Evaluation Reporting:  True Stories, Well Told

In a recent, themed issue devoted to the topic of validity in program evaluation, the journal New Directions in Evaluation (No. 142, Summer, 2014) revisited and commented on Ernest House’s influential 1980 book, Evaluating with Validity. House then argued that validity in evaluation must not just be limited to classic scientific conceptions of the valid (i.e. empirically describing things as they are), but that it must also include an expanded dimension of argumentative validity, in which an evaluation “must be true, coherent, and just.”  Paying particular attention to social context, House argued that “there is more to validity than getting the facts right.” He wrote that “…the validity of an evaluation depends upon whether the evaluation is true, credible, and normatively correct.” House ultimately argued that evaluations must make compelling and persuasive arguments about what is true (about a program) and thereby bring “truth, beauty, and justice” to the evaluation enterprise.

In the same issue of New Directions in Evaluation, in her essay, “How ‘Beauty’ Can Bring Truth and Justice to Life,”  E. Jane Davidson argues that the process of creating a clear, compelling, and coherent evaluative story ( i.e., a “beautiful” narrative account) is the key to unlocking “validity (truth) and fairness (justice).” To briefly summarize, Davidson argues that a coherent evaluation story weaves together quantitate evidence, qualitative evidence, and clear evaluative reasoning to produce an account of what happened and the value of what happened. She says that an effective evaluation—one that is truly accessible, assumption-unearthing, and values-explicit—enables evaluators to “arrive at robust conclusions about not just what has happened, but how good, valuable, and important it is.” (P.31)

House’s book and Davidson’s essay highlight how effective evaluations—ones that allow us to clearly see and understand what has happened in a program— rely on strong narrative accounts that tell a coherent  and revealing story. Evaluations are not just tables and data—although these are necessary parts of any evaluation narrative—they are true, compelling, and fact-based stories about what happened, why things happened the way they did, and what the value (and meaning) is of the things that happened.

Davidson writes:

“When I reflect on what has improved the quality of my own work in recent years, it has been a relentless push toward succinctness and crystal clarity while grappling with some quite complex, and difficult material. For me this means striving to produce simple direct and clear answers to evaluation questions and being utterly transparent in the reasoning I have used to get to those conclusions.” (p.39)

Davidson further observes that evaluation reports often are often plagued by confusing, long-winded, and academic jargon, that make them not only difficult to read, but obfuscate the often muddled and ill-reasoned thinking behind the evaluation process itself.  She argues that evaluation reporting must be clear, accessible, and simple—which does not mean that reports need to be simplistic, but that they must be coherent and comprehensible. I am reminded of a statement by the philosopher John Searle, ‘If you can’t say it clearly, you don’t understand it yourself’.

Reflecting on Davidson’s article, I realize that the best evaluation reports are the product of well thought-out and effectively conducted evaluation research, presented in a clear and cogent way. The findings from such research may be complex, but they need not be obscure or enigmatic. On the contrary, clear evaluation reports must be true stories, well told. To learn more about our evaluation reporting visit our Impact & Assessment reporting page.

Focus Groups

Pioneered by market researchers and mid-20th century sociologists, focus groups are a qualitative research method that involves small groups of people in guided discussions about their attitudes, beliefs, experiences, and opinions about a selected topic or issue.  Often used by marketers who obtain feedback from consumers about a product or service, focus groups have also become an effective and widely recognized social science research tool that enables researchers to explore participants’ views, and to reveal rich data that often remain under-reported by other kinds of data collection strategies (e.g., surveys, questionnaires, etc. ).

Organized around a set of guiding questions, focus groups typically are composed of 6-10 people and a moderator who poses open-ended questions that allow participants to address questions. Focus groups usually include people who are somewhat similar in characteristics or social roles.  Participants are selected for their knowledge, reflectiveness, and willingness to engage topics or questions.  Ideally—although not always possible—it is best to involve participants who don’t previously know one another.

Focus group conversations enable participants to offer observations, define issues, pose and refine questions, and create informative debate/discussions.  Focus group moderators must: be attentive, pose useful and creative questions, create a welcoming and non-judgmental atmosphere, be sensitive to non-verbal cues and the emotional tenor of participants.   Typically, focus group sessions are recorded or videoed so that researchers can later transcribe and analyze participants’ comments.  Often an assistant moderator will take notes during the focus group conversation.

Focus groups have advantages over other date collection methods.  They often employ group dynamics that help to reveal information that would not emerge from an individual interview or survey: they produce relatively quick, low cost data (they produce an ‘economy of scale’ as compared to individual interviews); allow the moderator to pose appropriate and responsive follow-up questions;  enable the moderator to observe non-verbal data; and often produce greater and richer data than a questionnaire or survey.

Focus groups also can have some disadvantages, especially if not conducted by an experienced and skilled moderator:  Depending upon their composition, focus groups are not necessarily representative of the general population; respondents may feel social pressure to endorse other group members’ opinions or refrain from voicing their own; group discussions require effective “steering” so that key questions are answered, and participants don’t stray from the questions/topic.

Focus groups are often used in program evaluations.  I have had extensive experience conducting focus groups with a wide-range of constituencies.  During my 20 years of experience as a program evaluator, I’ve moderated focus groups composed of:  homeless persons; disadvantaged youth; university pr ofessors and administrators; k-12 teachers; k-12 and university students, corporate managers; and hospital administrators.  In each of these groups I’ve found that it’s been beneficial to: have a non-judgmental attitude, be genuinely curious; exercise a gentle guidance; and respect the opinions, beliefs, and experiences of each focus group member.   A sense of humor can also be extremely helpful. (See our previous post: “Interpersonal Skills Enhance Program Evaluation,”  Also “Listening to Those Who Matter Most, the Beneficiaries” ) Or if you want to learn more about our qualitative approaches visit our Data collection & Outcome measurement page.

Resources:

About focus groups:

http://sociology.about.com/od/Research-Methods/a/Focus-Groups.htm

About focus groups:

http://www.cse.lehigh.edu/~glennb/mm/FocusGroups.htm

How focus groups work:

http://money.howstuffworks.com/business-communications/how-focus-groups-work1.htm

Focus group interviewing:

http://www.tc.umn.edu/~rkrueger/focus.html

‘Focus groups’ at Wikipedia

https://en.wikipedia.org/wiki/Focus_group

Understanding How Programs Work: Using Logic Models to “Map” Cause and Effect

A logic model is a schematic representation of the elements of a program and the program’s resulting effects.  A logic model (also known, as a “theory of change”) is a useful tool for understanding the way a program intends to produce the outcomes (i.e. changes) it hopes to produce.  Logic models typically consist of a flowchart schematic that shows the logical connection between a program’s “inputs” (i.e. invested resources), “outputs” (program activities and actions), “short-term outcomes” (changes), “medium-term outcomes” (changes), and “long range impacts” (changes).

When developing a logic model many evaluators and program staff rightly focus on inputs, outputs, and program outcomes (the core of the program).  However, it is critical to also include in the logic model the implicit assumptions that underlie the program’s operation, the needs that the program aspires to address, and the program’s environment, or context.  Assumptions, needs, and context are crucial factors in understanding how the program does what it intends to do.  Ultimately these are crucial to understanding the causal mechanisms that produce the intended changes of any program.

Without clearly understanding the causal mechanisms at work in a program, program staff may work ineffectively, placing emphasis on the wrong or inefective activities—and ultimately fail to correctly address the challenges the program intends to address.   Similarly, without a clear understanding of the causal mechanisms that enable the program to achieve its outcomes, the program evaluation may not measure the proper outcomes or fail to see the changes the program, in fact, brings about.

Brad Rose Consulting, Inc. works with clients to develop simple, yet robust, logic models that explicitly document the causal mechanisms that are at work in a program.  By discussing, and explicitly identifying the often implicit causal assumptions,  as well as highlighting the needs for the program and the social context of a program, we not only ensure that the evaluation is properly designed and executed, we also help program implementers to ensure that they are activating the causal processes/mechanisms that yield the changes that the program strives to achieve.

Other Resources:

1. “Monitoring and Evaluation: Some Tools Methods and Approaches, The World Bank.”

2. “The Logic Model Development Guide,” W.K. Kellogg Foundation.”

3. Logic Model Resources at the University of Wisconsin

4. A Bibliography for Program Logic Models/Logframe Analysis

Transforming “Data” Into Knowledge

In his recent article in the New York Times, “What Data Can’t Do” (February 18, 2013, visit here ), David Brooks discusses some of the limits of “data.”

Brooks writes that we now live in a world that is saturated with gargantuan data collection capabilities, and that today’s powerful computers are able to handle huge data sets which “can now make sense of mind-bogglingly complex situations.” Despite these analytical capacities, there are a number of things that data can’t do very well. Brooks remarks that data is unable to fully understand the social world; often fails to integrate and deal with the quality (vs. quantity) of social interactions; and struggles to make sense of the “context,” i.e., the real environments, in which human decisions and human interactions are inevitably embedded. (See our earlier blog post “Context Is Critical”.)

Brooks insightfully notes that data often “obscures values,” by which he means that data often conceals the implicit assumptions, perspectives, and theories on which they are based. “Data is never ‘raw,’ it’s always structured according to somebody’s predispositions and values.” Data is always a selection of information. What counts as data depends upon what kinds of information the researcher values and thinks is important.

Program evaluations necessarily depend on the collection and analysis of data because data constitutes important measures and indicators of a program’s operation and results. While evaluations require data, it is important to note that data alone while, necessary, is insufficient for telling the complete story about a program and its effects. To get at the truth of a program, it is necessary to 1) discuss both the benefits and limitations of what constitutes “the data”—to understand what counts as evidence; 2) to use multiple kinds of data—both quantitative and qualitative; and 3) to employ experience- based judgment when interpreting the meaning of data.

Brad Rose Consulting, Inc., addresses the limitations pointed out by David Brooks, by working with clients and program stakeholders to identify what counts as “data,” and by collecting and analyzing multiple forms of data. We typically use a multi-method evaluation strategy, one which relies on both quantitative and qualitative measures. Most importantly, we bring to each evaluation project our experience-based judgment when interpreting the meaning of data, because we know that to fully understand what a program achieves (or doesn’t achieve), evaluators need robust experience so that they can transform mere information into genuine, useable knowledge. To learn about our diverse evaluation methods visit our Data collection & Outcome measurement page.

Learning by Changing: Lessons from Action Research

What is Action Research?
Action Research is a method of applied, often organization-based, research whose fundamental tenet is that we learn through action, learn through doing and reflecting. Action research is used in “real world” situations, rather than in ideal, experimental conditions. It focuses on solving real world problems.

Although there are a number of strands of Action Research (AR) including: participatory action research, emancipatory research, co-operative inquiry, appreciative inquiry, all share a commitment to positively changing a concrete organizational or social challenge through a deliberate process of taking action, and reflecting on cycles of emergent learning. Kurt Lewin, one of the original theorists of AR said that: “If you want truly to understand something, try to change it.” Ultimately, Action Research is about learning though doing, indeed, learning through changing.

Collaboration and Co-Learning
Although Action Research uses many of the same methodologies as positivist empirical science (observation, collection of data, etc.), AR typically involves collaborating with, and gathering input from, the people who are likely to be affected by the research. As Gilmore, Krantz, and Ramirez point out in their article “Action Based Modes of Inquiry and the Host-Researcher Relationship,” Consultation 5.3 (Fall 1986): 161, “… there is a dual commitment in action research to study a system and concurrently to collaborate with members of the system in changing it in what is together regarded as a desirable direction. Accomplishing this twin goal requires the active collaboration of researcher and client, and thus it stresses the importance of co-learning as a primary aspect of the research process.”
(Retrieved from http://www.web.ca/robrien/papers/arfinal.html#_edn1)

Collaboration, Stakeholder Involvement, and Constructive Judgment for Program Strengthening
Brad Rose Consulting draws on the key ideas Action Research–collaboration and stakeholder involvement— to ensure that its evaluations are grounded in, and reflect the experience of program stakeholders. Because we work at the intersection of program evaluation (i.e. Does a program produce its intended results?) and organization development (i.e. How can an organization’s performance be enhanced, and how can we ensure that it better achieves it goals and purposes?), we know that the success of program evaluations depend in large part upon the involvement of all program stakeholders.

This means that we work with program stakeholders, (e.g. program managers, program staff, program participants, community members, etc.) to understand the intentions, processes, and experiences of each group. Brad Rose Consulting, Inc., begins each evaluation engagement by listening to clients and participants—including listening to their aspirations, their needs, their understanding of program objectives, and their experience (both positive and negative) participating in programs. Furthermore, we engage clients not as passive recipients of “expert knowledge,” but rather as co-learners who seek both to understand if a program is working, and how a program can be strengthened to better achieve its goals.

Ultimately we make evaluative judgments about the effectiveness of a program, but our approach to making such judgments is guided by our commitment to constructive judgments that help clients to achieve both intended programmatic outcomes (program results) and desired organizational goals. To learn more about our adaptive approach to evaluation visit our Feedback & Continuous improvement page.