WHAT IS NEEDS ASSESSMENT?
Communities struggle with many social problems and social service needs. Sometimes it is difficult to know which of the problems
isthe most important, which one should be at the
topof a list of priorities. Similarly, social service agencies recognize that their clients often have multiple problems—for some of these they may receive ser- vice, but for some problems no service may be available. Whether at the community or agency level, social workers are often involved in assessing the extent of problems people face and documenting possible unmet needs. Ideally, needs assessments should be conducted as the first step in responding to perceived needs. That is, needs assess- ments should be the first step in a program development process.
Needs assessments are attempts at estimating deficiencies. We can think of these deficiencies as unmet needs, gaps in services, or problems that have not been previ- ously recognized. Ickillip (1998) defines needs assessments as "decision-aiding" tools used for "resource allocation, program planning, and program development" based on the "assumption that planned programming can alleviate distress and aid growth"
(p. 262). Needs assessment is not a single methodology, but includes a wide range of approaches that will be discussed in this chapter. Sometimes known as "front-end analyses," needs assessments
aresimilar to market or feasibility studies.
For example, let's say that there are indications that drug abuse in Average
Town,USA, has been getting a lot worse. Key community leaders, parents, hospital staff, criminal justice officials, and school representatives all agree that something must be done about the drug abuse problems they see in the community. However, assume that it is not clear
tothis community how best to direct resources within that problem area. In` other words, should they address individuals involved in the crimi- nal justice system, who also have drug problems, for drug treatment? Should they
REASONS FOR CONDUCTING A NEEDs AssESsMENT ВОХ3R
1. To explore the extent of an ongoing problem with a client population or commu- nity; to determine whether there are enough clients/staff/community residents with a particular problem to justify creating a new program.
2. To prioritize the needs for services and help with resource allocation.
3. To determine whether other interventions and resources to address the problem exist.
4. To determine whether existing interventions are known to or are acceptable to potential clients.
5. To determine if there are barriers preventing clients from accessing existing services.
6. To obtain information for tailoring a program to a specific target population.
5б CHAPTCR 3
target kids in middle and high school with drug prevention? Or, should they focus on pregnant women who are abusing drugs and alcohol? Are more outpatient or inpa- tient programs needed? These are the kinds of questions that can be examined with a needs assessment.
Although as professional helpers, we may feel that we know the needs of our clients (or of certain neighborhoods and communities), this presumed knowledge is only sub- jective opinion until we obtain evidence of the unmet needs in our communities. Once needs have been documented, that information can be considered a baseline against which program implementation and outcome will be compared. Later the question can be asked, "Has XYZ program reduced the problem of drug abuse in our community?"
Needs assessments can provide valuable information for program planning, in- cluding what groups to target for services, the best ways to publicize or market services, estimates of the numbers of persons who could benefit from a specific program or ser- vice, information about the geographic distribution and sociodemographic characteris- tics of potential clients, and barriers that may encountered by clients (see Box 3.1).
For example, it may be important to ask questions like "How far will potential clients be willing to travel to reach the agency?" or "What hours would be most convenient for potential clients?" Thus, program planners and administrators might want to obtain potential clients' views regarding the agency's location and hours. ("Do clients view the agency as accessible and available?")
Similarly, an agency may already have a specific program designed, but the target
opulation may not be accessing the services as expected. For example, perhaps at one -Dint the waiting list was so long that potential clients no longer viewed that program -s a real source of help. Thus, a needs assessment might also target clients' knowledge 2 out the agency and its services in order to understand clients' awareness of services.
finally, an agency might want to know if community residents or potential clients view the agency's services as being of top or inferior quality. (This could be termed accept- ability of services.) In a world where budgetary concerns were not paramount, needs assessments would be conducted on a regular, ongoing basis to provide information to service providers and program planners to best respond to the needs of clients in their agencies and potential clients in the community. With such information, adjust- ments could be made as programs mature and evolve and target populations change.
DEFINITIONS OF NEED
Bradshaw (1977) conceptualized four types of need:
1. Normative need-a condition or situation defined by an expert. By looking at poverty figures for a given county
andthen the number of children enrolled in free lunch programs, an expert might recommend that additional steps be taken to inform families in a particular
schooldistrict where too few children are receiving meals. Normative estimates of need are strongly affected by the expertise and knowledge of the expert. However, normative estimates of needs can also be biased and nonrepresentative. In other words, only asking "experts" about needs may not accurately reflect the full picture of program needs because the
perspective is based on the opinions of only a few individuals.
2 Felt need—perceptions of need as reported by actual clients. Felt need data is collected by surveying clients either
inperson, by mail, or telephone. Each approach has drawbacks. For instance, clients may not want to tell an investi- gator what their most personal or private needs are. Mailed surveys
mayyield unrepresentative samples and low response rates.
3. Expressed need—a demand for service. For example, clients who have applied for services or received services in
agiven time period and those clients on a wait- ing list for service represent the demand for a particular program. The problem with this approach to defining need is that many clients who could benefit from services may not be receiving them because they do not meet eligibility guidelines, they do not have transportation, or they do not
knowwhere to apply for the services. Expressed need is always going to underestimate the true level of need
4. Comparative need—an inferred measure of need determined by examining the characteristics of those receiving services and then locating those characteristics
inthe population. For instance, a study of senior citizens living in one high-rise apartment building revealed that 12 percent needed transportation services.
This information might provide a useful estimate to apply to other high-rise apartment buildings populated primarily by senior citizens. This approach to defining need is vulnerable to problems of nonrepresentativeness. Clients of a particular program or agency (as well as those living in a specific apartment building) are, by definition, a self-selected sample. Their needs may or may not resemble those of other persons who have similar demographics (for example, age and ethnic group) in common.
From these different ways of thinking about need it should be obvious that conflict and disagreement can arise when individuals do not share the same definition of need. To take a ludicrous example, we all know intuitively when
we"need" some- thing like a sandwich or
asteaming hot cup of coffee. However, a neutral observer might take a look at us and conclude that, being 10 pounds overweight, we do not.
really need that fattening sandwich. The observer's recommendation might be that someone
elseneeds it more. Our physician might also take a contrary position and be of the opinion that we should cut caffeine from our diets—that we would be better off without that additional cup of coffee we so crave. Whose assessment of need is right? And
whois the best judge of need?
58 CHAPTER 3
One way to think about resolving differences between different views of "need" is to collect data from two different perspectives. That is, attempt to consider what a group of people might "want" versus what they really "need" by asking for both per- spectives and then subtracting the mean rating for "what is" from the mean rating for
"what should be." As an example of this gap-based needs assessment approach, Goldman and Schmalz (2007) provide an illustration of an item ("Parents teaching their adolescents about condoms") where respondents rated the "What should be" status as 2.07 (1= always, 2 occasionally, 3 = sometimes, 4 = never, 5 = don't know) and the
"What is" status as 2.78. Thus, there is
agap between these two positions of .71. This approach allows evaluators and planners to examine a variety of programs or activities and to prioritize them based
onthe size of the gap. For more detail see the Goldman and Schmalz (2007) article.
Evaluators who are conducting needs assessments should strive to be sensitive to different perceptions about need. This is an inexact science at best. However, when trying to determine what a client population needs, always try to involve clients in the process. Additionally, evaluators need to be mindful that there always are multi- ple stakeholders in a community. Stakeholders are all those who have an interest in
program and can include such groups as:
Funders Administrators Politicians
Community members Service providers and staff
Other programs and organizations that might be making or receiving referrals Businesses
Current, past, and potential clients
As Mika (1996) notes, each group of stakeholders can have a different per- spective on the program and may tend to interpret results from their own refer- ence point. Various groups may use the same data to argue that a program is a waste of public funds or a dire necessity-depending on whether they favor or oppose the program.
Needs assessment efforts can range from the simple to the complex. There is no one standardized approach. In some communities, it might make sense to start with the local United Way's resource inventories to determine what is missing in the way of programs and services. In other communities, where every imaginable service is available in some form, the task might be to examine issues of accessibility and avail- ability. A program that is available only in one neighborhood or only operates during selected daytime hours may not be of any benefit to potential clients without child care or transportation or those who cannot leave their employment in order to access services with limited hours. Needs assessments may also be contacted within an agency to shape the educational and training needs of its staff (see Cassidy, Rosen, Bolan, Walker, & Krinsky et al., 2005).
PLANNING A NEEDS ASSESSMENT
How would you go about evaluating whether there
issufficient need to justify the start of a new program?
Suppose you feel that there
is-a need for a latchkey program in your community.
You are particularly concerned about elementary school aged children who, because of working parents, are at home for several hours in the afternoon without adult supervision. You learn that a local foundation has expressed interest in funding a pilot latchkey program in your community during the next school year if it can be convinced of the need.
askyourself, "What information sources are available?" As you think about the information sources that would be helpful and obtainable, it occurs to you that among your friends are three elementary school prin- cipals.
Youcontact them and find that two are convinced that a latchkey program is needed, while the third is undecided. You
donot feel that this is sufficient information to take to the foundation. What more could you do? You could ask all of your friends and neighbors if they thought that this program was needed. Unfortunately, as one friend indicates
toyou, these opinions do not constitute objective information. Asking only people you know about their opinions will give you biased information-even if the number of people you have talked to is now up to 25.
What else could you do? If there is a "true" need for the latchkey program, it would be evidenced by parents who are interested
inhaving their children participate in the program. Their interest could be documented (with the principals' support) by sending home a brief questionnaire to every parent with elementary-school-aged children ex- plaining that a planning effort. is being conducted to gauge interest in
alatchkey pro- gram. When parents and guardians return the questionnaire, you will have objective information regarding the perceived need for a latchkey program. However, another information source could provide useful data. With the cooperation of the child protec- tion agency, you could survey the child protection staff in your community in order to learn if they, too, perceive the need for
alocal latchkey program.
As this example shows, much of what constitutes needs assessment revolves around thinking about what sources of useful information
youcould obtain. The emphasis is on useful. In this day and age, information from hundreds of sources is available and is as close as the nearest computer. But before gathering informa- tion, focus on the necessary steps in conducting a needs assessment (Box 3.2).
Although these steps may be similar for process and outcome evaluation as well as needs assessment, it is the focus on documenting needs that should be emphasized throughout each step.
Step 1: Define the parameters for the needs assessment. The more resources at your disposal, the larger your budget, and the more time
youhave, the more compre- hensive and sophisticated your needs assessment can be. Keeping the purpose of the needs assessment constantly in mind will help
youkeep your efforts focused. In addi- tion, you need to determine the level of your needs assessment: is it a statewide assess- ment, community, or neighborhood assessment? Current clients or potential clients?
Step 2: Identify the information needed for decision making. Many times key stake- holders or planning committees add specific questions or goals to the needs assessment that could provide interesting information. However, the test of whether these items
б0 CHAPTER 3
BOX З Ø
' STEPs IN NEEDS IASSESSMENT
Step 1. Clearly understand:
a. The purpose of the needs assessment
b. The level of assessment: Statewide, community, neighborhood
C. What stakeholders to include: Clients or potential clients, program staff, key community leaders, state officials, etc.
d. Budget and available resources e. Time allotted for the project
Step 2. Identify the specific information you need to acquire.
Step 3. Determine whether the information already exists or can be obtained with your resources.
Step 4. Design the methodology and instrumentation (if necessary).
Step 5. Collect and analyze the data.
Step 6. Prepare the report.
Step 7. Disseminate preliminary results to key stakeholders to obtain their feedback.
Step 8. Formally disseminate results.
should be included is whether they will provide information useful to program man- agers. If no direct use can be made of proposed information items, then they should not be added. What do you need to know?
Step 3: Determine what information already exists. Avoid "reinventing the wheel" or "spinning your wheels." Check with other social service agencies, look through past reports, talk with staff about the information they collect, and perhaps conduct some literature searches.
Step 4: Develop a needs assessment plan—that is, a methodology that will structure the data-gathering efforts. The methodology is a description of the proce- dures, the research subjects, key variables, and objectives in the proposed study.
Step 5: Collect the actual data. Depending on the design and scope of the effort, many different people could be involved with gathering the data. The coordinator of the needs assessment becomes, on occasion, a supervisor of others to ensure that important tasks are completed well and on time. The coordinator may also be engaged in data entry or data editing in addition to "number crunching" as the project moves into data analysis.
Step 6: Begin writing the needs assessment report. Multiple drafts of the written report may be necessary. The first rough draft simply gets the main ideas down. The sec- ond draft puts flesh to the bones. The third draft corrects spelling, grammar, and
lishes the written document. The more that is riding on this needs assessment, the
`amoreimportant it is that you have someone else proofread the final report for clarity .f expression, accuracy, and to protect you from any political blunders that it might contain.
Step 7: Share the preliminary results with key stakeholders. Not including prin- zipal stakeholders to provide feedback could be problematic for a number of reasons
(Amodeo & Gal, 1997). First, it takes time to conduct a needs assessment and the stakeholders may forge ahead without the results if they don't see the needs assess- ment's progress. Second, keeping key stakeholders involved in the process will keep the needs assessment a central focus, which is important when agencies are busy with other concerns. Third, it is important to involve key stakeholders in interpreting the results and to obtain their input into how the results are presented. For example, key stakeholders may have questions that are easily answered but the information is
notcontained in the report. It is also important that recommendations generated from the report are viewed as credible by the stakeholders and the key agency per- sonnel This will help to ensure utilization of the needs assessment results.
Step 8: Disseminate the results to interested and sympathetic parties. There is more information about disseminating results in the last chapter. However, it is probably important to write a report with an executive summary as well as to make a formal presentation of the final results to key stakeholders and, possibly, to the larger community or public as well.
SELECTING A NEEDS ASSESSMENT APPROACH WHAT RESOURCES ARE AVAILABLE?
There are a variety of ways to go about estimating the need for a human service program. And numerous questions and issues will affect the particular needs assess- ment strategy you select. Besides your budget, the resources available to you, and the amount of time you have to finish the project, you need to consider whether this assessment
willbe a one-time phenomenon
oran effort that will be built on—
perhaps even repeated each year. Is the purpose of the needs assessment to satisfy some bureaucrat in the state capital, or will the data really be used by the agency?
Is it seen as "busywork" or as a useful activity? Is your supervisor or agency direc- tor anxious to see the needs assessment, or are your instructions to just put "some- thing" on paper? Is it likely that the needs assessment will be used by others in the community, or will it simply be submitted to some government office and promptly forgotten?
How much can be spent on the needs assessment? Can you afford consultants and paid interviewers for
yourcommunity survey? What kind of technical expertise or staff resources will be available from your agency or cooperating agencies? If you are creative, low-budget approaches to needs assessment can be found. Stefl (1984), for instance, reported on a community survey that volunteers conducted. The way some of these volunteers were recruited is interesting. One of the agency's board members was a probation officer, and he was able to offer community service to a select group of offenders as an alternative to incarceration. These persons were screened very closely, trained as interviewers, and were said to have performed very well. They obtained 822 telephone interviews in 21 days. The overall refusal rate was comparable
tothose reported by professional survey organizations.
Agency staff are also resources. Besides yourself, who else can be asked to assist with the needs assessment? What skills
Planning a community survey obviously takes more time than contacting a hand- ful of key informants. If you are working against a rapidly approaching deadline, your
62 CHAPTER 3
choice of a needs assessment approach may be justifiably influenced by what can be accomplished in a short period of time.
Closely related to the resources available for the needs assessment is the issue of the amount of detail or information desired—whether the assessment has to be objec- tive and data-oriented or whether it can be "softer." How "hard" does the data need to be? Will your audience be skeptical or eager to accept your findings? If you cannot directly assess need, what surrogate measures are available to you—what programs are most similar to the one you must assess? How was need for those programs deter- mined? The amount of time you have and what information your employer or sponsor wants collected influences the choice of methodology. Deciding on the right design is not always a left-to-right, linear process.' Planning a needs assessment may require simultaneously considering resources, deadlines, information desired, and so forth.
The next section discusses several different approaches available for assessing needs.
SECONDARY DATA ANALYSIS
Secondary data refers to existing information that comes from census data, public documents, and reports. Even data generated by other researchers or surveys can be reexamined for relevance to the new program. Census data, for instance, con- tains a wealth of information, and because it is readily available
inmachine- readable form and from the Internet (www.census.gov), it should be reviewed before collecting any other data. At a minimum, census data can provide you with estimates of the population likely to need or to benefit from the program you are proposing.
Census data are available for geographical units known as census tracts and census blocks. (Note that the block data are available only for large metropolitan areas.) By referring to census data, it is possible to learn how many school-aged children reside in a defined geographical area. You can learn the race and sex of these children, and the number living in poverty. There is even a category that pro- vides information on the number of females in the labor force with children under age 6, and between 6 and 17 years of age.
To switch examples for a moment, census data can also be used to provide estimates of the number of persons 55 and older, the areas within the community where these older adults tend to reside,. and the number of older adults living in poverty. Census data can also be used to provide such information as the general level of affluence in a community, the average level of educational attainment, the number of substandard dwellings, and the number of persons with work disabilities.
In order to protect the confidentiality of information supplied to it, the Census Bureau suppresses data that could be used to identify specific individuals or fami- lies. Census data cannot be used to gain personal information on a specific family or families. It cannot supply you with the names, addresses, or phone numbers of families having school-aged children or living in poverty. You can, however, use census data to plot a map of those areas
inthe community that have the highest concentrations of older adults, or children of families living in poverty.
In addition to census data, every state maintains a wealth of useful data for planners and evaluators. It is possible to learn from the state health department information
such as the number of births, marriages, deaths, and suicides that occurred in a county in a given year. If you were developing a prenatal program for teenage mothers, it would be possible
tofind both the number of babies born
toteenage mothers and the number of infant deaths in the years prior to the start of the program. Persons interested in starting an alcoholism prevention or treatment program may want to document the number of persons who have died as a result of cirrhosis of the liver. You may find studies that have used drug and alcohol-related arrests, drug and alcohol-related hospital discharges, drug and alcohol treatment admissions or even other variables to estimate substance abuse treatment gaps for specific areas.
These and maný other categories of information are available through various local, state, and federal agencies. (These variables and others that help gauge the extent of social problems are known as social indicators.) For instance, from the state department of education you can find such information as the number of school dropouts, the number of ninth graders reading at grade level, and school enrollments. Other state departments keep records of such social indicators
asthe number of children receiving food stamps, medically indigent children, free
reduced-cost school breakfast recipients, child abuse allegations, substantiated abuse allegations, delinquency cases, unemployment, psychiatric admissions
In addition to state-produced reports, a little Internet searching may uncover national reports that provide detailed information on
astate-by-state basis. For ex- ample, the Annie E. Casey Foundation annually publishes the Kids Count Data Book, which compares states on a number of variables, such
asinfant mortality rate, percentage of births to single teens, juvenile violent crime arrest rate, percentage graduating from high school,
andpercentage of children not living with a parent. It is also available online (www.aecf.org).
Information from over 70 federal government agencies that produce statistics can be located at www.fedstats.gov. A good many of these agencies provide . their own search engines to make it easier for you to find the specific statistics you are seeking. Most human services agencies report annually on the characteristics of those who have been clients in the past year. These data can be reviewed to see what groups within the
communityare being served (and underserved).
Administrative records, reports, and files within your own agency are a source of already existing data that should not be overlooked for needs assessment pur- poses. Table 3.1 shows how the data from one counseling agency could be used for needs assessment purposes.
From this table, we can identify potentially underserved segments of the community based on the numbers of clients who have received service. The table shows greater de-
mandfor the adult program than the children's program and more usage of it than the older-adult program. These figures could be used to understand expressed need-that is, official requests for service. We can also
seethat the drug abuse treatment program appears to need additional staff. In 2007, almost 25 percent of their current caseload
wasawaiting service and the need for services grows about 8 percent over the three years. Clearly, this program is in need of additional staff or resources in order to reduce to an acceptable level the number of clients waiting for service, None of the other pro- grams had so many clients awaiting service
orhad the increased need for services over time like this particular program.
б4 CHAPTER 3
CLIENT UTILIZATION DATA, PUBLIC COUNSELING SERVICES, INC.
2007 2008 2009
Children served 363 383 407
Number on waiting list 16 19 21
Adults served 785 791 818
Number on waiting list 14 12 15
Older adults served 63 72 84
Number on waiting list 0 ' 3 9
Drug abusers served 302 414 545
Number on waiting list 75 124 183
Total clients served 1,513 1,660 1,854
Clients on waiting list 105 158 228
Client data can also be used for such purposes as locating neighborhoods or streets with the highest prevalence rates of certain problems (e.g., drug abuse). Multicolored pins representing client families can be placed in maps to help staff focus their outreach and education activities. Street addresses from client admissions over the past year can also be plugged into GPS software to produce a map of the city or county revealing pockets of the county that are underserved.
Secondary data sources are generally convenient to access and easy to under- stand and use. Anyone can rank counties or census tracts in terms of those having the most or least of some characteristic. Anyone can identify the county with the highest unemployment rate or determine what the unemployment rate has been in a selected county for the past 5 years. In metropolitan areas, census tracts or blocks may be ranked in terms of percentage of families living in poverty or number of older adults.
There are, however, a few drawbacks that are associated with secondary data.
The data may be outdated, unreliable, incomplete, or from an agency that is not really similar (because it is in a different geographical region, serves a different pop- ulation, or has different eligibility guidelines). Even with the best agency data, there is always the problem of counting those potential clients who qualify for services but who have never been referred and who do not self-refer. Anytime you rely solely on client utilization data, you are likely to be underestimating the problem to some extent.
Last, using social indicators to estimate the health or social conditions of persons living in a certain geographic area can, at times, result in a misinterpretation known as an ecological fallacy. The fact that a neighborhood with colossal rates of drug overdoses is also highly populated with a particular ethnic group does not mean that every member of that ethnic group is going to overdose, or even that the ethnic group has the highest rate of drug overdose. It is possible that the absolute number of
inthat neighborhood may be elevated because of the extremely high rates by other ethnic populations, thus giving a misleading impression.
After you have consulted the census data or other secondary data and have some be- ginning data, additional information can come from consulting with service providers and other key informants. Key informants are those persons who are informed about a given problem because of training or work experience—
usuallybecause they are involved in some sort of service with that population. In an after-school program, key informants could be the principals, guidance counselors, social workers, and teachers. One person conducting a key-informant needs assessment for the after- school latchkey program could easily contact all of these personnel in a single ele- mentary school. Key informants could also include child protection workers and their supervisors, and include the parent-teacher organization and others.
Impressionistic approaches have a subjective quality to them. That is, these ap- proaches are not as accurate or scientific as large-scale community surveys. Why not? For one thing, the sample sizes are often too small to be representative of the larger population. Think about the situation where the needs assessment involved talking with three principals. Even if the needs assessment were expanded to include three teachers, three single mothers, and three leaders from the school's parents' or- ganization, we still would not have a sample necessarily representative of the opin- ions of all the principals, teachers, and parents in the community. Our data would not be scientific—especially if we chose these individuals because we knew them.
(We will discuss sample size and representativeness more fully in Chapter 8.) Although the opinions of these people
maybe well founded and based on a superb knowledge of the problem, they may also be based on nothing other than personal bias, beliefs, or values. Suppose, for example, that two of the principals
youselected strongly believed that women should not beemployed outside of the home. These principals may be less likely to acknowledge the need for
alatchkey program than a principal with more egalitarian values.
A cynical view of a key-informant type of needs assessment might caution against warmly embracing the perceptions of key informants without first examining what they stand to gain or lose by
over-or understating a particular problem. Thus, a principal might be expected to be in favor of a latchkey program if it makes his or her life easier—even if there does not appear to be great need for such
aprogram in the community. There could, for instance, be one or two "problem children" who must be watched after school because their parents are always late in picking up their children, which
causesteachers to complain to the principal. The problem of dealing with subjective opinions is also present (and perhaps more visible) in another type of impressionistic needs assessment.
Public hearings and community forums are a type of needs assessment that are grassroots oriented. What is more democratic than acquiring a public meeting room and posting
anotice or advertising that anyone concerned with the problem of (fill in the blank) is invited to attend and share their concerns? This approach has the advantage of being reasonably inexpensive, not requiring a lot of preplanning, and again, needing little research expertise to interpret or summarize the results.
6б CHAPTER 3
There are also some serious difficulties with public hearings. For one, the "public"
seldom seems to attend. Unless the issue is a controversial one, rarely would potential clients attend the public hearings that are supposed to generate planning data. Often, the only attendees are the planning staff and a few service providers from other agencies who have an interest in working with that specific population..
A second problem with community forums and public hearings is that even when citizens from the community attend, there is no guarantee that they represent the larger community. Sometimes certain interest groups can "pack" the meeting so that the opinions of others are not represented. Numerically small but vocal groups can dominate meetings. And persons most in need of the proposed service (e.g., families in poverty, juvenile delinquents, teenage parents) probably will not be in attendance at all.
A third problem that can result from impressionistic approaches such as public hearings is that because they are usually focused ori a single issue, other major prob- lems or needs in the community may not be addressed even though more lives or citizens might be affected.
NOMINAL GROUPS, THE DELPHI TECHNIQUE, AND Focus GRouPs
Several other impressionistic techniques provide good information from small groups.
The nominal group technique (Delbecq, Van de Ven, & Gustafson, 1975) involves a small group of persons who, in response to a common problem or question, work independently at first, and then share their ideas. The group leader asks each person to offer one idea in round-robin fashion. These ideas are recorded in front of the group on a chalkboard or a large sheet of paper. This process continues until all new ideas are exhausted. A discussion period follows when participants can elabo- rate, eliminate, combine, and add new ideas to the list. Next, each participant pri- vately ranks the five most important ideas from those remaining on the list. The group compiles the individual rankings in order to arrive at the most popular ideas or solutions to the question posed. The group then discusses the anonymous rankings to resolve any misunderstandings. After the discussion, group members are asked to give a final independent rating.
The Delphi technique (Delbecq et al., 1975) involves the use of a question- naire that is distributed to a panel of key informants or experts. (They do not meet together in person and may remain anonymous.) Their ideas are solicited, and their replies are compiled. If there are areas of disagreement, a second ques- tionnaire is developed based on the responses. This new questionnaire is sent to the panel, and their opinions solicited once again. This process continues until consensus is reached in all areas. For more insight into this approach, see Hung, Altschuld, and Lee (2008). The authors have recently discussed what they learned in conducting an electronic Delphi study associated with an educational program evaluation.
Focus groups represent another perception-oriented approach to assessing needs.
Although focus groups originally evolved from market research, their popularity has grown greatly in recent years to the point where it is not at all uncommon for social
service agencies to make use of them for needs assessment, program evaluation, and so forth.
Focus groups usually involve six to eight individuals
whoparticipate in structured discussion (Krueger, 2000). For needs assessment, a moderator facilitates a dialogue with constituent members of a client or target group. The goal is not to have these persons arrive at consensus but
toidentify and delineate their particular needs—some of which will be common to all group members and others unique to one individual.
The moderator obtains in-depth information by probing and asking clarifying ques- tions. For recent examples of literature describing the use of focus groups for needs assessment, see Henderson, Bainbridge, Keaton, Kenton, Guz, and Kanis (2008) and Chernesky and Gutheil (2007) or Buckelew, Pierrie, and Chabra (2006). Also, see Chapter 4 for additional discussion on focus groups.
Although impressionistic approaches have much to recommend them (they in- volve the community, are inexpensive, relatively quick to implement, and require no special knowledge of needs assessment), it is difficult to assess the accuracy of the obtained data. Those who are invited or chosen, or who elect themselves, to partic- ipate may not be truly representative of the larger community. Their views may be atypical and not reflect those of the majority. If this is
'amajor concern, a commu nity survey would provide less biased and more accurate information. Perhaps the best use of impressionistic approaches is to add the "personal" or "human angle"
to those approaches that have relied heavily
Surveys are familiar to most of us. Businesses use surveys to learn why we choose the brand of toothpaste that we buy; politicians use them to identify who will vote for them in the next election. Social scientists use surveys to determine the preva- lence of such social problems as elder abuse. Surveys can also be used for program development and needs assessment purposes.
Surveys are exceptionally valuable tools to use for needs assessment. Although they require more planning and resources than the impressionistic approaches, they provide information that is much more objective and scientific. When a careful prob- ability sampling design has been used to ensure a representative sample, it is possible to talk very precisely and confidently about the extent of a problem in a community.
You could find, for instance, that only 42 percent of the respondents knew that there was a mental health center in their community. Similarly, what if only 24 percent knew that counseling could be obtained for children who were not doing well in school or getting along with their families? Because the sample was derived from a probability sampling design, the researcher was 95-percent confident that the results were accurate within plus or minus 5.5 percent. That is the type of accuracy that the other needs assessment approaches cannot provide.
On the other hand, a large representative survey may not be possible or necessary.
If you are interested in a specific population you may need to survey representatives from that population. For example, Acosta and Toro (2000) surveyed a sample of 301 homeless adults in New York two times over a period of 6 months. Their sample was recruited from homeless shelters, soup kitchens, inpatient hospitals, outpatient
68 CHAPTER 3
facilities, and homeless service agencies. These researchers found that although housing was rated as an important need, participants rated other needs as at least as important as housing, including safety, education, transportation, medical/dental care, and job training/placement. Many of these needs were also identified as difficult to obtain.
They also found that needs for formal mental health and substance abuse services were rated as relatively unimportant and easy to obtain, although satisfaction with the services in these areas was low.
As another example, you may be interested in assessing training needs of staff for a particular agency or organization. One study surveyed a random sample of substance abuse treatment providers working in New England substance abuse treatment facilities (Hall, Amodeo, Shaffer, & Bilt, 2000). These researchers found that although partici- pants reported limited previous training and barriers to current training opportunities, the respondents had high levels of knowledge and skills regarding substance abuse treatment. However, providers did identify a need for more training on assessment strategies, more advanced clinical techniques, and dual diagnosis and treatment.
Regardless of whether you use a probability sample of households or a targeted sampling strategy, the precise estimate of the community's needs, beliefs, values, or behavior can come about only when the survey methodology has been sound. This type of needs assessment requires knowledge of both research methodology and sampling procedure. And, you must consider trade-offs or strengths and limitations in whatever sampling strategy you choose. For instance, a larger sample will give you a lower margin of error but will also cost more to complete.
Although it is possible to conduct surveys with persons who are near at hand and easy to access, those who are chosen merely because it is convenient may not adequately represent the community. For instance, one could choose to survey par- ents who have children on the school's swim team to see if they would support the school district raising taxes in order to constuct a new swimming pool. Their views, however, may not represent the opinions of all the parents with children attending the school or even those of others in the community living on fixed incomes and whose children are grown.
Biased samples might result when surveys are distributed at meetings and some in- dividuals might not be able to attend because they lack transportation. Additionally, others, especially single parents, have the inconvenience of having to arrange for babysi
tters—still another financial burden for some households.
In order to be representative of the whole community, every person in the pop- ulation must have an opportunity to provide input. With small populations it may be necessary to contact everyone, or at least a majority, to have a representative sample of the population. With large populations, random sampling can be em- ployed so that perhaps less than 10 percent of the population is contacted (but ev- ery person in the population still had an equal chance of being selected to provide their opinions).
Although we are going to spend more time discussing sampling later in Chapter 8, you can understand the importance of representativeness if you think about a large metropolitan community of about one million persons. Suppose a friend of yours from another country is interested in the quality of life as perceived by persons in this country who live in large cities. Your friend (who knows nothing about sampling) asks you to send her the address of one person from your metropolitan community so that
she can mail this person a questionnaire. Can any one person adequately reflect the diversity of opinions, experiences, and lifestyles that are found in large cities?
What if your friend later asks for
youto send the addresses of two more people?
alarge metropolitan area be represented by the opinions of three persons?
What if you sent 30 addresses? How many addresses
wouldyou have to send
order for this sample to be representative of the opinions held by the majority of per- sons living in your metropolitan urea? The return or response rate is also important.
A low response rate is the equivalent of inviting a large number of people to a party and only one
ortwo showing up.
The purpose of the previous illustration is to show that an evaluator needs to have a good grasp of sampling before beginning a community survey. There is noth- ing wrong with conducting small convenience surveys of 20, 40, or even 50 respon- dents as pilot studies to provide for some beginning estimates of need. However, remember that unless all the members
arecontacted or have an op- portunity to be selected, it will not be possible to assess the accuracy of the results.
Unfortunately, even if you go to a
toobtain a random sample, you still may end up with biased results.
For instance, Thompson, Ruma, Authier, and Bouska (1994) mailed a community needs assessment questionnaire to every household in an Iowa county
(n =1,850) and got a response rate of only 23 percent. They found that those who returned completed surveys were largely college-educated people who were middle-aged and in the upper- income categories. Young adults and lower income groups were the most clearly underrepresented.
Epidemiologic surveys seek to learn the extent of problems (usually diseases and injuries) in a community and typically express these as rates within the popula- tion. Epidemiologic surveys must, by definition, be based on scientific, representa- tive samples and for this reason tend
tobe much more expensive than smaller scale, more informal surveys.
Ciarlo, Tweed, Sherri, Kirkpatrick, and Sachs-Ericsson (1992) have reported that a comprehensive psychiatric survey in Colorado generating need estimates for just 6.4 percent of the population cost over $700,000, or approximately $150 per respondent.
These costs did not include questionnaire design, computer editing, and data analysis costs. At this rate, it would have cost $11 million to survey the whole state.
Done properly, large-scale community surveys are more expensive, more time consuming, and tend to require more research expertise than secondary data analy- sis or impressionistic approaches. But they offer a level of precision and confidence not found with the other needs assessment approaches.
CONVERGENT ANALYSIS AND MULTIMETHOD APPROACHES
Because any one approach to needs assessment may provide a somewhat incomplete picture of the true need for a program, convergent analysis should be the focus of the needs assessment effort (Siegel, Attkisson,
&Carson, 1978; Warheit, Bell, & Schwab, 1977). Convergent analysis involves using multiple sources of information and at- tempting to explore the need for the program or confirm information about the program by means of different assessment strategies and perspectives. One can also
70 CHAPTER 3
attempt to converge findings by involving various stakeholder groups. For instance, Hiatt, Stelle, Mulsow, and Scott (2007) obtained multiple perspectives concerning hospice care. They report involving physicians, family members, patients, volunteers, bereaved family members, staff, and community stakeholders in face-to-face inter- views and focus groups.
When needs assessment data are obtained from more than one source, areas of agreement may not always be immediately identifiable. For that reason, needs assessors attempt to converge the data by looking for patterns and areas of agreement. This strat- egy is similar to a process in navigation and surveying called triangulation where multiple reference points are used to locate an exact position. Information from various sources is integrated and synthesized to provide a "reasonably viable portrait" of the community's perceptions (Nguyen, Attkisson, & Bottino, 1983, p. 104). Including key stakeholders and key agency personnel to help interpret the results can be very important.
Because all needs assessment approaches have conceptual, empirical, and inferential problems (Fiorentine, 1994), experts often recommend a multimethod approach to needs assessment in order to reduce the underlying problems found with a single method.
A CONVERGENT, HULYINEYHOD ILLUSTRATION
Let us imagine that as
aschool social worker you first became convinced of the need for a latchkey program when you learned of an 8-year-old child who started a fire
his bedroom and barely escaped serious injury. Because both parents were at work and the child had been regularly without adult supervision from 3:00 P.M. until 5:30 P.M., legal and child protection authorities had become involved.
As you talk about your idea of
alatchkey program with several elementary schoolteachers during lunch hour, they become excited and each names about four children who could benefit from such a program. The school principal agrees that a latchkey program is needed and suggests that
youtalk with the parents' organization scheduled to meet the next evening. The parents' organization wholeheartedly en- dorses the concept and asks the principal if a questionnaire can be sent to every child's home. The principal agrees. A small planning committee meets with you and designs a questionnaire that looks something like the one in Figure 3.1.
The needs assessment questionnaires are prepared and given to each child in the elementary school to take home. About 60 percent of the questionnaires are re- turned. The results are as follows:
Q1: Twenty percent of the parents would enroll their children in an after- school program if there were no charge. A total of 105 children would be expected to participate if there were no charge for the after-school program.
Q2: Ten percent of the parents would enroll their children in an after-school program if there were a $25 a week charge per child. Approximately 50 children could be expected to participate if there were a $25 a week charge per child.
Q1 and Q2: Thirty percent of the parents were undecided about enrolling their children and wanted additional information.
Each of the informational sources explored in this fictitious example leads us to believe that there is a definite need for an after-school program. The need for a
FOXGLOVE PARENTS' ORGANIZATION AFТER-SCHOOL CARE QÚESTIONNAIÀE FIGURE 3.i
Because of the recent fire in our Foxglove community and the narrowly averted tragedy, we believe that there is a need for an after-school program. Our children would be supervised at school by teachers. Tutoring, games, and special "fun"
classes could be arranged—if there is sufficient interest from the parents. The school board may agree to pay several teachers for two hours each day after school if sufficient need can be documented. Please take five minutes to complete the following survey and have your child return it tomorrow to his or her homeroom teacher:
1. If an after-school program were available January 15 and there were no charge for enrolling your child, would you enroll one or more of your children (kindergarten through sixth grade)?
— YES, I WOULD ENROLL — (number of children) _ NO, I WOULD NOT ENROLL ANY OF MY CHILDREN
— UNDECIDED, I NEED MORE INFORMATION ABOUT THE PROGRAM 2. If the school board does not have sufficient funds and there is a charge of
$25.00 per week for each child, would you still make use of an after-school program?
— YES, I. WOULD ENROLL (number of children)
— NO, I WOULD NOT ENROLL ANY OF MY CHILDREN
— UNDECIDED, I NEED MORE INFORMATION ABOUT THE PROGRAM з. If you want to make sure that we reserve a place for your children, please
write your name and address below. However, please return the question- naire whether or not you want us to reserve a place at this time.
latchkey program could be further supported by including secondary data such as the number of calls to police or rescue squads by unsupervised children. Such a needs assessment would make a strong and compelling argument for the proposed latchkey program. However, not every needs assessment could be expected to find such a high degree of convergence.
For instance, it is entirely possible that the parents organization may not feel that a latchkey program is needed. Or, perhaps the principal, cognizant of the increased costs in utilities, janitorial services, and staff for the program, is not supportive be- cause of an inadequate school budget.
There are no easy solutions as to what you should do if your sources of infor- mationdo not converge. Sometimes you can explain away the lack of support from one sector (for example, the parents' organization is not representative of those parents who might make use of the latchkey program). On the other hand, the principal's concern must be addressed even if everyone else you contacted thought the project was viable.
Convergent needs assessment is a model that strives to include multiple sources of data. These efforts are more comprehensive than other forms of needs assessment, and
ANEEDS ASSESSMENT PRACTICE SAMPLER
I. A county health agency surveyed high school, middle school, and community youth (n = 1,567) to learn what these young people knew about HIV/AIDS and the sources of their information.
Seventy-five percent of the middle school sample but only 58 percent of the high school sample indicated they got information about НIV/AIDS from their families.
In response to the question, "If someone were to talk to you about AIDS, who would you be more likely to listen to?" 24 percent of the middle school youth but only 7 percent of high school youth chose the category "Parent." Both groups reported they would be more likely to listen to a person with AIDS than to anyone else.*
II. To find out what the greatest needs of fathers were in San Mateo County, California, 1200 surveys were distributed to 16 agencies and organizations. About 20% were re- turned (n = 240) and nine focus groups were also conducted. Of those responding to a question asking them to prioritize their greatest needs, finances was first, followed by health care, shelter/housing, and food. The authors concluded that it was important to increase the visibility of existing programming.**
III. Homeless mentally ill clients' perceptions of service needs were compared to those of their service providers. At the time of referral to case management services, outreach workers identified on the referral form which of 17 services they felt clients needed and then selected the 3 they thought were most important for that client. Clients were asked to identify which of the same 17 specific services they needed.
Seventy-four percent of the providers but only 44 percent of the consumers thought mental health services were most needed.
Twenty percent of the providers but only 8 percent of the clients thought substance abuse services were most needed.
More clients than providers thought job assistance was needed.* * *
*Shields, G., & Adams, J. (1995). HN/AIDS among youth: A community needs assessment study.
Child and Adolescent Social Work Journal, 72(5), 361-380.
**Buckelew, S.M., Pierrie, H., & Chabra, A. (2006). What fathers need: A county-wide assessment of the needs of fathers of young children. Maternal and Child Health Journal, 10(3), 285-291.
***Rosenheck, R., & Lam, J. A. (1997). Homeless mentally ill clients' and providers' perceptions of service needs and clients' use of services. Psychiatric Services, 48(3), 381-386.
72 CHAPTER 3
the many "reference points" give them more credibility. At the same time, they are more complex in the data collection phase as well as in the analysis. Box 3.3 provides some ex- amples of needs assessments that involved more than one sample of respondents.
THINKING CREATIVELY ABOUT NEEDS ASSESSMENT
Needs assessments do not have to be terribly complex. Sometimes the simplest of docu- mentation procedures provides useful data for estaЫishing that there is a need for a new program or facility. Once, after lecturing to a class about needs assessment, a student
who believed that she did not know enough about needs assessment to do an assignment came to my office to talk. She worked at a diversion program for juveniles who had been arrested and she mentioned a desperate need that she encountered every day. There
notenough temporary shelter for status offenders (young people who were picked up by the police for running away, being out too late, or being in possession of alcohol). Because of the lack of suitable shelter, young people were placed in jail until beds became available elsewhere.
I asked her how many times a month the jail was used inappropriately. She said about 18 to 20 times. I asked if she could document this, and she indicated that it would be easy to do because a special form had to be completed each time. In a few minutes, we outlined an approach she could use to convince county and state offi- cials about the need for additional shelter beds for juveniles with status offenses.
In another example, a community mental health agency was required by a state agency to conduct a needs assessment. However, practically
nofunds were avail- able during the fiscal year. The agency improvised using data that was available from other counties. This is demonstrated in the next illustration.
A NEEDS ASSESSMENT ILLUSTRATION
Bluebird County (population 99,570) hired a consultant to conduct
acommunity survey using a standardized instrument derived from earlier studies. This instrument contained scales from which could be inferred the extent of the population having either
aprobable need for mental health services. Various scales made up the instrument, but for the purposes of illustration, the relevant data from the needs assessment effort have been simplified in Table 3.2
A short time later, River County (population 90,831) also retained the consultant to conduct a needs assessment of their county. Once again, a probability sample was obtained, and the same instrument used. The agency executive in adjoining Franklin County wanted very much to have a needs assessment conducted in his county, but
TABLE 3.2 BLUEBIRD COUNTY NEEDS ASSESSMENT DATA
Scale Counseling Percent
Anxiety Possible 6.2
Depression Possible 8.5
Psychosocial Possible 3.6
74 CHAPTER 3
TABLE 3.3 I ESTIMATED NEEDS IN FØKLIN COUNTY
Scale Need Bluebird Co. River Co. Franklin Co.
Population 99,570 90,831 85,422
Anxiety Possible 6.2% 6.9% 6.5%
Probable 5.3%о 5.1% 5.2%
Total 11.5% 12.О% 11.7%о
Depression Possible 8.5%о 9.6% 9.0%о
Probable 4.5%о 5.1% 4.8%о
Total 1З.0% 14.7%о- 13.8%
Psychosocial Possible 3.6% 4.5%о 4.0%о
dysfunction Probable 5.5% 6.0% 5.8%
Total 9.1%о 10.5%о 9.8%
a severely limited budget could not be stretched to encompass a community survey.
However, he contacted the agency directors in the neighboring counties, and both were cooperative and shared the data produced from their needs assessments. With this information, it was possible to estimate needs that a similar study might have found in Franklin County (Table 3.3).
Note how similar the percentages are for Bluebird and River counties.
Although there are some minor variations, the percentages of persons in need of mental health services in both counties are almost equal. We can arrive at the es- timated number of persons in Franklin County in need of services by averaging the Bluebird and River County data. Thus, Franklin County would expect a slightly larger proportion of depressed persons than Bluebird County, but less than River County.
This approach uses survey data, but in its methodology it is most akin to the use of secondary data. Of course, the problem with this approach is that the data did not come from Franklin County. In actuality, 17 percent of the population of Franklin County might be depressed and 20 percent might score above normal levels of anxiety. We would not know the "true" level of these dimensions with- out conducting a probability survey in Franklin County. However, if a convinc- ing case for the similarities among these three counties can be built, then this type of estimation is as good as any of the other indirect approaches. How would we know if the counties were similar? We would begin by making comparisons of such variables as the average age of the population, average income per capita, the percentage of families in poverty, and the percentage of divorced and sepa- rated persons, and by looking at the racial and religious mix of the counties.
Sometimes it is relatively easy to know if two counties are similar or dissimilar.
If one county borders on a large metropolitan area and the other is rural and re- mote from any large city, then the two counties probably should not be compared.
If two counties are primarily rural, in close proximity to each other, and compare well
ondemographic variables, then it is reasonable to use them to estimate needs in a third similar county.
Even if you do not have access to needs assessment data from other counties, it may be possible to use social indicators or rates as developed by others. Ciarlo et al. (1992) recommend that other states consider using percentage of persons in poverty and percentage of divorced males as indirect measures for estimating total need for alcohol, drug, and mental health services. However, it should be remem- bered that such methods are only estimates and that substantial variations across different geographical areas can be expected. See also Herman, Wiesen, and Flewelling's (2001) description of using social indicators to estimate county-level substance use intervention and treatment needs.
Needs assessments do not have to be, financially burdensome to a human service agency. Some surveys (e.g., key informant studies)
canbe conducted without major expenditures of monetary or personnel resources. On occasions when a needs as- sessment
willcost several thousand dollars, cooperative efforts among social service providers or funders (for example, United Way) should be explored. Partial funding may also be possible with creative planning. For instance, in a needs assessment of mental health services in one community, additional questions were incorporated that asked respondents about their favorite radio stations (during day, evening, and "drive" time). Because advertising is based on the number of listeners, radio stations were interested in this data and several purchased that portion of the results dealing with their listening audiences and thus underwrote part of the cost for the study.
If you are given the assignment of designing a needs assessment, think creatively.
Shifman, Scott, Fawcett, and Orr (1986), for example, reported on the use of a game called "Family Few," modeled after the television program "Family Feud," to assess attitudes, beliefs, and knowledge about sexuality among female adolescents. In the process of obtaining a needs assessment profile on these adolescents, they were also able to provide didactic instruction!
Given the assignment of conducting a needs assessment, spend some time brain- storming all of the various ways one could go about examining the needs of the clientele. Make a list of these—whether they are feasible or not. When you run out of ideas, review the list and choose the best approach. Following is a list of sev- eral different types of needs assessments that one mental health agency conducted during a 6-year period. This list will give you some idea of the variety of ways in which needs assessment information can be obtained.
Community survey. Over 300 questionnaires were mailed to elected officials, school principals, attorneys, and other "key informants." In addition, over 500 questionnaires were mailed to randomly selected community respondents.
• Clergy survey. Over 100 questionnaires were mailed to clergy to ascertain their knowledge and perceptions of the community mental health system.
• Client utilization study. The characteristics of present and past clients of the mental health system were examined. Potential groups who were not being adequately served (e.g., minorities, low-income families) were of special
76 CHAPTER 3
interest. This client utilization study was felt to be so useful that it was subsequently prepared in an annual report for several successive years.
• Key informant study. Representatives from 35 human service organizations were contacted by phone and letter about their perceptions of the community mental health system.
• Community awareness survey. Over 300 respondents were contacted by telephone to discover the extent of their knowledge about the availability of local mental health services.
When the needs assessment is finished and it suggests there is a need for a specific type of program, another consideration is the community's readiness to respond.
For example, let's say that a needs assessment for a drug abuse prevention program examined such data as the number of teen drug users, teen pregnancies, and juvenile arrests. The community's rates of drug use and related problems were higher than national rates, and there is no drug use prevention program locally. Before a new pre- vention program is designed and implemented, it can be very important to know if the community will be receptive. Resistance can make the implementation difficult or possibly even cause the program to fail.
Because communities differ in their resources, strengths, challenges, and politi- cal climates, what works in one may not be fully endorsed in another. Fortunately, the community readiness model is a method for assessing a community's willingness to support a new program.
The community readiness model identifies nine stages or levels of readiness (letting et al., 1995; Plested, Smitham, Thurman, letting, & Edwards, 1999).
These stages are briefly described as follows.
1. No awareness. In this stage, community leadership tolerates the problem be- havior in certain contexts. The community climate may overtly or covertly en- courage the problem behavior. There are no formal or informal policies in place regarding the behavior.
2. Denial. In this stage, the community leadership recognizes that the behavior is or can be a problem, but not a local problem. Even if it is perceived as a local problem, there is a feeling that nothing can be done about it. The attitudes of community leaders are passive and apathetic.
3. Vague awareness. Community leaders in this stage generally recognize it as a local problem and that something ought to be done about it, but have no immediate motive to do anything. Ideas about who has the problem tend to be stereotyped and/
or vague. Policies exist but may be inconsistently followed. The leadership lacks energy or motivation for dealing with the problem, and the community climate does not motivate the leaders.
4. Preplanning. At least some leaders clearly recognize that there is a local problem and that something should be done about it. There are some identifiable leaders, maybe even a committee, but efforts are not focused or detailed. Policies are known and generally followed. Community leaders