been done in that regard. As an example, Kennedy and Krahn (1984), using a common measure, "How safe do you feel walking alone in your neighborhood at night," do not specifically address crime. A person might be afraid of attack by dogs, being hit by a car in a high traffic area, or simply afraid of the dark. Moeller (1989:212) also notes the weaknesses of this single-item question but defends it as the most frequently used, thus allowing her to make comparisons with other similarly-operationalized data sets. Teske and Hazlett (1988), using a set of similar measures, acknowledge their problematic nature and yet still insist that the items have face validity, largely because the larger questionnaire itself was about crime issues (1988:289).
In another example of a broad approach to fear of crime, Hepburn and Monti (1979) attempt to measure fear of crime among teens by using only the notion that someone might "hurt you or bother you" at school. This approach neglects the assessment of property crime although the authors refer to a poll that shows that more of those surveyed suffered some type of property crime, property damage, or destruction at school than feared for their personal safety. More recently, McCoy et al. (1996) used "fear of crime" as one of thirteen items under a question asking about the seriousness of various problems. At least the measure has the benefit of being directly related to fear of crime, but it is less about being fearful of crime than about how serious fear of crime is relative to other problems faced by people.
Thus there is no consensus on a "best" measure for fear of crime, nor even on exactly what may be measured by a particular approach. The most common approach to fear of crime measurement uses some variation of the National Crime Survey's "walking in the neighborhood" questions, which offer a very general approach to crime fear.3 Few researchers have been concerned with the exact type of fear they are tapping and those few have generally attempted to make the distinction in an ex-post-facto fashion. In reality, the term "fear of crime" is an artifact of a broad interest in what is presumed to be the psychological effect of crime. As a result, there is no clear rationale behind its use. There are some, however, who prefer to reserve fear as the effect of anticipated physical violence and use "worry" as a characterization of concern over property crime (Maxfield 1984; Garofalo 1981). In a study of the effect of victimization on fear, Skogan (1987) focuses on attitudinal measures using "concern" and "worry" rather than the term "fear." His rationale is that there is a problem of multiple meanings in the term "fear." Such a distinction nonetheless fails to define the element in fear itself which conveys a separate state of mind from other more general forms of fear. If, in fact, there is a distinct element which distinguishes fear of crime from other psychological states, this element is not yet adequately specified in the criminological literature and thus the term still retains its generic flavor.
"Fear" has been reported to be a problematic concept with potentially differential responses caused by reticence to disclose vulnerability (Newburn and Stanko 1994) or even emotional distress (LaGrange and Ferraro 1987). Thus, it seems that a less psychologically intrusive but similar concept might be useful in overcoming such problems. In the first alternative approach of which we are aware, a Canadian victimization study (Sacco 1982, 1985; Sacco and Glackman 1987) in Vancouver and other Alberta cities used the concept of "worry" to create a zero-to-ten point scale by which respondents rated their worry about crime. Their scale appears to create a unidimensional concept and has a high reliability coefficient. A second study, presented to the 1991 meetings of the British Society of Criminology by Williams, Akers, and McShane (1991) make a preliminary argument for substituting the terms "concern" or "worry" in items that purport to measure fear of crime. Subsequent to that paper, the Scottish and British Crime Surveys, as well as some German surveys, develop Likert-based measures around concern or worry about crime. The primary architect of the Scottish Crime Poll, Jason Ditton and his colleagues (1999a, 1999b) have now begun asking questions about another alternative measure, "anger," in comparison to "fear" measures. Other research, such as Williams and Dickinson's (1993) examination of the relationship between newspaper crime reporting and the fear of crime, also use worry as the basis for measurement.
Reliability. Even assuming that one knows the specific fear of crime that is being measured, there is still the problem of reliability. It is worth noting that in the vast majority of studies,4 there is no mention of the issue of reliability, much less a reported estimate for the fear measure. Two issues surface here: the number of measurement items and the way in which those items fit together.
Reliability and validity become questionable when fear of crime is measured with only a single question. In a review of the fear of crime literature from 1971-1984, Ferraro and LaGrange (1985) cite 28 studies that use only single-item indicators of fear of crime. Since that time, the single-item criticism has not sufficed to preclude research using such measures (Ferraro 1995). As Taylor and Hale (1986:158) note, avoiding single-item scales will build analyses that are "less noisy than single item..." and "...avoid developing an extensive research tradition limited to very narrow outcome measures." From the viewpoint of reliability, measurement error is decreased when the number of similar items is increased. Therefore, more reliable measures of fear of crime will be found among multiple-item measures or scales.
However, the use of multiple items does not rule out simplicity. A series of questions can be easily understood and still allow the respondent a range of answers that provide a "close fit" choice. The common use of "yes/no" or trichotomy responses should be discouraged in the construction of fear of crime instruments. Similarly, we should now be beyond commonly used four-item Likert scales and headed toward better interval-level measurement. The fact is that multi-point, metric-ordinal rating scales yield lower measurement error and provide a closer approximation of true response positions.
A final concern in developing fear of crime measurements is that the items presented for analyses be unidimensional, measuring a single concept. Fear of crime questions have normally resulted in an aggregation of items with no concern for reliability or unidimensionality. For example, two forms--fear of walking in one's neighborhood at night and fear experienced while alone in the house are frequently combined. There is reason to believe that the fear expressed in these two items is derived from separate sources and should not be included in the same measure.
To compensate for current weaknesses in measuring of fear of crime, attention should be paid to the operationalization of the term as well as the design of survey questions. The optimum construction of a fear of crime measure would describe a unidimensional concept but have several similar items combined to measure that idea (and, importantly, offering simultaneous measurement of concrete fear items contained within the general measure). The individual questions themselves should offer multi-point responses, minimally at the metric-ordinal level, thus increasing accuracy of measurement. Finally, the constructed scale should have a high degree of reliability.
This article addresses some of these measurement problems and reports the results of a survey research project that is specifically designed to examine fear of crime measures using the suggested operationalization of "worry" about crime. The instrumentis constructed with the express intention of determining the efficacy of several of the more common fear of crime measures as they are translated to a "worry" dimension. In addition, and following our own critique, we create a multi-point, metric-ordinal, multiple-item rating scale that narrowly focuses on crime. We then test that scale in a general population survey.
DATA COLLECTION
The data were compiled from a statewide survey conducted in a Southwestern state in 1986.5 The sampling protocol involved drawing a simple random sample (with replacement) of 10,000 names from the state listings of drivers licenses. This sampling frame provides a reasonable guarantee that residents in both metropolitan and rural areas are adequately represented in the sample. The state agency responsible for issuing drivers licenses estimates that about ninety percent of adults are present in the data base. In addition, the data base represents both those with drivers licenses and a small percentage of individuals who merely hold identification cards issued by the agency.
The survey contains two stratified samples, both of which are random samples drawn from separate original samples of 10,000 names. The first sample is of 1,000 individuals who were randomly selected from those aged 65 and over. All of these individuals are used in the final survey because of a secondary purpose of examining fear of crime among the elderly. The second sample is of 1,000 individuals randomly selected from those under age 65. Thus, the final combination of the two samples contains the names and addresses of 2,000 people.
These 2000 individuals received a 24 page survey booklet that includes questions designed to explore issues in victimization and fear of crime. The survey begins with a preliminary letter, followed by the survey package one week later. The package contains a cover letter, the survey booklet, and a postage-paid return envelope. Approximately two weeks later we mailed a reminder letter to those who had not returned the instrument. Ten days later, another complete survey package was mailed and ten days after that we mailed a final follow-up letter. Thus, there was a possible total of five contacts with potential respondents.
Usable responses (or some set thereof) were provided by 1,152 individuals. Correcting for nonforwardable addresses and those who were recently deceased, we obtain an adjusted return rate of sixty percent (a raw return rate of 57.6 percent). The data were also weighted to achieve population parity between the elderly and non-elderly strata. Since problems arise from modified sample sizes when calculating probability levels, we adjusted the entire sample to reflect a .20 proportion of elderly while maintaining the same size as the original sample.
THE FEAR OF CRIME MEASURES
The survey instrument containes a variety of fear of crime measurements, ranging from an adaptation of the standard National Crime Survey questions to a series of anticipated victimization questions. Some are single items, while others include multiple items. In all cases, however, we replace any use of the words "fear" or "afraid" with "worry." As a result, these data may differ from those obtained by some fear of crime research. Indeed, where we use the common phrase "fear of crime" in discussing our analyses, it should be understood that the real concept is worry about crime.6
Single-Item Measures
There are three major single-item measures. Each of these variables make use of a zero to ten ranking scale to increase the response variance. Compared to traditional Likert scales, such a response set enhances the measurement quality of the data and approaches the concept of magnitude. The latter is especially helpful since the use of a numbered scale with standard boundaries boosts what is normally ordinal measurement to a close approximation of interval scales.
The first single-item question seeks to measure a nonspecific worry about crime. The wording of that question is, "On a scale of 0 to 10, how concerned are you about crime in general?"7 The 0 point is defined as "Not worried at all about crime in general," and the 10 point is labeled "Very worried about crime in general." The second question concerns the respondents' perception of their victimization risk. That item asks, "On a scale of 0 to 10, what do you think your chances are of being a victim of any type of crime during the next year?" and provides labels of 0, "I will not be a victim of crime," and 10, "I will certainly be a victim of crime." The final question follows four separate listings of the same fifteen offenses and is designed to measure an overall worry about victimization. The wording is "Overall, how worried are you about becoming a victim of any of these fifteen crimes during the next year." The end points of the 0 to 10 scale are labeled "Not worried at all" and "Very worried."
These three single-item measures are not strongly intercorrelated. The highest amount of shared variance is less than 25 percent (see Table 1). Thus, at a bivariate level, the measures appear to be relatively independent of each other. In addition, a reasonable amount of variance is observed (a standard deviation of 2.4 to 3.0 over a range of 10) and all the distributions are within the bounds of normality with standardized skewness and kurtosis8 values inside of +/- 1.0 (see Table 2). Except for reliability problems inherent in single-item measures, the three variables appear to be individually useful, and relatively independent, approaches to the measurement of worry about crime.
Multiple-Item Measures
There are five multiple-item measures of worry about crime present in the survey instrument. They cover questions about walking around one's home and neighborhood, anticipated victimization, and behavioral adaptations due to crime. The wording of each question appears below. The discussion that follows focuses on the reliability and usefulness of these multiple-items as worry-about-crime scales.
NCS "walking" worry scale. The first series uses the National Crime Survey approach by asking about walking alone or with someone else within four blocks/one block of home and about being at home alone. The base questions are worded as follows: "Is there any area within (four blocks/one block) of your house where you would be worried about walking (alone/even if someone else were with you)?" and "Are you worried about being in your home alone?" Responses are elicited for both night and day with choices of "no" (0), "maybe" (1), and "yes" (2).
The individual items demonstrate reasonable variance but the modal response pattern suggests a potential problem. In all except one item (four blocks, alone, night), the "no worry" response category accounts for more than fifty percent of the respondents. A reliability analysis finds that all items except those measuring in-home worry have item-total correlations of .6 or higher with an initial alpha of .902 for all ten items. However, the two in-home items are poorly correlated with the scale, suggesting that they are measuring some other dimension. A factor analysis confirms this and they were dropped from the final cumulative eight-item scale, yielding item-total correlations no lower than .66 and an alpha of .898. An examination of the scale distribution shows a "normal" kurtosis value under one (.619). There is, however, an abnormal positive skew (1.196), with a large number of respondents who report no worry over walking in their neighborhoods. This suggests that the scale may cause problems when used in multivariate statistical techniques assuming joint normal distributions. However, a similar scale developed by Teske and Hazlett (1988) on the usual fear-based measures using a Guttman procedure produced skewness values less than 1.0. Therefore, the non-normal kurtotic score of our data may be an artifact of the sample (as could the finding in the Teske and Hazlett data).
Victimization worry scale. The second series employs a group of fifteen offenses for which respondents were asked to estimate their chances of victimization during the coming year. The base question is, "We would now like to know how you feel about your chances of being a victim of any of these offenses during this coming year. On a scale of 0 to 10, how worried are you about being a victim?" The 0 to 10 rating scale is provided with end points of "not worried at all" and "very worried." Four of the offenses are of a white collar variety and are dropped from further analysis, leaving eleven offenses for examination as a scale.
The number of offenses on the victimization worry scale was further reduced when, after an examination of the correlation matrix, one item (assault without a weapon) was found to be colinear (correlation higher than .8) with other items. All ten remaining items have item-total correlations of .67 or higher. One of the remaining ten items is "rape or attempted rape" and thus was not appropriate for males. We split the scale into one for males and one for females, removing rape from the male scale. A re-examination of the separate male and female scales produces alphas of .9118 and .9311, respectively. In order to achieve numerical parity at a maximum of 100 points each, the 9-item male scale was multiplied by 10/9ths. Both scales were then combined into the final, composite scale with an alpha of .9337 (see Table 2). A factor analysis of the final scale (see Table 3) locates two readily-identifiable dimensions: violence and property factors. This seems to be appropriate, given other factor analytic studies of the fear of crime concept. In addition, the "victimization worry" scale demonstrates a normal distribution, with skewness and kurtosis values well under one.
General behavioral precautions. This is the first of three series of items that elicited information about actual measures respondents took for safety against crime. Ten items focus on precautions that respondents might be expected to take when engaged in routine outside activities. Responses to the precautions indicate whether they are taken "most of the time," "sometimes," or "never." All correlations among the items are low, as are the item-total correlations. The initial alpha level was .7127 and several attempts to reduce the items to a more unidimensional level failed to produce a higher alpha. A factor analysis produces two dimensions (carrying something for protection and more general precautionary measures). The entire ten-item scale is, nonetheless, quite close to a normal distribution (see Table 2), indicating that it may have some utility as a measure of worry about crime. We do caution, however, that the scale's reliability coefficient indicates a potential for almost as much random error (49 percent) as stability (51 percent).
Home-oriented behavioral precautions. This item-series is comprised of fourteen questions about precautions taken around the home. Items include a range of precautions from locking doors and windows to installing a burglar alarm. Responses are either "no" or "yes." As with the general behavioral precautions, bivariate correlations and item-total correlations are all low (maximum of .39) and remain so although several variations were attempted. The highest alpha level we were able to produce is .6091 for the entire 14-item scale, which is not particularly high for so many items. A factor analysis also uncovered five separate dimensions within the scale, thus leaving its utility as a scale in doubt. The distribution is, however, well within the bounds of normality (see Table 2).
Trip-oriented behavioral precautions. The final behavioral series is composed of a list of four potential precautions taken when leaving home for a weekend or more. The respondents were asked to answer "yes" or "no" to each of the items. After examining correlations among the items and an initial reliability analysis, the third (someone to watch house) and fourth items (automatic light timer) were removed because of low item-total correlations. The alpha for the final two-item scale is .8245 (see Table 2), providing a moderately reliable cumulative rating scale. With only two items, however, the fact that most of the respondents appear to take precautions before leaving on a trip suggests that the scale might not be useful in discriminating between levels of crime worry. The scale distribution is also at the bounds of normality (skewness of -1.064 and kurtosis of -.268).
General worry "walking" scale. In addition to the scales measuring worry about crime, we create a final scale that provides a measure of the respondent's non-crime worry about the neighborhood. The items are duplicates of the NCS "walking" questions, except that they precede the crime questions and request that respondents provide answers based on worrying about something other than crime. Thus, the scale is designed to tap a more general worry dimension than that captured by crime-focused questions.
The two in-home items of this scale represent a separate dimension of worry, as was the case with the NCS crime questions. They also have low item-total correlations and were dropped from the scale. One correlation was high enough to introduce a concern about potential colinearity, but there were no other problem correlations across the matrix associated with the two variables. The remaining eight item-total correlations are generally high, with .63 being the lowest. The alpha for the final scale is .884, indicating that the scale is reasonably reliable. However, the scale suffers the same problem as the NCS crime scale, with skewness above 1.0 (see Table 2). A further indication that the scale is not normally distributed is found in a kurtosis value well above one (1.84).
MEASUREMENT SCALES AND WORRY ABOUT CRIME DIMENSIONS
We approach the issue of worry about crime dimensions in two ways. First, we produce a bivariate correlation matrix so that the individual relationships between measures can be examined. Second, a factor analysis is performed to assist in the identification of potential worry dimensions. In general, the observed dimensions appear to support previous speculation in the literature.
The first observation of interest from the correlation matrix (see Table 1) is that only two correlations exceed a shared variance of 25 percent. The first, at .7322 (r-square = .53), is between the victimization worry scale and the one-item victimization worry scale that follows the items that make up the victimization worry scale. This might be interpreted as a real difference between specific offenses (even if summed) and a more general question referring collectively to those same offenses. Conversely, one could view the true relationship as perfect and attribute the difference to error as a result of (1) unreliability within the single-item general question or (2) the inability of a single-item scale to pick up the full variance of a multi-item scale. We prefer to take the middle ground and suggest that the correlation might be higher, but not perfect, were it not for reliability problems. Therefore, we conclude that some differences exist between offense specific scales and general scales.
The second, and slightly larger, correlation (.7557) is between the "walking" worry scale and the more general, non-crime worry scale (r-square = .57). Based on previous commentary in the literature, the most promising interpretation here is that an NCS type of scale may tap into a generalized set of concerns that reflect more than just a specific worry about crime. In fact, it may be reasonable to view concern over walking in one's neighborhood as indicative of a world view. Crime itself is merely one form of worry-generating stimuli.
All worry about crime scales (plus the general worry scale) were put through a principle-components factor analysis and the factors were rotated using a normalized varimax criterion in order to simplify the dimensions. The results (see Table 4) yield three easily interpreted factors with eigenvalues over 1.0. The first, and strongest (35 percent of the variance), of the factors is loaded by the scales which directly ask about crime or victimization. The second factor is comprised of the NCS-type walking scales (15 percent of the variance), and the final factor rests upon the behavioral precaution scales (14 percent of the variance).
If our interpretation of the common dimensions underlying the scales is reasonable, where worry about crime is concerned, it would appear that only the first factor focuses directly on crime. The third factor relates to a behavioral component attached to crime, while the second factor most clearly expresses a dimension of general worry about safety that is clearly not crime specific. Both of the latter two factors are relatively weak, resulting in a combined variance estimate lower than that of the first "crime" dimension.
Elaboration of the Victimization Worry Scale
With the combination of results presented here, it appears that the scales most commonly used by fear of crime researchers are less than desirable measures of the concept9 for one or more of the following reasons. Some of the "scales" have relatively low reliability estimates. This, of course, automatically includes all of the single-item scales. Several of the scales also have a limited number of response items that create only gross distinctions between levels of fear or worry. Others are questionable because they do not appear to directly measure crime and instead tap into more inclusive concepts of fear, worry, and concern. Finally, because many of the most sophisticated analytical measures commonly used today are parametric, linear tools, the lack of normality in a fear/worry scale could introduce error that is potentially greater than traditionally observed outcomes. All of this argues for a measure that clearly focuses on crime, is normally distributed, is composed of several items with a simple multi-point response format, and is reliable. We believe that, among those tested here (and we have been careful to include those measures in common use), the best option is the 10-offense victimization worry scale.
Any reliable scale must possess a reasonable degree of factorial invariance and subgroup tolerance, two major assumptions of reliability estimates. Where factorial invariance is concerned, the analyses produces the same results regardless of whether a principle-axis or a principle components approach is used. Similarly, we changed the method of rotation using varimax, quartimax, and equamax methods, and all produced identical rotated factor matrices.
For an examination of subgroup tolerance, we ran a separate reliability analysis on male and female respondents. One problem should be noted with the use of the scale for males and females: the rape item necessitated separate scales (the rape item was dropped for males, resulting in a 9-item scale with a 10/9ths multiplier to maintain the 100 point scale maximum). Therefore, any comparison of males and females along the property dimension must be controlled by recalculating the total male score. Since the property items were originally equal, the effect of the original 10/9ths multiplier must be reversed (i.e., multiply by 9/10ths). Failure to correct this will produce artificially inflated male worry scores for property victimization. After this correction, the female and male scales yield alphas of .9311 and .9118, respectively (see Table 3). For these two subgroups, then, the scale remains reliable.
A final examination involved potential subscales for violent and property victimization. The two potential subscales clearly reflect the expectations of the literature on dimensions of crime and the main question involving their utility is reliability. Reliability analyses on the two subscales yield reasonably strong alpha coefficients (.8921 and .9235). Further, the skewness and kurtosis coefficients are within bounds of normality, suggesting that both subscales are reliable and useful in parametric statistics (see Table 3).
DISCUSSION
The findings and comparisons offered in this research report suggest that the common measures of fear of, and worry about, crime leave much to be desired. Our results lead us to believe that the best approach to measuring crime fear/worry is a multiple-item scale comprised of items that are all focused directly on crime. However, the term fear of crime remains vague and we argue that it may be preferable to substitute the concept of worry. Until the field reaches agreement on the elements of fear of crime, both psychological and physiological, we believe that the term fear of crime should be put aside in favor of the term "worry about crime" (or victimization).
The scale described here offers simplicity as well as utility. It has ease of administration, the instructions are simple and understandable and the zero-to-ten rating scale for each item appears to be a culturally intuitive approach for respondents. The construction of the final scale from the individual items is relatively straightforward. In summing the items, a researcher must only remember that the rape item necessitates the temporary creation of separate male and female scales that are then recombined after an adjustment.
The victimization worry scale also seems to meet most of the criteria for a usable and reliable scale. The scale has an approximately normal distribution and approaches the interval level of measurement, both of which allow for its use in linear multivariate tools. All tests of reliability yield respectable alpha levels of .89 or higher and those coefficients hold up across various groups and samples.
Where unidimensionality is concerned, we do not claim the scale to be unidimensional, except under the general rubric of "crime." Even this is an improvement over other commonly used scales because the scale is at least focused on crime rather than other worry/fear producing factors. The probable existence of two subscales for violence and property victimization may even be desirable and one can certainly envision research that would concentrate on one or the other of those two dimensions. In sum, we believe those interested in tapping a concern or worry about crime will find this scale to be both reliable and useful.
On the other hand, those who remain focused on "fear" of crime should use this scale with care. We do not know precisely what effect substitution of the term "fear" for our "worry" will have on factorability or reliability. In particular, it is possible that fear-based versions of the various measures we compare here might produce different results. Until such time as fear and worry/concern can be directly compared on the same instrument with the same subjects, these cautions remain. We do note that the few attempts to establish reliability of fear-based scales fare no better (and usually worse) than the results for the worry scale and the literature consistently acknowledges a validity problem in the use of the fear concept. The overridding problem, as Ferraro and LaGrange (1987:71) noted over a decade ago, is "the phrase 'fear of crime' has acquired so many divergent meanings that its current utility is negligible." Thus, we reiterate our earlier concerns about the measurement of fear and believe the "worry about victimization" approach may be a viable way for researchers to make a fresh start.
1. Fear of specific forms of crime is now being referred to as "concrete fear." See, e.g., LaGrange and Ferraro (1989) and Carcach et al. (1995).
2. See Fattah and Sacco (1989:207-212) for an excellent discussion of differential operationalizations of fear of crime.
3. As a few worldwide examples see the 1982 and 1985 Canadian Urban Victim Surveys, the 1991 Queensland Crime Victims Survey, the 1993 and 1995 Scottish Crime Surveys, and the 1994 and 1998 British Crime Surveys.
4. There are some exceptions. See, for instance, Sacco (1982), Skogan (1987), Taylor and Hale (1986), Teske and Hazlett (1988), van der Wurff et al. (1989), Smith and Hill (1991) and Thompson et al. (1992).
5. The survey contains self-report information on sensitive, and some potentially illegal, behavioral topics. As an ethical consideration, we are merely identifying the state as a "Southwestern" state.
6. Our use of "fear of crime," even though we focus on concern and worry about crime, is consonant with much of the research literature. Indeed, there are so many approaches to fear of crime, both in measurement and imputed meaning, that the term may now be of little use. One reason, then, for changing to "worry" about crime may be to provide a modicum of specificity.
7. The main question uses the term "concern" but the response categories use the term "worry." We realize that this error could create a potential difference in response. However, the greater differences between the term "fear" and "worry/concern" lead us to believe that the relative error will be minimal.
8. We used the skewness and kurtosis formulas within SPSS for the calculations. As a result, the kurtosis values are modified (original formula value minus three) to set a perfect mesokurtic curve at 0.0 rather than the central point of 3.0 in the original kurtosis formula.
9. There is, of course, the possibility that these results are specific to concern and worry about victimization rather than fear of crime. Such a degree of uniqueness seems unlikely, however. A thorough reading of the fear of crime literature would easily lead one to anticipate our results. Therefore, we believe that the general direction of our findings also applies to measures of fear of crime as well.
Akers, Ronald L., Anthony J. LaGreca, Christine Sellers, and John
Cochrane. 1987. "Fear of Crime and Victimization among the Elderly in
Different Types of Communities." Criminology
25:487-506.
Antunes, G. E., F. L. Cook, T. D. Cook, and Wesley G. Skogan. 1977. "Patterns of Personal Crime against the Elderly: Findings from a National Survey." Gerontologist 17:321-327.
Baldassare, Mark. 1986. "The Elderly and Fear of Crime." Social Service Review 70:218-220.
Banks, C., E. Maloney, and H. D. Willcock. 1975. "Public Attitudes to Crime and the Penal System." British Journal of Criminology 15:228-240.
Bennett, Trevor. 1989. Contact Patrols in Birmingham and London: An Evaluation of a Fear Reducing Strategy. Cambridge, England: Institute of Criminology, University of Cambridge.
Carcach, Carlos, Peta Frampton, Kaye Thomas, and Matthew Cranich. 1995. "Explaining Fear of Crime in Queensland." Journal of Quantitative Criminology 11:271-287.
Chiricos, Theodore, Sarah Eschholz, and Marc Gertz. 1997a. "Crime, News and Fear of Crime: Toward an Identification of Audience Effects." Social Problems 44:342-357.
Chiricos, Theodore, Michael Hogan, and Marc Gertz. 1997b. "Racial Composition of Neighborhood and Fear of Crime." Criminology 35:107-131.
Clarke, Alan H., and Margaret J. Lewis. 1982. "Fear of Crime among the Elderly: An Exploratory Study." British Journal of Criminology 22:49-62.
Clemente, Frank, and Michael B. Kleiman. 1976. "Fear of Crime among the Aged." Gerontologist 16:207-210.
__________ . 1977. "Fear of Crime in the United States: A Multivariate Analysis." Social Forces 56:519-531.
Ditton, Jason, Jon Bannister, Elizabeth Gilchrist, and Stephen Farrall. 1999a. "Afraid or Angry? Recalibrating the 'Fear' of Crime." International Review of Victimology 6:83-99.
Ditton, Jason, Stephen Farrall, Jon Bannister, Elizabeth Gilchrist, and Ken Pease. 1999b. "Reactions to Victimisation: Why Has Anger Been Ignored?" Crime Prevention and Community Safety: An International Journal 1(3):37-54.
Donnelly, P. G. 1988. "Individual and Neighborhood Influences on Fear of Crime." Sociological Focus 22:69-85.
Doob, A. N. and G. E. MacDonald. 1979. "Television Viewing and Fear of Victimization: Is the Relationship Causal?" Journal of Personality and Social Psychology 37:170-179.
Fattah, Ezzat A., and Vincent F. Sacco. 1989. Crime and Victimization of the Elderly. New York: Springer-Verlag.
Ferraro, Kenneth F. 1995. Fear Of Crime. Albany: State University of New York Press.
Ferraro, Kenneth F., and Randy LaGrange. 1985. "The Measurement of Fear of Crime: A Review and Some Suggestions." Paper presented at the Southern Sociological Society, Charlotte, North Carolina.
__________. 1987. "The Measurement of Fear of Crime." Sociological Inquiry 57:70-101.
Fitzpatrick, Kevin M., Mark E. LaGory, and Ferris J. Ritchey. 1993. "Criminal Victimization among the Homeless." Justice Quarterly 10:353-368.
Freudenburg, W. 1986. "The Density of Acquaintanceship: An Overlooked Variable in Community Research?" American Journal of Sociology 92:27-63.
Furstenberg, Frank F., Jr. 1971. "Public Reaction to Crime in the Streets." American Scholar 40:601-610.
Garofalo, James. 1979. "Victimization and the Fear of Crime." Journal of Research in Crime and Delinquency 16:80-87.
Garofalo, James. 1981. "The Fear of Crime: Causes and Consequences." Journal of Criminal Law and Criminology 72:714-726.
Garofalo, James, and John Laub. 1978. "The Fear of Crime: Broadening our Perspective." Victimology 3:242-253.
Gerbner, G., L. Gross, M. Elley, M. Jackson-Beeck, S. Jeffries-Fox, and N. Signorelli. 1977. "TV Violence Profile No. 8: The Highlights." Journal of Communications 27:171-180.
Gerbner, G., L. Gross, M. Elley, M. Jackson-Beeck, S. Jeffries-Fox, and N. Signorelli .1978. "Cultural Indicators: Violence Profile No. 9." Journal of Communications 28:176-207.
Gibbs, John J. and Kathleen J. Hanrahan. 1993. "Safety Demand and Supply: An Alternative to Fear of Crime." Justice Quarterly 10:369-394.
Gibbs, John J., Charles M. Puzzanchera, Kathleen J. Hanrahan and Dennis Giever. 1998. "The Influence of Personal Safety and Other Environmental Concerns on Sense of Control and Emotional Well-being." Criminal Justice and Behavior 25:403-425.
Gilchrist, Elizabeth, Jon Bannister, Jason Ditton, and Stephen Farrall. 1998. "Women and the 'Fear of Crime'." British Journal of Criminology 38:283-298.
Gubrium, J. 1974. "Victimization in Old Age." Crime and Delinquency 20:245-250.
Hale, C. 1996. "Fear of Crime: A Review of the Literature." International Review of Victimology 4: 79-150.
Hepburn, John R. and D. J. Monti. 1979. "Victimization, Fear of Crime and Adaptive Responses among High School Students." In Perspectives in Victimology, edited by W. H. Parsonage. Beverly Hills: Sage.
Hindelang, Michael J., Michael R. Gottfredson, and James Garofalo. 1978. The Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. Cambridge, MA: Ballinger.
Holloway, Wendy and Tony Jefferson. 1997. "The Risk Society in an Age of Anxiety: Situating Fear of Crime." British Journal of Sociology 48:255-266.
Hough, M. 1995. Anxiety about Crime: Findings From the 1994 British Crime Survey, Home Office Research Study No. 147. London: Home Office.
Kennedy, Leslie W. and Harvey Krahn. 1984. "Rural-Urban Origin and Fear of Crime: The Case for Rural Baggage." Rural Sociology 49:247-260.
Krahn, Harvey and Leslie W. Kennedy. 1985. "Producing Personal Safety: The Effects of Crime Rates, Police Force Size and Fear of Crime." Criminology 23:697-710.
Krannich, R., E. Berry, and T. Greider. 1989. "Fear of Crime in Rapidly Changing Rural Communities: A Longitudinal Analysis." Rural Sociology 54:195-212.
Krannich, R., T. Grieder, and R. Little. 1985. "Rapid Growth and Fear of Crime: A Four Community Comparison." Rural Sociology 50:193-209.
Kuttschreuter, Margot and Oene Wiegman. 1998. "Crime Prevention and the Attitude Toward the Criminal Justice System: The Effects of a Multimedia Campaign." Journal of Criminal Justice 26:441-452.
LaGrange, Randy L. and Kenneth F. Ferraro. 1987. "The Elderly's Fear of Crime: A Critical Examination of the Research." Research on Aging 9:372-391.
_________. 1989. "Assessing Age and Gender Differences in Perceived Risk and Fear of Crime." Criminology 27:697-719.
Lebowitz, B. D. 1975. "Age and Fearfulness: Personal and Situational Factors." Journal of Gerontology 30:696-700.
Lee, Gary R. 1982a. "Sex Differences in Fear of Crime among Older People." Research on Aging 3:284-298.
_________. 1982b. "Residential Location and Fear of Crime among the Elderly." Rural Sociology 47:655-669.
_________. 1983. "Social Integration and Fear of Crime Among Older Persons." Journal of Gerontology 38:745-750.
Liska, Allen and William Baccaglini. 1990. "Feeling Safe by Comparison: Crime in the Newspapers." Social Problems 37:360-374.
Maguire, M. 1980. "The Impact of Burglary upon Victims." British Journal of Criminology 20:261-275.
Maxfield, Michael G. 1984. Fear of Crime in England and Wales. London: H.M. Stationary Office.
Markson, E. and B. Hess. 1980. "Older Women in the City." Signs 5:127-141.
McCoy, H. Virginia, John D. Wooldredge, Francis T. Cullen, Paula J. Dubeck, and Sandra L. Browning. 1996. "Lifestyles of the Old and Not So Fearful: Life Situation and Older Person's Fear of Crime." Journal of Criminal Justice 24:191-205.
Moeller, Gertrude L. 1989. "Fear of Criminal Victimization: The Effect of Neighborhood Racial Composition." Sociological Inquiry 59:208-221.
Newburn, T. and Elizabeth A. Stanko. 1994. "When Men are Victims." In Just Boys Doing Business? Men, Masculinities and Crime, Edited by T. Newburn and E. A. Stanko. London: Routledge.
Normoyle, J. and P. J. Lavrakas. 1984. "Fear of Crime in Elderly Women: Perceptions of Control, Predictability, and Territoriality." Personality and Social Psychology Bulletin 10:191-202.
Ortega, Suzanne L. and Jessie L. Myles. 1987. "Race and Gender Effects on Fear of Crime: An Interactive Model with Age." Criminology 25:133-152.
Parker, Keith D. and Melvin C. Ray. 1990. "Fear of Crime: An Assessment of Related Factors." Sociological Spectrum 10:29-40.
Reiss, Albert J., Jr. 1983. "Crime Control and the Quality of Life." American Behavioral Scientist 27:43-58.
Riger, Stephanie, Margaret Gordon and Robert Le Bailly. 1978. "Women's Fear of Crime: From Blaming to Restricting the Victim." Victimology 3:274-284.
Rountree, Pamela Wilcox and Kenneth C. Land. 1996a. "Perceived Risk Versus Fear of Crime: Empirical Evidence of Conceptually Distinct Reactions in Survey Data." Social Forces 74:1353-1376.
__________. 1996b. "Burglary Victimization, Perceptions of Crime Risk, and Routine Activities: A Multilevel Analysis Across Seattle Neighborhoods and Census Tracts." Journal of Research in Crime and Delinquency 33:147-180.
Sacco, Vincent F. 1982. "An Exploratory Analysis of the Conceptual Meaning of Perceptions of Crime." Canadian Journal of Criminology 24:295-306.
Sacco, Vincent F. 1985. "City Size and Perceptions of Crime." Canadian Journal of Sociology. 10:277-293.
Sacco, Vincent F., and W. Glackman. 1987. "Vulnerability, Locus of Control and Worry about Crime." Canadian Journal of Community Mental Health 6:99-111.
Skogan, Wesley G. 1987. "The Impact of Victimization on Fear." Crime and Delinquency 33:135-154.
Smith, Lynn Newhart and Gary D. Hill. 1991. "Perceptions of Crime Seriousness and Fear of Crime." Sociological Focus 24:315-327.
St. John, Craig and Tamara Heald-Moore. 1996. "Age, Racial Prejudice, and Fear of Criminal Victimization in Public Settings." Sociological Focus 29:15-31.
Stanko, Elizabeth A. 1997. "Safety Talk: Conceptualizing Women's Risk Assessment as a 'Technology of the Soul'." Theoretical Criminology 1:479-499.
Taylor, Ralph B. 1988. Human Territorial Functioning. New York: Cambridge University Press.
Taylor, Ralph B. and Jeanette Covington. 1993. "Community Structural Change and Fear of Crime." Social Problems 40:374-397.
Taylor, Ralph B. and Margaret Hale. 1986. "Testing Alternative Models of Fear of Crime." Journal of Criminal Law and Criminology 77:151-189.
Teske, Raymond H. C. and Michael H. Hazlett. 1988. "A Scale for the Measurement of Fear of Crime." American Journal of Criminal Justice 12:274-292.
Thompson, Carol Y., William B. Bankston, and Roberta St. Pierre. 1992. "Parity and Disparity Among Three Measures of Fear of Crime: A Research Note." Deviant Behavior 13:373-389.
Toseland, R.W. 1982. "Who is Most Vulnerable?" Journal of Criminal Justice 10:199-209.
Tseng, Ben Y. 1986. Fear of Crime: A Test of a Theoretical Model. Unpublished doctoral dissertation. Huntsville, TX: Sam Houston State University.
Vander Ven, Thomas M. 1998. "Fear of Victimization and the Interactional Construction of Harassment in a Latino Neighborhood." Journal of Contemporary Ethnography 27:374-398.
van der Wurff, A., L. van Staalduinen, and P. Stringer. 1989. "Fear of Crime in Residential Environments:Testing a Social Psychological Model." Journal of Social Psychology 129:141-160.
Vogel, Ron and Sam Torres. 1998. "An Evaluation of Operation Round Up: An Experiment in the Control of Gangs to Reduce Crime, Fear of Crime and Improve Police Community Relations." Policing 21:38-53.
Walklate, Sandra. 1998. "Excavating the Fear of Crime: Fear, Anxiety or Trust?" Theoretical Criminology 2:403-418.
Warr, Mark. 1984. "Fear of Victimization: Why are Women and the Elderly More Afraid?" Social Science Quarterly 65:681-702.
Warr, Mark. 1990. "Dangerous Situations: Social Context and Fear of Victimization." Social Forces 68:891-907.
Weinrath, Michael and John Gartrell. 1996. "Victimization and Fear of Crime." Violence and Victims 11:187-197.
Williams, Frank P. III, Ronald L. Akers, and Marilyn D. McShane. 1991. "Measuring Worry about Crime: The Construction of a Reliable Scale." Paper presented at the biannual meeting of the British Society of Criminology, York, England.
Williams, Paul and Julie Dickinson. 1993. "Fear of Crime: Read All About It? The Relationship Between Newspaper Crime Reporting and Fear of Crime." British Journal of Criminology 33:33-56.
Winkel, Frans W. 1998. "Fear of Crime and Criminal Victimization." British Journal of Criminology 38:473-484.
Winkel, F. W., and A. Vrij. 1990. "Fear of Crime and Mass Media Crime Reports: Testing Similarity Hypotheses." International Review of Victimology 1:251-265.
The wording of items on the survey instrument and the order in which they appear (original question numbers have been retained) are as follows:
GENERAL CONCERN ABOUT SAFETY
In the following questions, we are NOT CONCERNED ABOUT CRIME. Please tell us if you are worried about something OTHER THAN CRIME.
25. Is there any area within four blocks of your home where you would be worried about walking alone?
A. during the day? “yes “maybe “no
B. at night? “yes “maybe “no
26. Is there any area within four blocks of your home where you would be worried about walking even if someone else were with you? (responses as above)
27. Is there any area within one block of your home where you would be worried about walking alone? (responses as above)
28. Is there any area within one block of your home where you would be worried about walking even if someone else were with you? (responses as above)
29. Are you worried about being in your home alone? (responses as above)
CONCERN ABOUT CRIME
We would now like to know how worried you may be about CRIME. Please answer the following with ONLY CRIME IN MIND.
30. On a scale of 0 to 10, how concerned are you about crime in general?
0 = NOT WORRIED AT ALL about crime in general
10 = VERY WORRIED about crime in general
“1 “2 “3 “4 “5 “6 “7 “8 “9 “10
37. Is there any area within four blocks of your home where you would be worried about walking alone?
A. during the day? “yes “maybe “no
B. at night? “yes “maybe “no
38. Is there any area within four blocks of your home where you would be worried about walking even if someone else were with you? (responses as above)
39. Is there any area within one block of your home where you would be worried about walking alone? (responses as above)
40. Is there any area within one block of your home where you would be worried about walking even if someone else were with you? (responses as above)
41. Are you worried about being in your home alone? (responses as above)
CRIME PRECAUTIONS
Do you take any of these measures for SAFETY against crime?
48. Get someone to go with you when you go out after dark?
“most of the time “sometimes “never, almost never
49. Plan your route to avoid certain dangerous places? (responses as above)
50. Take something with you at night that could be used for protection like a dog, whistle, knife or gun? (responses as above)
51. When riding or sitting in a car, do you keep the doors locked? (responses as above)
52. Telephone back to a friend or relative to say that you've arrived safe at home? (responses as above)
53. Get someone to go with you during the day? (responses as above)
54. Carry a gun in the car? (responses as above)
55. Carry a weapon with you? (responses as above)
56. Carry Mace or a spray with you to repel attackers? (responses as above)
57. Avoid carrying much cash with you? (responses as above)
HAVE YOU DONE ANY OF THESE THINGS AROUND THE HOUSE?
58. Keep a gun in the house “no “yes
59. Installed a burglar alarm? (responses as above)
60. Always keep doors and windows locked? (responses as above)
61. Formed a crime watch group with your neighbors? (responses as above)
62. Installed better or extra door locks? (responses as above)
63. Installed a door security chain? (responses as above)
64. Installed window locks? (responses as above)
65. Put burglar bars on your windows? (responses as above)
66. Installed a security fence? (responses as above)
67. Put in an outside security light? (responses as above)
68. Got a guard dog? (responses as above)
69. Put identification numbers on your property? (responses as above)
70. Put warning stickers on your windows/doors? (responses as above)
71. Changed the landscaping around your home? (responses as above)
NOW THINK OF THE LAST TIME YOU WENT AWAY FOR A WEEKEND OR MORE. DID YOU:
72. Stop delivery of newspapers or have some bring them in? (responses as above)
73. Stop delivery of mail or have someone bring it in? (responses as above)
74. Have someone watch you home? (responses as above)
75. Set an automatic time to switch lights on and off? (responses as above)
VICTIMIZATION AND CRIME SERIOUSNESS
76. On a scale of 0 to 10, what do you think your CHANCES are of being a VICTIM OF ANY TYPE OF CRIME DURING THE NEXT YEAR?
0 = I will NOT be a victim of crime 10 = I will CERTAINLY be a victim of crime
“1 “2 “3 “4 “5 “6 “7 “8 “9 “10
ANTICIPATED VICTIMIZATION EXPERIENCES
We would like to know how you feel about your chances of being a victim of any of these offenses DURING THIS COMING YEAR. On a scale of 0 to 10, how worried are you about being a victim?
122. Robbery
0 = NOT WORRIED AT ALL 10 = VERY WORRIED
“1 “2 “3 “4 “5 “6 “7 “8 “9 “10
123. Assault with a weapon (responses as above)
124. Assault without a weapon (responses as above)
125. Rape or attempted rape (responses as above)
126. Arson or attempted arson (responses as above)
127. Burglary (responses as above)
128. Motor vehicle theft (responses as above)
129. Other theft (responses as above)
130. Fraud (responses as above)
131. Vandalism or malicious mischief (responses as above)
132. Murder (responses as above)
133. Drunk driving accident (responses as above)
134. False advertising (responses as above)
135. Unsafe products (responses as above)
136. Overcharging (responses as above)
143. Overall, how worried are YOU about becoming a victim of any of these fifteen crimes during the NEXT YEAR?
0 = NOT WORRIED AT ALL 10 = VERY WORRIED
1 2 3 4 5 6 7 8 9 10
Frank P. Williams III
Franks Williams is professor of Criminal Justice at California State University, San Bernardino, where he is director of a criminal justice research center, and he is affiliated with Prairie View A & M University. His most recent research has been in the area of parole classification systems and the most recent book is Imagining Criminology: An Alternative Paradigm.
Contact information: Department of Criminal Justice, California State University, San Bernardino, 5500 University Parkway, San Bernardino, CA 92407; phone: 909.880.5552; e-mail: fwilliam@csusb.edu.
Marilyn D. McShane
Marilyn McShane is a professor of criminal justice and author or coauthor of books on criminological theory, community corrections, and prisons. Her most recent research has been related to parole prediction instruments.
Contact information: Department of Criminal Justice, Northern Arizona University, Flagstaff, AZ, and Prairie View A & M University, Texas; e-mail: Marilyn.Mcshane@nau.edu.
Ronald L. Akers
Ronald Akers is a professor of criminology and sociology and director of the Center for the Studies in Criminology and Law. He has authored several books on criminological theory and is well known as the major proponent of social learning theory in criminology.
Contact information: Department of Sociology, University of Florida; e-mail: rla@soc.ufl.edu