This step on the eightfold path may not seem of obvious significance, when you focus only on the specific occupational categories that are to be avoided (1. dealing in weapons, 2. dealing in living beings (including raising animals for slaughter as well as slave trade and prostitution), 3. working in meat production and butchery, and 4. selling intoxicants and poisons, such as alcohol and drugs. Adapted from The Bigview).
This step is still of vital importance if we look at it in another context. The precepts of Right Livelihood are to support the practitioner engaging in Right Speech and Right Action. Similarly, we can develop a precept for Right Livelihood with respect to data analysis. We should seek to work for and with employers that support the practice of our profession in the right way. We should strive to promote environments where carrying out the principles embodied in Right Speech and Right Action become normative, and are not seen as threatening or disruptive. We should encourage the organizations that we are associated with to manifest these ideals in every aspect, not just in relation to their data analysis practices.
Tuesday, January 31, 2012
Tuesday, January 24, 2012
Right Action
Right Action involves some precepts, such as sexual behavior, which are beyond the scope of duties of the professional data analyst (at least I think so - if your experience is different you are welcome to share). But other precepts are relevant "to abstain from harming sentient beings, especially doing harm intentionally or delinquently; to abstain from taking what is not given, which includes stealing, robbery, fraud, deceitfulness, and dishonesty". adapted from The Bigview
Like Right Speech, we can also see this as building upon the concepts discussed in Right Vision. In work involving the public good, we are often called upon to balance the needs of society as a whole versus the needs of the individuals in society. In actions involving direct physical intervention with an individual, such as medical treatment, the harms or negative consequences are expected to be discussed, but in interventions that are less intrusive or are not aimed at specific individuals the quantification and discussion of harms, intended or unintended, is too often minimized. In order to make informed judgments and decisions, we should be prepared to openly investigate and present all the effects of a program, policy or intervention, not only the those that are most directly related to desired outcomes.
In thinking about taking what is not given, in addition the more obvious interpretations, with respect to data acquisition and analysis, we should always be striving towards appropriate interpretations of ownership. From the extreme examples such as Henrietta Lacks to more insidious examples such as not making analyses available to subjects who provided the data (most community members do not have access to medical libraries, for example), there are many opportunities to learn and continue to evolve.
Like Right Speech, we can also see this as building upon the concepts discussed in Right Vision. In work involving the public good, we are often called upon to balance the needs of society as a whole versus the needs of the individuals in society. In actions involving direct physical intervention with an individual, such as medical treatment, the harms or negative consequences are expected to be discussed, but in interventions that are less intrusive or are not aimed at specific individuals the quantification and discussion of harms, intended or unintended, is too often minimized. In order to make informed judgments and decisions, we should be prepared to openly investigate and present all the effects of a program, policy or intervention, not only the those that are most directly related to desired outcomes.
In thinking about taking what is not given, in addition the more obvious interpretations, with respect to data acquisition and analysis, we should always be striving towards appropriate interpretations of ownership. From the extreme examples such as Henrietta Lacks to more insidious examples such as not making analyses available to subjects who provided the data (most community members do not have access to medical libraries, for example), there are many opportunities to learn and continue to evolve.
Tuesday, January 17, 2012
Right Speech
"To tell the truth, to speak friendly, warm, and gently and to talk only when necessary (from The Bigview )." When conducting data analysis, the "numbers" often spring forward as the primary focus of attention. The language we use when formulating and conducting analysis is also a critical foundation. We can build on one of the concepts we discussed in Right Vision, "commitment to framing questions and presenting results in a non-stigmatizing way, especially when focused on those with less power". One of the specific ways that we can put this intention into practice is through our use of language. Science is popularly characterized as being objective, but we have to develop an active awareness of how our place in the world influences how we see and describe things.
For example, I grew up with lactose intolerance being defined as a racially associated illness or deficiency. This may seem accurate if you are defining normality as the dominant traits of populations of European heritage, but from a larger human perspective, dairy products as a regular part of the diet of adults is the exception, not the rule. Lactose intolerance is the normative state of the species and is not intrinsically associated with ill health.
In public health, we commonly invoke biased social constructs such as "births to unmarried mothers" and "teen mothers". Unless our species has developed parthenogenesis, births involve two parents. By only describing mothers, and making recommendations and policy analysis directed towards women, we add life to a long standing misogynist framework that puts women as the gatekeepers of men's sexuality.
In our language, we should strive to be slow to pathologize differences, recognize and acknowledge the limits of our perspective, use descriptive language and concepts that we would be comfortable sharing with the subjects of our analyses as well as our colleagues or sponsors.
For example, I grew up with lactose intolerance being defined as a racially associated illness or deficiency. This may seem accurate if you are defining normality as the dominant traits of populations of European heritage, but from a larger human perspective, dairy products as a regular part of the diet of adults is the exception, not the rule. Lactose intolerance is the normative state of the species and is not intrinsically associated with ill health.
In public health, we commonly invoke biased social constructs such as "births to unmarried mothers" and "teen mothers". Unless our species has developed parthenogenesis, births involve two parents. By only describing mothers, and making recommendations and policy analysis directed towards women, we add life to a long standing misogynist framework that puts women as the gatekeepers of men's sexuality.
In our language, we should strive to be slow to pathologize differences, recognize and acknowledge the limits of our perspective, use descriptive language and concepts that we would be comfortable sharing with the subjects of our analyses as well as our colleagues or sponsors.
Wednesday, January 11, 2012
Right Intention
Let us now consider right intention from a data analysis perspective. Think of the Intention of Renunciation - "resistance to the pull of desire" (from The Bigview) . Most of us go into data analysis projects with some sort of emotional investment in the results, we will have proof to gain resources for something we care about, to fight against something that is harmful, to gain recognition and prestige. At our best we should detach ourselves from the hoped for results of the analysis. We should simply want the analysis to be correct- that we will pick the correct analysis for the question that needs to be answered, and that our execution is correct. The desire for specific results can cloud our judgement.
Let us think of the Intention of Good Will - "resistance to feelings of anger and aversion" (from The Bigview). Extending upon our first reflection, conducting analysis in the right manner means that we have to be open to the possibility of having the results not turn out the way that we expect. We can find nothing, or even evidence that supports something contrary to our prior beliefs. We should not rejecting or suppress analysis solely because it doesn't support our beliefs. This does not mean that we can't challenge findings, indeed we should always be cautious and prudent when interpreting any results, but we have to maintain sufficient detachment from existing ideas that we can entertain new ones.
Finally, let us think of the Intention of Harmlessness - "not to think or act cruelly, violently, or aggressively, and to develop compassion (from The Bigview) ." I believe the principles of ethical conduct in data collection and analysis, and a commitment to framing questions and presenting results in a non-stigmatizing way, especially when focused on those with less power, are relevant to this point.
Let us think of the Intention of Good Will - "resistance to feelings of anger and aversion" (from The Bigview). Extending upon our first reflection, conducting analysis in the right manner means that we have to be open to the possibility of having the results not turn out the way that we expect. We can find nothing, or even evidence that supports something contrary to our prior beliefs. We should not rejecting or suppress analysis solely because it doesn't support our beliefs. This does not mean that we can't challenge findings, indeed we should always be cautious and prudent when interpreting any results, but we have to maintain sufficient detachment from existing ideas that we can entertain new ones.
Finally, let us think of the Intention of Harmlessness - "not to think or act cruelly, violently, or aggressively, and to develop compassion (from The Bigview) ." I believe the principles of ethical conduct in data collection and analysis, and a commitment to framing questions and presenting results in a non-stigmatizing way, especially when focused on those with less power, are relevant to this point.
Saturday, January 7, 2012
Right Vision
Many people I encounter want numbers to show what impact their intervention or program or policy is having in the world. The most important thing you can ever do is to have an accurate perception of what are the expected outcomes of your actions.
To see and understand things as they really are. What is that in practice? The first analogy is a valley. You can create a valley in a relatively short period of time with intensive sustained application of earth moving equipment and resources. You can also create a valley by the force of water flow over time, over a much longer period. Each drop of water, by itself, cannot create the valley, but the combined and sustained force of water is one of the most powerful energies on the planet.
Many programs and policies are designed to tackle massive problems, so even with what may be on the face a lot of resources, it is much more like adding drops of water to the stream than going out with dynamite and bulldozers. You cannot expect to measure the impact of your individual effort by short term changes in population measures. Creating a famine will show short term measurable changes in population weight. Nutritional labeling won't.
The second analogy continues with water. If we add our drops of water into a stream, we can have confidence that we are ultimately helping to create a valley, even though our contribution might be small. Let us take those same drops of water and sprinkle them on the desert. Our water drops, in isolation, will not be enough to substantially alter the biota of the desert. Understanding the context in which we are placing our actions is critically important to assessing their likely impact.
To see and understand things as they really are. What is that in practice? The first analogy is a valley. You can create a valley in a relatively short period of time with intensive sustained application of earth moving equipment and resources. You can also create a valley by the force of water flow over time, over a much longer period. Each drop of water, by itself, cannot create the valley, but the combined and sustained force of water is one of the most powerful energies on the planet.
Many programs and policies are designed to tackle massive problems, so even with what may be on the face a lot of resources, it is much more like adding drops of water to the stream than going out with dynamite and bulldozers. You cannot expect to measure the impact of your individual effort by short term changes in population measures. Creating a famine will show short term measurable changes in population weight. Nutritional labeling won't.
The second analogy continues with water. If we add our drops of water into a stream, we can have confidence that we are ultimately helping to create a valley, even though our contribution might be small. Let us take those same drops of water and sprinkle them on the desert. Our water drops, in isolation, will not be enough to substantially alter the biota of the desert. Understanding the context in which we are placing our actions is critically important to assessing their likely impact.
Thursday, January 5, 2012
The Noble Eightfold Path of Data Analysis
I title this post with the greatest admiration and respect for the Buddhist community. The Noble Eightfold Path is a very powerful system of principles, and while I would not apply it directly to the work of data analysis, I have been motivated by my professional experiences to try to develop a set of guiding principles for data analysis that would mirror the progressions we need to make if we are doing our work at its' highest level. I will develop these ideas in the upcoming series of posts.
Subscribe to:
Posts (Atom)