Making Data Patterns Visible

Embed Size (px)

Citation preview

  • 8/12/2019 Making Data Patterns Visible

    1/20

    Volume 27, No. 3 Spring 2009

    As the second quarter of 2009 moves at a rapid pace,

    there are a number of updates, congratulations, andmost importantly thanks to share. I am always

    pleasantly surprised by the many accomplishments of our

    member leaders, and I applaud you for your continued

    enthusiasm and energy toward making outstanding

    contributions to the Statistics Division.

    First, I would like to announce the Executive Leadership

    Committee for the 2009-2010 year. The incoming Chair is

    Vijay Nair, Chair of the Department of Statistics at the

    University of Michigan. Chair-Elect will be Christine Anderson-Cook, Research

    Scientist and Project Leader at Los Alamos National Laboratory. Bill

    Rodebaugh will continue as Treasurer and Mindy Hotchkiss will continue as

    Secretary. We all will do a great job of running our Division as we all havepassion and commitment to the Division.

    Chairs Messageby Daksha Chokshi

    Link to page 3

    Chairs Message . . . . . . . . . . . . . . . . . . . . 1

    Editors Corner . . . . . . . . . . . . . . . . . . . . . 1

    30th Anniversary Logo Contest . . . . . . . . . 4

    Long Range Planning Meeting MinutesHighlights . . . . . . . . . . . . . . . . . . . . . . . 5

    MINI PAPER

    Making Data Patterns Visible . . . . . . . . . 7

    Statistics Division Narrated Slide Shows. . 14

    Statistical Resources on the Web:

    Visual CV . . . . . . . . . . . . . . . . . . . . . . . 14

    Awards Showcase . . . . . . . . . . . . . . . . . . 15

    Nominations Sought for William G.

    Hunter Award . . . . . . . . . . . . . . . . . . . 16

    Statistics Division Fall Technical

    Conference Scholarships . . . . . . . . . . . 16

    53rd Annual Fall Technical Conference . . 17

    Treasurers R eport . . . . . . . . . . . . . . . . . . 18

    Statistics Division Committee Roster . . . . 20

    In This Issue

    Attending the division Long Range Planning meeting was a privilege for one of us (Ted Allen).

    The opportunity made visible the deep commitment many division leaders have for ASQ

    and the Statistics Division. As described in this newsletter, there was concerted effort at the

    meeting and afterward to keep the divisions service of its membership up-to-date and data-driven.

    There is also a developing vision for roles the division can play in major issues relating to the body of

    knowledge for accreditations including those related to six sigma.

    This newsletter also contains an accessible article on data visualization by Anil Menon and

    Nadhamuni Nerella. This mini paper discusses some easy and common sense approaches to looking

    at table data. In addition, the division will be well-represented at the upcoming WCQI and FTC

    meetings. As described in this newsletter, the FTC is counting on division participation in Indianapolis

    and has resources to support student participation. Additional features describe the Divisions work to

    help communicate our body of knowledge through narrated slide shows, support member growth

    and visibility through awards and scholarships, and leverage statistical resources on the web.

    Editors Cornerby Ted Allen and Shih-Hsien Tseng

    Ted Allen

    Daksha Chokshi

    Shih-Hsien Tseng

  • 8/12/2019 Making Data Patterns Visible

    2/20

    2

    Criteria forBasic Tools and

    Mini PaperColumns

    Basic Tools

    Purpose: To inform/teach the quality

    practitioner about useful techniques thatcan be easily understood, applied and

    explained to others.

    Criteria:

    1. Application oriented/not theory2. Non-technical in nature

    3. Techniques that can be understood

    and applied by non-statisticians.

    4. Approximately five pages or less inlength (8 1/2 x 11 typewritten,

    single spaced.)

    5. Should be presented in how to use itfashion.

    6. Should include applicable examples.

    Possible Topics:

    New SPC techniques

    Graphical techniquesStatistical thinking principles

    Rehash established methods

    Mini-Paper

    Purpose: To provide insight into

    application-oriented techniques of

    significant value to quality professionals.

    Criteria:

    1. Application oriented.2. More technical than Basic Tools, but

    contains no mathematical derivations.

    3. Focus is on insight into why atechnique is of value.

    4. Approximately six to eight pages or less

    in length (8 1/2 x 11 typewritten,

    single spaced.)Longer articles may be submitted and

    published in two parts.

    5. Not overly controversial.6. Should include applicable examples.

    General InformationAuthors should have a conceptual

    understanding of the topic and should be

    willing to answer questions relating to thearticle through the newsletter. Authors do

    not have to be members of the StatisticsDivision.

    Submissions may be made at any timeto the Statistics Division Newsletter Editor.

    All articles will be reviewed. The editor

    reserves discretionary right indetermination of which articles are

    published.

    Acceptance of articles does not implyany agreement that a given article will be

    published.

    VISION Data Driven Decisions Through Statistical Thinking We are the recognized forum that advances data-driven decision making through Statistical Thinking.

    MISSION Advance data-driven decision making through Statistical Thinking. Improve the publics perception and understanding of statistical methods and data-driven decisions. Be the source for the statistical components of the ASQ body of knowledge. Support the growth and development of ASQ Statistics Division members. Increase the credibility, marketability and influence of ASQ Statistics Division members.

    STRATEGIC FOCUS1. BODY OF KNOWLEDGE What it is? Where is it? How to categorize it? Disseminate via Web page Keep current Partner with HQ Goals to understand, organize, make accessible,

    inventory, gap analysis

    2. COMMUNICATION Newsletter E-Zines Align both to vision and mission Gap analysis with primary audiences Discussion boards Promote via E-Zine, conference booths Align discussion boards to vision and mission

    Evaluate whether to continue

    3. VOICE OF THE CUSTOMER Members, other divisions, audiences Proactive way to engage (go, see listen)

    4. DATA DRIVEN DECISIONS How do we advance? Do we broaden the audience? AQC session? Partnerships?

    DESIRED END STATE Our members will be proud to be part of the Statistics Division. Our Divisions operations will be a model for other organizations. We will be a widely influential authority on scientific approaches to quality and productivity improvement.

    PRINCIPLES

    Our customers needs will be continuously anticipated and met (i.e. Customer focused rather thancustomer driven). Our market focus for products and services is weighted as follows:

    Greatest weight on intermediate level. Nearly as much weight on basic level. Much less weight on advanced level.

    Focus on a few key things. Balance short-term and long-term efforts. Value diversity (including geographical and occupational) of our membership. Be proactive. Recognize that we exist for our customers. View statistics from the broad perspective of quality management. Apply Statistical Thinking ourselves; that is, practice what we preach. Uphold professional ethics. Continuously improve.

    MEETING GROUND RULES Respect and listen to all participants. No speeches.

    No side-bar discussions. Decisions by consensus, if possible. We will be open and honest, even if it hurts. Support your ideas, dont defend them. We will delegate word-smithing to small groups. All help facilitate, although we will have a formal leader, facilitator, scribe, and timekeeper (including at

    breakouts). We will rotate scribes. We will keep a separate flipchart for To-Dos. Mission, Vision, Principles, Strategy, Ground Rules should be visible.

    DisclaimerThe technical content of material published in the ASQ Statistics Division Newsletter may not have been refereed to the same extent as the rigorous refereeing that isundergone for publication in Technometrics orJ.Q.T. The objective of this newsletter is to be a forum for new ideas and to be open to differing points of view. Theeditor will strive to review all articles and to ask other statistics professionals to provide reviews of all content of this newsletter. We encourage readers with differingpoints of view to write to the editor and request an opportunity to present their views via a letter to the editor. The views expressed in material published in thisnewsletter represents the views of the author of the material, and may or may not represent the official views of the Statistics Division of ASQ.

    StrategyPrincip

    les

    Prin

    ciple

    s

    Vision

    Mis

    sion

    Strategy

    Strategy

    Strategy

    Strategy

    Strategy

    Return Home

  • 8/12/2019 Making Data Patterns Visible

    3/20

  • 8/12/2019 Making Data Patterns Visible

    4/20

    4 Return Home

    3030thth ANNIVERSARYANNIVERSARYLOGO DESIGN CONTESTLOGO DESIGN CONTEST

    Many thanks to all who participated!The Executive Council selected the logo shown here after carefullyconsidering how and where the logo would be used you will beseeing more of this as the Statistics Division celebrates its 30th

    anniversary during the coming year!

  • 8/12/2019 Making Data Patterns Visible

    5/20

    5Return Home

    In Attendance: Ted Allen, Christine Anderson-Cook, Daksha Chokshi, Gordon Clark, Doug Hlavacek, Mindy Hotchkiss,Mark Johnson, Mark Kiel, Vijay Nair, Bill Rodebaugh, and Geoff Vining.

    Agenda: Division Vision, Mission, and Strategies: Brainstorming & Reconciliation Operational Topics, Upcoming Activities, & Roundtable

    The purpose of this meeting was to review & update the Division vision, mission, and strategic plan. This meeting isheld every 3-5 years. The previous LRP, chaired by Geoff Vining, was held in 2005.

    Division Vision, Mission, & Strategies: Brainstorming & ReconciliationReview of past LRPs identified common themes: communication, infrastructure, assessing member needs, maintenanceof the organization, promotion of statistical thinking, and positioning the Division as an accessible resource formembers.

    The current image of the ASQ Statistics Division was discussed. Results were grouped using the Affinity process.Member feedback from the open-ended survey was discussed and incorporated. Geoff Vining facilitated thereconciliation process.

    The primary roles of the division were identified as follows: Technical Leadership in Statistical Thinking & Statistical Methods for System Improvement Shepherding & Dissemination of Statistical Resources for System Improvement Outreach Career and Leadership Development Learning Opportunities

    Each of the identified roles was expanded upon in order to identify key focus areas and target areas where the Division

    could concentrate its future activities and resources or expand offerings. These were reviewed against the currentvision, mission, and strategy. The Council decided that while the underlying concerns of the Division arefundamentally unchanged, the vision, mission, and strategies could be better articulated in order to more explicitlyhighlight certain issues.

    Revised Vision:The ASQ Statistics Division is the recognized leader in promoting statistical thinking and data-driven decision-makingfor system improvement.(short version) Advancing data-driven decision-making through statistical thinking

    Revised Mission: Improve practitioners understanding of statistical concepts and methods Be the source for the statistical components of the ASQ body of knowledge Support the growth and development of ASQ Statistics Decision members and the quality community

    Nurture and grow a vibrant, engaged statistical community of practice Increase the marketability and influence of ASQ Statistics Division members Demonstrate the economic benefits of statistical thinking to organizational leaders and the community

    Revised Key Strategies: Organize, enhance, and disseminate the Statistics Divisions body of knowledge Develop and deliver useful and useable communication vehicles Proactively engage the membership in Division decision-making Foster career development of members Develop future leaders of the Division and the profession Develop and enhance alliances with other organizations within and outside of ASQ

    ASQ Statistics DivisionLong Range Planning Meeting Minutes: Highlights

    March 21-22, 2009

    Orlando, FL

    Link to page 6

  • 8/12/2019 Making Data Patterns Visible

    6/20

  • 8/12/2019 Making Data Patterns Visible

    7/20

    7Return Home

    Abstract: Objectives of a well composed table are the same as that of a well-written manuscript to organize andpresent the key findings to a reader in a simple and illustrative manner. Often our tables reflect the order in which wecollect the data rather than in the order that best tells the story. Using examples we show how we can compose datatables for effective communication. The six rules of Order, Round, Average, Columnize, Layout, and Explain (ORACLE)help make the patterns and the exceptions in the data visible. These rules enable authors to avoid the curse ofknowledge, and allow the readers to see the trends and patterns in the data easily, instead of trying to figure out thetabular arrangement.Keywords: data tables, data patterns, ordering, rounding, ORACLE

    1. WHAT IS THE ISSUE?In a population demographic study (Robbins 2005) investigators raised two questions: (1) Where do most Asian-Americans live? (2) Which county is the most racially diversified? They presented their data as shown in Table 1. Inanswering the first question, we need to go across the row for Asian-Americans and find the County column having the

    largest number of Asian-Americans. While answering the first question is easier than the second, this table does noteasily enable us to see other relevant data patterns. But what patterns might we see if we keep looking at Table 1? Wehave to digest the data in Table 1 to understand the data patterns. Just as we think about our audience when writingor presenting, we need to think of the consumers of our data and the questions that need answering, including thefollowing:, what is the key conclusion? Which patterns in the data are relevant to the investigation? What are some ofthe surprises or exceptions? Does this agree with previous findings, or across different states?

    Table 1. Original Data Table from Robbins (2005): Population Demographics Organized by County and Race

    County Bronx Kings New York Queens Richmond Nassau SuffolkWestchester Rockland Bergen Hudson Passaic Total

    White 194,000 855,000 703,000 733,000 317,000 986,000 1,118,000 592,000 205,000 638,000 215,000 252,000 6,808,000

    Latino 645,000 488,000 418,000 556,000 54,000 133,000 149,000 145,000 29,000 91,000 242,000 147,000 3,097,000

    Black 415,000 845,000 233,000 420,000 40,000 129,000 92,000 123,000 30,000 43,000 73,000 60,000 2,503,000

    Asian American 38,000 184,000 143,000 392,000 24,000 62,000 34,000 41,000 16,000 94,000 57,000 18,000 1,103,000

    All Others 40,000 93,000 39,000 128,000 9,000 24,000 26,000 23,000 6,000 18,000 22,000 12,000 440,000

    Total 1,332,000 2,465,000 1,536,000 2,229,000 444,000 1,334,000 1,419,000 924,000 286,000 884,000 609,000 489,000 13,951,000

    Table 2: Modified Data Table based on Robbins (2005): Population Demographics Organized by County and Race

    (Population x10,000, Counties Ordered by Total Population, County with Maximum Concentration Identified for each Racial Group)

    County Kings Queens New York Suffolk Nassau Bronx Westchester Bergen Hudson Passaic Richmond Rockland Total

    White 86 73 70 112 99 19 59 64 22 25 32 21 681

    Latino 49 56 42 15 13 65 15 9 24 15 5 3 310

    Black 85 42 23 9 13 42 12 4 7 6 4 3 250

    Asian American 18 39 14 3 6 4 4 9 6 2 2 2 110

    All Others 9 13 4 3 2 4 2 2 2 1 1 1 44

    Total 247 223 154 142 133 133 92 88 61 49 44 29 1,395

    Table 2 is laid out similarly to Table 1 but the improved design now makes the data patterns more visible. Roundingand ordering the data help us identify the data patterns. Values in the table are truncated by dividing by 10,000.Counties with the highest population of a given race are highlighted. We can see that Kings County has the largestpopulation overall, while Queens has the greatest concentration of Asian-Americans.

    MINI PAPERMaking Data Patterns Visible

    Anil Menon, Sanofi-AventisNadhamuni Nerella, Merck & Co.

    Link to page 8

  • 8/12/2019 Making Data Patterns Visible

    8/20

    8 Return Home

    Depending on the complexity of the data, a single tabular layout may not be enough to portray the data patterns.Table 2 has been further modified (Table 3) to show percentages instead of absolute population counts. Table 3 showsthat Whites are the largest racial group in all counties except in the Bronx and Hudson counties where Latinos are thelargest race. Queens could be considered most racially diversified county since it has comparable percentages ofpopulation across the four major racial groups.

    Table 3. Alternative Modified Data Table based on Robbins (2005): Population Demographics Organized by County and Race(Percentage of County Population, Counties Ordered by Percent White Population)

    County White Latino Black Asian All Others

    Suffolk 79% 11% 6% 2% 2%

    Nassau 74% 10% 10% 5% 2%

    Bergen 72% 10% 5% 11% 2%

    Rockland 72% 10% 10% 6% 2%

    Richmond 71% 12% 9% 5% 2%

    Westchester 64% 16% 13% 4% 2%

    Passaic 52% 30% 12% 4% 2%

    New York 46% 27% 15% 9% 3%

    Hudson 35% 40% 12% 9% 4%Kings 35% 20% 34% 7% 4%

    Queens 33% 25% 19% 18% 6%

    Bronx 15% 48% 31% 3% 3%

    Total 49% 22% 18% 8% 3%

    Tables 2 and 3 effectively answer the demographic questions posed above. One display cannot do it all and so weneed to display the data in different ways to highlight the patterns of the data and answer different questions. Creatingan effective table is an iterative process of successively improving on draft versions.

    2. WHAT CAN WE DO TO MAKE OUR DATA PATTERNS VISIBLE?Based on Ehrenberg (1981, 1982, and 1998) and Wright (1973) there are six simple rules to designing an effectivetable. We Order, Round, Average, Columnize, Layout, and Explain the data. The reader will note the sentence

    condenses to the acronym ORACLE (Menon and Nerella, 2000).

    O Order rows and columns by averages or some other useful measureR Round numbers to two effective digitsA Provide averagesor totals to provide a visual focusC Columnize: Put the numbers we want to compare in columns rather than rowsL Layout the table to help comparisonsE Explain the patterns of the data

    ORDERING the rows or columns by some summary measure like the average rather than a chronological oralphabetical order helps us see the data patterns at a glance.

    ROUNDING to two effective digits (Ehrenberg, 1981) simplifies mental arithmetic and makes the numbers easier toremember and easier to see the patterns. For example, we find it difficult to divide 51.6 by 16.7 in our minds, butrounded to two digits, the number 52 is about three times 17. Wainer (1984) calls the practice of presenting digits farbeyond the number that a reader can understand as "more is murkier".

    AVERAGING rows and/or columns can help provide a visual focus and a summary of the data. This provides a basis forcomparison for each individual datum.

    Reading and comparing data down a COLUMN is easier than across rows. We also notice exceptions to the overalldata patterns.

    MINI PAPER Continued from page 7

    Link to page 9

  • 8/12/2019 Making Data Patterns Visible

    9/20

    9Return Home

    Table LAYOUT characteristics can facilitate understanding. An uncluttered table is easier to read and follow. Avoid darkgrid lines because they dominate the table, interrupt eye movements, and distract us from seeing data patterns. Single-spaced tables keep the data close to one another (Bigwood and Spore, 2003). In long tables insert a blank line everyfive lines to help us keep our place. By avoiding abbreviations we spare the user from going back and forth from thekey to the abbreviations.

    EXPLAINING the data patterns by giving brief verbal summaries helps authors and the readers understand the data andsee the exceptions. This also provides the reader mental hooks.

    A Financial Budgeting ExampleManagers routinely prepare financial budgets and track the financial performance by comparing the actual expensesincurred to the planned budget. The goal is to not exceed or underspend the budget. The deviation (or variance)between actual expenses and planned expenses is a measure of the effectiveness of the planning process.

    Table 4 is a typical report of financial budget data for a technical department, sorted by default on a six-digit expensecode. The data reflect the precision to the closest dollar even though performance evaluation and managementbudgetary decisions are based on higher thresholds ($100,000). In Table 4, data patterns are obscured that wouldhave been obvious had it been ordered by a relevant measure (Wainer 1984). Questions managers are likely to ask arethe following: (1) how did the department perform against our plan? and (2) what specific expense types show major

    deviations from the plan?

    Table 4: Original Department Budget Data Table

    Deviation $

    Expense Type Actual$ Plan$ (Actual Plan)

    050020 - Headcount Support 100,598 17,501 83,097

    050060 - Travel 729,427 284,511 444,916

    050070 - Payroll Expenses 2,767,152 3,022,676 -255,524

    050080 - Scientific Meetings 29,675 50,000 -20,325

    050090 - External Outsourcing 413,598 0 413,598

    050100 - Contract Labor 55,400 0 55,400

    050120 - Technical Consultants 16,952 147,824 -130,872

    050140 - Non-Technical Consultants 24,594 0 24,594

    050220 - Training 33,116 13,681 19,435

    050410 - Other Outside Services 41,545 10,333 31,212

    050500 - Miscellaneous 82,934 14,192 68,742

    050610 - Car Rental 530 0 530

    050620 - IT Hardware and Software 62,450 30,900 31,550

    050700 - Equipment Maintenance 17,500 20,000 -2,500

    050800 - Capital Purchase 205,000 98,000 107,000

    050520 - Lab Supplies 84,354 75,000 9,354

    Grand Total 4,664,825 3,784,618 880,207

    Table 5 shows the table once redesigned using ORACLE principles, keeping in perspective the performance evaluation

    threshold. We round the data to the closest thousands of dollars and sort them by the largest to smallest deviation. Aquick scan reveals that travel, external outsourcing and capital purchases contributed significantly to over-spending thebudget, while technical consultants and payroll expenses contributed to under-spending the budget. Manager cannow ask follow-up questions to understand or justify the deviations, and to improve the planning process.

    MINI PAPER Continued from page 8

    Link to page 10

  • 8/12/2019 Making Data Patterns Visible

    10/20

    10 Return Home

    Table 5: Redesigned Department's Financial Budget (x$1000) showing the expense types with

    largest to smallest deviations between Actual and Planned budget

    Deviation $

    Expense Type Actual$ Plan$ (Actual Plan)

    1 Travel 729 285 445

    2 External Outsourcing 414 0 4143 Capital Purchase 205 98 107

    4 Headcount Support 101 18 83

    5 Miscellaneous 83 14 69

    6 Contract Labor 55 0 55

    7 IT Hardware and Software 62 31 32

    8 Other Outside Services 42 10 31

    9 Non-Technical Consultants 25 0 25

    10 Training 33 14 19

    11 Lab Supplies 84 75 9

    12 Car Rental 1 0 1

    13 Equipment Maintenance 18 20 -3

    14 Scientific Meetings 30 50 -20

    15 Technical Consultants 17 148 -13116 Payroll Expenses 2,767 3,023 -256

    Grand Total 4,665 3,785 880

    Paint Pigment Dispersion Experiment ExampleA split plot experiment to study pigment dispersion in paint evaluated four different mixes of a particular pigment(Montgomery, 2005). The experimenters applied the pigment mix to a test panel by one of three different applicationmethods - brushing, spraying, and rolling coded as application methods 1, 2, and 3 respectively. The percentagereflectance of the pigment was studied over three days. Table 6 shows the data as shown. No attempt has been madeto bring out any patterns in the data which are not obvious because the data seem to be in a small range and thedecimal place is a distraction. It is difficult make a quick assessment about whether a particular application method isbetter than the others or a particular mix is better than the others.

    Table 6: Original Data Table from Pigment Dispersion Experiment (Montgomery, 2005)

    Day Application Mix

    Method 1 2 3 4

    1 1 64.5 66.3 74.1 66.5

    2 68.3 69.5 73.8 70.0

    3 70.3 73.1 78.0 72.3

    2 1 65.2 65.0 73.8 64.8

    2 69.2 70.3 74.5 68.3

    3 71.2 72.8 79.1 71.5

    3 1 66.2 66.5 72.3 67.7

    2 69.0 69.0 75.4 68.6

    3 70.8 74.2 80.1 72.4

    MINI PAPER Continued from page 9

    Link to page 11

  • 8/12/2019 Making Data Patterns Visible

    11/20

    11Return Home

    Table 7 shows the pigment dispersion data table after being redesigned using ORACLE principles. Note that theredesigned table does not present the split-plot experimental arrangement as performed.

    Table 7: Modified Data Table from Pigment Dispersion Experiment (Montgomery, 2005)

    ApplicationDay

    Mix

    Method 3 2 4 1 AverageBrushing 1 74 66 67 65 68

    2 74 65 65 65 673 72 67 68 66 68

    Rolling 1 78 73 72 70 732 79 73 72 71 743 80 74 72 71 74

    Spraying 1 74 70 70 68 702 75 70 68 69 713 75 69 69 69 71

    Average 76 70 69 68 71

    Days 1, 2, and 3 are similar for all application methods. The percent reflectance was highest for the rolling application

    method (#3) followed by the spraying (#2) and brushing (#1) application methods. Mix number 3 produced thehighest reflectance. Mix numbers 2, 4, and 1 are similar within an application method. While no statements ofstatistical significnce can be made directly from reviewing the table, presenting results in this way facilitates recognitionand understanding of the data in the table. The analysis of variance (Table 8) confirms our conclusions and properlyaddresses the split-plot format of the experiment. Knowing the significant factors from the analysis of variance willprompt questions about which levels of the factor are driving that significance, and how scientifically relevant, orpractical are the differences between the effects.

    Looking at Table 7 the experimenter starts thinking what to do next, We may need to measure the thickness of thecoat for each application method; what are the common ingredients or characteristics between mixes 2, 4, and 1;which ingredient is contributing to the refractive index; what are the cost, time, equipment availability, and scalabilityissues using spraying? On studying the visible data patterns in Table 7 the experimenter has gone throughcomparison, possible cause, and considered multivariate data of time, effort, cost, availability, scalability. The analysisof variance by itself is not effective in deciding what to do next.

    Table 8: Analysis of Variance (ANOVA) Results for Evaluating the Pigment Dispersion Experiment Data from Montgomery

    Source of Degrees of Sum of Mean

    Variation Freedom Squares SquareFo p-value

    Day 2 2.04 1.02

    Mix 3 307.48 102.49 135.78 < 0.01

    Day*Mix 6 4.53 0.75

    Method 2 222.10 111.05 226.28 < 0.01

    Day*Method 4 1.96 0.49

    Mix*Method 6 10.04 1.67 2.28 0.105

    Error 12 8.79 0.73Total 35 556.93

    Weld Repaired Cast Fatigue ExperimentHunter, Hodi, and Eager (1982) used a 12 run Plackett-Burman design to study the effect of seven factors on fatiguelife of weld repaired castings. Table 9 shows the factors and levels where + denotes the high level of each factor and- denotes the low level.

    MINI PAPER Continued from page 10

    Link to page 12

  • 8/12/2019 Making Data Patterns Visible

    12/20

  • 8/12/2019 Making Data Patterns Visible

    13/20

    13Return Home

    3. WHY DOES MAKING OUR DATA PATTERNS VISIBLE MATTER?Making the data pattern obvious is like having good lighting to read; it makes seeing easier (Krug, 2006). Just like goodlighting is needed to read, good design of a data table is needed to obviate the underlying patterns. A data displaythat allows us to think about important aspects of the data results in effective communication.

    When we are looking at a data display that is clear, we start thinking Why is the studied factor important? What

    would happen if we increase the factor? Should we run repeats to confirm? Is there a bias in the data? Is this anintrinsic property of the material? Are the findings too good to be true? Why does this disagree with prior findings? Inthis way we maximize the content reasoning time and minimize the time we spend figuring out the design and layout.

    When we are looking at a data display that is unclear, we puzzle over What was that abbreviation? What is differencebetween 57.56 and 24.39? What was the value in row 10? Where is the most important characteristic listed? What isthe author's finding? Which of these factors are important? Using a data display that distracts us and makes us thinkabout unimportant design features saps our energy, enthusiasm, and time.

    When creating a data display, get rid of the distractions in design or content that do not help in connecting ideas orencourage us to consider what comes next. Comments that do not provoke exploration, opportunities, or possibilitiesadd to our cognitive workload, distracting our attention from reasoning about the data. These distractions may beslight but they add up, and sometimes it doesnt take much to derail a conversation or prolong a meeting. As

    explained by David Sless (Wyatt and Wright, 1998) Information design is about managing the relationship betweenpeople and information so that the information is accessible and usable...

    4. WHY DO WE FAIL TO MAKE OUR DATA PATTERNS VISIBLE?The hazard is that we are already familiar with our data. Unlike the reader we have spent plenty of time to understandthe patterns and exceptions. We assume that the reader can easily see the patterns in the data the way we see itbecause If it is clear to me, it must be clear to them. Our subject matter knowledge acts as a blinder and prevents usfrom seeing problems in our tabular displays. We attribute our mistaken bias to the curse of knowledge, whichCramerer, Lowenstein and Weber (1989) argue that when assessing others knowledge, we are unable to ignoreknowledge that we have but others do not.

    5. WHAT ARE THE LIMITATIONS OF ORACLE?We have found the ORACLE guidelines not as effective for data tables that extend beyond one page because theinformation is not within our visual field all at once. Readers may be better served in such a case if the table is brokenup into meaningful smaller tables. ORACLE is not effective for visualizing data patterns in time series or sequential data.

    6. CONCLUSIONSOur displays should match the thinking task, whether it is comparison, change, mechanics, cause, or multivariatecomplexity (Tufte, 1997). Readers want to make sense of the data, classify the data, understand why the data patternoccurs, see if it agrees with a prior conclusion, think about the impact of the data patterns and decide what to do next.When a thinking task requires the use of tabular data, we suggest using the six rules of Order, Round, Average,Columnize, Layout, and Explain to help make the patterns and the exceptions in the data visible. The rules are simple,practical, and effective in most situations. It is not hard to apply ORACLE, but neither is it natural or instinctive becausewe are unaware that most tables can be improved to communicate better.

    7. REFERENCESBigwood, S., and Spore, M. (2003), Presenting Numbers, Tables, and Charts, Oxford University Press, pp. 26-36Cramerer, C .F, Lowenstein, G., and Weber, M. (1989), The curse of knowledge in economic settings: An experimental analysis,Journal of Political Economy, 97, 1232-1254.Ehrenberg, A.S.C. (1981), The problem of numeracy, American Statistician, 35, 67-71.Ehrenberg, A.S.C. (1982), A Primer in Data Reduction, Wiley, pp. 219-231Ehrenberg, A.S.C. (1998), Making data user-friendly, R&D Initiative, Marketing Learnings 2.Hamada, M., and Wu, C.F.J. (1992), Analysis of designed experiments with complex aliasing,Journal of Quality Technology, 24, 130-137.Hunter, G.B., Hodi, F.S., and Eager, T.W. (1982), High-cycle fatigue of weld repaired cast Ti-6l-4V, Metalurgical Transactions, 13A, 1589-1594.Menon, A., and Nerella, N. (2000), Communicating tabular data using ORACLE, Pharmaceutical Development Technology, 5, 423-431.Montogmery, D. (1991), Using fractional factorial designs for robust process development, Quality Engineering, 3, 193-205.Montgomery, D. (2005), Design and Analysis of Experiments, Wiley, p. 557.Robbins, N. (2005), Creating More Effective Graphs, New York, Wiley, pp. 322-327.Tufte, E.R. (1997), Visual and Statistical Thinking, CT, Graphics Press, p. 31.Krug, S. (2006), Dont Make Me Think: A Common Sense Approach to Web Usability, Berkeley, New Riders, pp. 11-19.

    Wainer, H. (1984), How to display data badly, American Statistician, 38, 137-147.Wright, P. (1973), Understanding Tabular Displays, Visible Language, VIII4, 351-359.Wyatt, J.C., and Wright, P. (1998), Design should help use of patients data, The Lancet, 352, 1375-1378

    MINI PAPER Continued from page 12

  • 8/12/2019 Making Data Patterns Visible

    14/20

    14 Return Home

    Statistics DivisionNarrated Slide Shows

    T

    he Statistics Division is in the process ofdeveloping Narrated Slide Shows as an approachto presenting Quality Body of Knowledge (Q-BoK)

    topics in statistics. Scott Kowalski is the Statistics Divisionleader responsible for enlisting people to develop thedifferent modules.

    A module consists of several presentations on a similartheme. Each narrated presentation within a module isapproximately 20 minutes long and focuses on aparticular statistical tool or concept. The presentations areviewed using a Web browser and each slide in thepresentation has stop/pause/rewind/forward buttons. Thisallows the person viewing to go through the presentationat their own pace.

    The Basic Statistics, Gage R&R, and Capability Analysismodules are complete. Each module has five slide shows.

    Members can download these modules from ourwww.asqstatdiv.org Web site. The cost is $5 for a slideshow or $20 for an entire module (1 slide show for free).The slide shows contained in each module are as follows:

    Basic Statistics Module Gage R&R Module Capability Analysis

    Tests for a Single Mean Introduction to Gage Overview of Capability

    R&R Studies Analysis

    Two Sample t-test Gage R&R for Crossed Capability Analysis for

    Designs Normal Data

    Paired t-test Gage R&R for Nested Capability Analysis for

    Designs Nonnormal Data

    Proportion Tests MSA: Bias and Linearity Capability Analysis for

    Attribute Data

    Power and Sample Size Gage R&R for Attribute Confidence Intervals

    Data for Capability Indices

    Examples of Capability for

    Normal and Nonnormal

    Data

    There are also 2 free narrated slide shows; one is on thebasics of hypothesis testing and the other is an invited talkgiven at the 2005 Fall Technical Conference. The StatisticsDivision hopes to add more free presentations fromvarious sponsored conferences.

    In addition, there is a module on Statistical Thinkingwhich contains 2 presentations and costs $5. TheStatistics division is currently developing the following fivemore modules:

    Regression Analysis

    Analysis of Variance (ANOVA) Statistical Process Control (SPC) Design of Experiments (DOE), and Soft Six Sigma ToolsThe presentations help the viewer gain a thorough

    understanding of how and where the statistical tool isapplied. By enlisting subject matter experts, the StatisticsDivision believes these presentations provide an excellentuse of technology to disseminate the Q-BoK. Go to thefollowing URL to learn morehttp://www.asqstatdiv.org/narrated.htm

    Statistical Resources onthe Web: Visual CV

    http://www.visualcv.com/

    By Mindy Hotchkiss

    This column highlights resources available on the web of

    particular interest to industrial statisticians or quality andreliability engineers. Please feel free to contact me with anycomments at [email protected] or if you know ofany particularly useful sites or tools that you would like torecommend.

    In times like this, with layoffs an ugly reality for many,we realize that not all resources that statisticians mayneed are strictly statistical. In this spirit, this month

    we highlight VisualCV.com, which allows users toconstruct interactive internet-based resumes thatshowcase their experience, abilities, and accomplishmentsand share them using online social networking tools.Registration is free to both students and professionals.

    VisualCV helps users to organize information abouttheir work history, educational background, andexperiences in a secure online environment. Users canupload files to their Portfolio to provide both examples oftheir work and supporting evidence of theiraccomplishments, e.g. papers, presentations atconferences, newspaper articles, or video/ audio files.Resume-writing guidance is available on the website tohelp all users put their best face forward.

    Users are not limited to one CV either multiple CVscan be created for different purposes. Users can control

    and monitor access to each CV independently as well.CVs are offline during construction and become availableto others only when shared. For example, a publicversion can be made available to the world at large, butmore detailed alternatives can be set up to specificallytarget a particular job. VisualCVs can also be linked toother networking sites, such as LinkedIn, emailed topotential employers, or posted on job sites.

    Companies can also create an online presence usingVisualCV. Basic membership is free, but additionalfeatures are available through a paid Career Connectionmembership. At the time of writing, 691 companies haveVisualCV profiles.

    VisualCV is available at http://www.visualcv.com/.

  • 8/12/2019 Making Data Patterns Visible

    15/20

    Awards ShowcaseRoger Hoerl receives ASQs 2008 distinguished Shewhart medalThe Shewhart Medal is awarded to one individual each year for outstanding contributions to the science andtechniques of quality control or demonstrated leadership in modern quality control.

    The Shewhart medal for 2008 was awarded to Roger Hoerl, PhD. Roger currently leads the Applied Statistics Lab atGeneral Electric in Schenectady, NY, which specializes in the creation and application of advanced statistical methodsfor research and development. Dr. Hoerl is a fellow of the American Society for Quality, the America StatisticalAssociation and has been elected to the International Statistical Institute. Roger won ASQ's 2001 Brumbaugh Awardfor the published paper with the greatest impact on industrial quality control applications. He also received the 1999William G. Hunter Award, named and presented annually in honor of ASQ Statistics Division's founding chair.

    For outstanding technical leadership in the development and effective use of quality improvement andstatistical methodology; for innovative contributions to Statistical Thinking and Six Sigma; and forworldwide communication of these ideas through numerous thought provoking publications, books,courses, and presentations.

    Five Division members receive ASQs Testimonial AwardTestimonial Awards are given out to Division leaders and members as an acknowledgement of their dedicated serviceand contributions to the organization.

    So what do these five division members have in common? All five have been past chairs of the Statistics Division. Allfive have promoted statistical methods and statistical thinking. And they have worked tirelessly to understand andsupport the needs of our members. Congratulations to them for their leadership and accomplishments!

    Statistics Division achieves 2007-2008 J.S. McDermond Total Quality LevelThe J.S. McDermond Total Quality Level recognizes excellence in division management and is the highest level ofachievement attainable by a division. Divisions receiving the J.S. McDermond Total Quality Level are recognized fordemonstrating dedication and commitment to serving the needs of their members. For the 2007-2008 fiscal year, theStatistics Division was chaired by Doug Hlavacek. Congratulations to Doug and his team!

    Statistics Division recognizes 30th Anniversary Logo Contest ParticipantsMany thanks to all of our logo contest participants! All contributions have been printed in this newsletter. Theselected logo will be utilized on a variety of special Statistics Division commemorative materials! Once these materialshave been printed, all contributors will be receiving a complementary set in the mail.

    Statistics Division recognizes Long Range Planning Survey ParticipantsMany thanks to everyone who responded to our Long Range Planning Survey! The Division Council was thrilled toreceive so many thought-provoking responses. Your time and effort is greatly appreciated. All contributors will bereceiving a token of our appreciation in the mail.

    Congratulations to all our Award winners!

    Doug Hlavacek

    Chair 2007-2008

    Gordon Clark

    Chair 2006-2007

    Geoff Vining

    Chair 2005-2006

    Mark Kiel

    Chair 2004-2005

    Davis Balestracci

    Chair 2003-2004

    15Return Home

  • 8/12/2019 Making Data Patterns Visible

    16/20

  • 8/12/2019 Making Data Patterns Visible

    17/20

    17Return Home

    The 53rd Annual Fall Technical Conference will be held in Indianapolis, Indiana on October 8-9, 2009. Thetheme of the conference is Quality and Statistics: Accelerating to Higher Performance. The conference willfeature the latest developments in statistical methods as they relate to quality improvement and decision-

    making and will highlight discoveries in unique and innovative statistical methodologies and quality tools.The technical program will include presentations on the following topical areas: multivariate SPC, measurement

    systems analysis, screening methodologies, generalized linear models, DOE theory and case studies, design algorithms,strip and split plot case studies, advanced process control, reliability, advanced visualization, and biometricidentification.

    Greg Piepel, Laboratory Fellow, Statistics and Sensor Analytics, Pacific Northwest National Laboratory has beenselected to present the Youden Address. Preceding this will be a special Statistics Division invited session on thehistory, current activities and long range plans of the division. An open reception celebrating the 30th anniversary ofthe Statistics division will be held immediately following the Youden address.

    When most people think of Indianapolis they think of car racing and the Indy 500. Attendees will learn about thelargest single day sporting event in the world from Thursdays luncheon speaker Mr. Joie Chitwood, President andCOO of the Indianapolis Motor Speedway.

    American Statistical Association President Sally Morton will speak at Fridays luncheon. She will discuss therelationship of ASQ and ASA and challenges for both organizations.

    Indianapolis is also home to the NCAA Hall of Champions and the Indiana State Museum. The Childrens Museumof Indianapolis is the largest childrens museum in the world. The Indianapolis Zoo is close to the downtown locationof the conference.

    For more information contact Frank Rossi at [email protected] or visit the conference websitehttp://www.indyasq.org/stats/for the conference program as the event draws nearer.

    53rd Annual Fall Technical Conference

    ASQ Indianapolis Proudly Presents53rd Annual Fall Technical Conference

    Quality and Statistics: Accelerating to

    Higher Performance

    October 8th & 9th, 2009

    Hilton Indianapolis / Indianapolis, IN

  • 8/12/2019 Making Data Patterns Visible

    18/20

    18 Return Home

    2008-2009

    Budget

    July - Feb

    Actual

    Expenses (con't.)2008-2009

    Budget

    July - Oct

    Actual

    $60,000 $20,871 Web Design & Maintenance $4,500 $1,259

    $1,000 $0 Narrated Slideshows $500 $92

    $1,600 $455 Virtual Academy $0 $0

    $0 $0 Outreach Projects $6,500 $3,000

    $0 $0 FTC Sponsorship $3,500 $0

    $0 $0 ISBIS Conference Short Courses $3,000 $0

    Miscellaneous $0 $1,235 Colombian National Meeting $0 $3,000

    $62,600 $22,560 Tactical Plans Sub-Total $11,500 $4,351

    Hunter/Nelson Awards (plaque) $500 $329

    $1,500 $0 Hunter Awardee Honorarium (travel) $1,000 $0

    $500 $0 Youden Speaker (travel) $500 $0

    General Fund $2,000 $0 FTC Student Grants $1,500 $1,774

    $6,000 $0 ASQ Testimonials ($50 each) $100 $0

    Travel, Hotel $6,000 $0 Service Awards (WCQI, FTC Reps) $700 $0

    $6,600 $1,081 Outgoing Chair's Gift $500 $112

    WCQI Meeting/Hospitality $4,100 $500 Awards Sub-Total $4,800 $2,215

    WCQI Travel $2,500 $581

    $7,500 $13,539 Misc/postage $100 $0

    Travel, Hotel, Meals $7,500 $13,539 Misc/travel $500 $584

    $3,000 $2,698 Misc/other (teleconferences) $100 $87

    FTC Meeting/Hospitality $500 $1,357 Misc- Sub-Total $700 $671

    FTC Travel $2,500 $1,341

    $7,500 $0 Total Expenses (amended) $70,100 $28,627

    LRP Meeting/Hospitality $2,500 $0 Total Expenses (original) $60,600

    LRP Travel $5,000 $0

    $30,600 $17,318$0 $0

    $0 $0 Ott Scholarship

    $0 $0 Assets (as of Feb 28, 2009)

    $0 $0 Scholarship Fund 200,000 180,990

    $2,000 $300 Expenses

    $0 $0 Scholarship (2) 10,000 10,000

    $2,000 $0

    $2,000 $300

    $2,500 $715 Ending Balance (as of Feb 28, 2009)

    $2,000 $715 Checking $63,790

    $500 $0 Money Market $111,399

    $10,000 $0 Accounts Receivable $1,642

    $6,000 $0 ASQ $1,642

    $2,500 $0 Dividends

    $1,000 $0

    $500 $0 Current Assets $176,831

    $12,500 $715 Capital Assets $6,413

    0 0 depreciated to $0

    0 0 Long Term Assets $238,239

    0 0 from reserve fund $57,249

    6,000$ 3,058$ Ott fund $180,990

    -$ -$ Total Assets

    51,100$ 21,391$ *10% unrealized loss in assets due to stock market volatility.

    Amendment approved 03/21/09: LRP $7500; WCQI Event $2000

    $415,070

    Committees Sub-Total

    Newsletter Comm

    Nominating Comm

    Programs Comm

    Publications Comm

    Standards Comm

    Promotions Comm

    Postage/Misc

    Special Publication (every year)

    Sp Pub Printing

    Sp Pub Postage

    Sp Pub Reprints

    Sp Pub Honorarium

    WCQI Exhibitor Fees

    WCQI Promotional Items

    WCQI Recognition Event

    Membership Comm

    Regular Newsletter (3)

    Printing (Layout, pdf files)

    Long Range Planning (3 -5 yrs)

    Planning CommAuditing Comm

    Bylaws Comm

    Certification Comm

    Examining Comm

    New Member Mailings

    DAC/WAC Meetings (Nov, May)

    WCQI Tactical Meeting (May)

    Operational Planning (July)

    FTC Council Meeting (Oct)

    Interest/Royalties

    Teleclass Revenue

    AQC Tutorials

    FTC Short Courses

    Total

    Expenses (as of Feb 28, 2008)

    Revenue (as of Feb 28, 2009)

    Dues

    Retail Sales

    TREASURERS REPORT7/1/08 2/28/09

  • 8/12/2019 Making Data Patterns Visible

    19/20

    19

    Committee Name Division Position E-mail address Telephone

    OFFICERS

    Daksha Chokshi Division Chair [email protected] 561-796-8373

    Vijay Nair Chair-Elect [email protected] 734-763-8018

    Bill Rodebaugh Treasurer [email protected] 215-743-0406

    Mindy Hotchkiss Secretary [email protected] 561-796-8146

    Doug Hlavacek Past Chair [email protected] 651-293-4465

    STANDING

    Examining

    Chair Geoff Vining Examining Chair [email protected] 540-231-5657

    Auditing

    Chair Daksha Chokshi Division Chair [email protected] 561-796-8373

    By-Laws

    Chair Doug Hlavacek Past Chair [email protected] 651-293-4465

    Nominating Chair Doug Hlavacek Past Chair [email protected] 651-293-4465

    Program

    Co-Chair Mindy Hotchkiss Secretary [email protected] 561-796-8146

    Co-Chair Scott Kowalski Vice Chair - Products & Services [email protected] 407-328-9609

    Publications

    Co-Chair Mindy Hotchkiss Secretary [email protected] 561-796-8146

    Co-Chair Scott Kowalski Vice Chair - Products & Services [email protected] 407-328-9609

    Voting Member Ted Allen Newsletter Co-Editor [email protected] 614-292-1793

    Voting Member Shih-Hsien Tseng Newsletter Co-Editor [email protected]

    Voting Member Willis Jensen Special Publications Editor [email protected] 928-864-3041

    Non-Voting Member Rudy Kittlitz Glossary & Tables Editor [email protected] 432-837-9937

    Non-Voting Member Steve Schuelka How To... Series Editor [email protected] 219-689-3804

    Strategic Planning

    Chair Daksha Chokshi Division Chair [email protected] 561-796-8373

    CONSTITUTED

    Tactical Planning

    Chair Vijay Nair Chair-Elect [email protected] 734-763-8018

    Promotions

    Co-Chair Bill Rodebaugh Treasurer [email protected] 215-743-0406

    Co-Chair Mark Kiel Vice Chair - Outreach [email protected] 219-888-3788

    Non-Voting Member Small Web Solutions Web Master Contact [email protected] 219-988-3139

    Non-Voting Member Vijay Nair Body of Knowledge Chair [email protected] 734-763-8018

    Membership Needs

    Co-Chair Bill Rodebaugh Treasurer [email protected] 215-743-0406

    Co-Chair Mark Kiel Vice Chair - Outreach [email protected] 219-888-3788

    Voting Member Brian Sersion Membership Chair [email protected] 513-363-0177

    V oting Member Mark Johnson Standards Chair [email protected] 407-823-2695

    Voting Member Harry Koval Certification Chair [email protected] 651-776-9503

    AwardsChair Doug Hlavacek Awards Chair [email protected] 651-293-4465

    Non-Voting Member Lynne Hare Ott Scholarship [email protected] 908-897-0610

    Non-Voting Member Paul Prew FTC Student Grants [email protected] 651-795-5942

    Non-Voting Member Robert Mitchell Hunter Award Chair [email protected] 651-736-8684

    Non-Voting Member Ted Allen Nelson Award Chair [email protected] 614-292-1793

    Non-Voting Member Doug Hlavacek Youden Address Chair [email protected] 651-293-4465

    Non-Voting Member Daksha Chokshi Testimonial Awards [email protected] 561-796-8373

    Activity Chairs

    Non-Voting Member Frank Rossi FTC Representative [email protected] 847-646-5196

    Non-Voting Member Robert Mitchell WCQI Session Manager [email protected] 651-736-8684

    Non-Voting Member Bob Brill Short Course Chair [email protected] 785-749-8124

    STATISTICS DIVISION COMMITTEE ROSTERMembers of STAT Council

    2008-2009

  • 8/12/2019 Making Data Patterns Visible

    20/20

    The ASQ Statistics Division Newsletterispublished in three quarters of the yearby the Statistics Division of theAmerican Society for Quality.

    Al l commun icat ions rega rd ing th ispublication, EXCLUDING CHANGEOF ADDRESS, should be addressed to:

    Ted AllenThe Ohio State Univeristy1971 Neil Avenue210 Baker SystemsColumbus, OH 43221email: [email protected]

    Other communications relating to theStatistics Division of ASQ should beaddressed to:Daksha Chokshi, Division ChairASQ Fellow; P&W Technical FellowStatistical Methods and Quality

    TechnologiesM/S 731-84, P.O. Box 109600, West

    Palm Beach, FL 33410-9600(561) 796-8373(860) 755-8598email: [email protected]

    Communications regarding change ofaddress should be sent to ASQ at:

    American Society for QualityP.O. Box 3005Milwaukee, WI 53201-3005

    This will change the address for allpublications you receive from ASQ.You can also handle thi s by phone(414) 272-8575 or (800) 248-1946.

    STATISTICS DIVISION AMERICAN SOCIETY FOR QUALITY

    U P C O M I N GN E W S L E T T E RD E A D L I N E S F O RS U B M I S S I O N SIssue Vol. No. Due Date

    Fall 2009 28 1 August 31, 2009

    Winter 2010 28 2 Nov. 30, 2010

    VISIT THE STATISTICS DIVISION WEBSITE:

    www.asqstatdiv.org

    Other Periodicals for Applied Statistics

    http://www.asq.org/pub/jqt/