Graphs: An effective tool, but use them carefully

Ahh…the graph.  Where would the business world be without them?  While some of us are just as content looking through a giant spreadsheet full of numbers, graphs can help to illustrate the story more effectively for number geeks and math haters alike.  However, while graphs can be a great tool, there are certainly times when graphs can make interpreting your data even more difficult to understand or (intentional or not) even misleading.  Here are a few things to think about when creating graphs for your data.

Line charts should represent something linear!

Line charts are a very common way of representing data.  However, in most cases, line charts should only be used if there is some sort of linear relationship in the categories displayed on the horizontal axis of your chart.  For example, if you want to see how responses vary by age of respondents, the year the data was collected, or even satisfaction on a numeric scale, a line chart can be a great way of representing this data.

DO: Customer Satisfaction by Year

However, if you are instead dealing with categorical data, using a line chart suggests a relationship between the categories that may not be true.   In the example below, a line chart implies that Colorado is related to Nebraska in the same way that Nebraska is related to Wyoming.  Clearly this isn’t true, so in this case, a bar chart would likely be a more effective way of presenting the data.

DON'T: Customer Satisfaction by State

Pie charts (and split-bar charts) should add to 100%!

Everyone loves pie charts.  Not only are they the best type of graph to use if you want to represent your favorite food (or video game character), they are an excellent way of presenting the distribution of data in which every data point belongs to one category.

DO: Gender

However, pie charts can cause all sorts of problems in interpretation if the data is not mutually exclusive (that is, if a single data point can belong to multiple categories).  In the example below, the pie chart implies that the chart represents the total population in terms of pet ownership, but some people may have multiple types of animals.  Again, in cases such as these, a bar chart would be a much more clear way of presenting this data.

DON'T: Pets at Home

Be careful with the scales you use!

While bar charts can be a good option for a wide variety of data, the scale you use for your charts can cause confusion in interpretation if you aren’t careful.  In particular, if you let your graphing software choose your scale for you, you may end up with results that tell a very different story than reality!  For example, let’s say you wanted to compare customer satisfaction across customer segments.  If you pulled out trusty old Excel and graphed this data with no modifications, here’s the output you would get:

DON'T: Satisfaction with Automatic Scaling

Even with the percentages listed on the graph, it looks like there is a HUGE difference in the satisfaction of Segments A and B compared to the others.  However, the difference in satisfaction between Segments A and E is really only 13 percent.  Here’s the same data, but using a fixed scale from 0 to 100 percent:

DO: Satisfaction with Fixed Scaling

By ensuring that the scale represents the entire range of possible responses, we can more accurately convey the true differences between segments.

Other things to consider

These are, of course, just a few things to consider when presenting your data.  We haven’t even touched on other topics like overall charting philosophies, how newer visualization techniques can result in pretty, but dysfunctional graphs, or the nuances of more advanced types of visualizations, such as cartography.  However, by keeping in mind what your data is meant to represent and ensuring that your approach avoids some of these pitfalls, you’ll be on your way to more meaningful and accurate graphs.

Welcome, Gregory!

We are delighted to introduce Gregory Hornback as the newest Welcome
member of the Corona Insights team! To learn more about Gregory’s role at Corona Insights, check out his bio here.

We also thought it’d be fun to include a little Q&A to help us and our readers learn more about Gregory. Enjoy!

Q. What is your favorite hobby?
A. I have a bunch of odd hobbies – like mechanical keyboards and wet shaving – but my favorite would probably be computers. I built my first computer when I was in high school and have enjoyed messing with the technology ever since. Just like most other computer enthusiasts, my computer really gets put through its paces when I’m enjoying the most recent and demanding games.

Q. What are you looking forward to most now that you live in Colorado?
A. Colorado has been my favorite state ever since I was young. Being able to drive up into the mountains just to ski or snowboard for the day is a lifelong dream of mine, so I would definitely say the mountains. Also, it’s not well known that Indiana is brutally humid, so every day I am looking forward to enjoying the pleasant, low-humidity weather here in Colorado.

Q. What is your favorite book?
A. Despite my love for high-fantasy books (magic, mystical creatures, etc.), my all-time favorite book would have to be The Stand by Stephen King. I found it to be a fascinating portrayal of humanity and survival in a post-apocalyptic world. Though the book might find its roots in horror, it is truly one that everyone should pick up.

Q. Favorite movie?
A. No Country for Old Men definitely takes the cake.

Q. What is your perfect pizza?
A. My perfect pizza has a few key criteria: 1) it must have a fantastic crust that has a nice and crisp, but not burnt, outer edge with a firm yet spongy inside 2) the sauce must have some sweetness to it but not taste overly sweet 3) it should have a saltier taste baked into the cheese and 4) it should contain extra cheese on one half and pepperoni, ham, bacon, and Canadian bacon on the other.

Q. Favorite college course?
A. My favorite undergraduate course was almost definitely my consumer psychology course. It was a course that combined my love of studying people and my interest in consumerism and economics. It contained the context (within the larger field of psychology) that really spoke to me. In my graduate studies, my favorite course was Online Interaction, a course detailing the history of the internet, how communities formed, and everything on how individuals interact with one another via computer-mediated-communication.

Q. Why Market Research?
A. I originally started out my college career majoring in Chemistry. After I quickly discovered chemistry (and to a greater degree, the required physics courses) was not for me, I fell back to a field I knew I loved, but I wasn’t sure how to make a career out of – Psychology. Eventually, after focusing most of my education around social-psychology, I stumbled upon the field of Consumer Behavior. A string of video games were released that absolutely bombed, despite insanely high budgets, and I was extremely confused; I knew these companies could spend even a small percentage of their budgets on researching how people react to certain aspects of their games in development, but they didn’t seem to be. It was then that I found my passion. Studying how people act as consumers has been my focus ever since, eventually leading to a graduate degree in the field. In my mind, market research is a way in which I can utilize my passion and make it a career.

Wondering if you’re normal? The answer is probably yes.

Katherine Brown is Corona’s 2015 summer intern. She has one year left at Macalester College where she has chosen to study anthropology because it gives her an excuse to talk to people from all over the place and call it work. After living in Germany for a calendar year and Turkey for an academic one, Katherine is able to pass for a German and haggle for coffee sets in Turkish. She’s also well on her way to writing a senior thesis about the huge German Turkish population. When she’s not busy looking for people she’s never met to tell her about their lives in detail, she’s likely spending time with some portion of her Denver-based extended family in restaurants, the mountains, or both. 

Wondering if you’re normal? The answer is probably yes.
Written by Katherine Brown

Early this summer in the midst of looking for jobs and internships, I decided to make a documentary about my extended family. When I mentioned this idea to my eighteen year old cousin, she immediately wanted in. Since I’m all about teamwork and she’s more like a sister than a cousin to me anyway, it didn’t take much convincing at all before we were a small team of two.

Our first (and, I will add, only) meeting was a discussion of how to start the project. It’s hard to make a documentary about your own opinionated and very close family without offending anyone. We decided the best thing to do would be to write up a short questionnaire, send it out to the family, see where people had strong opinions, and go from there. Coming from an anthropology background and hours of training on how to frame questions, how to stay neutral, and how to make people comfortable enough to share information, coming up with five questions for my family seemed like nothing. “What are the biggest generational gaps?” “Do you see major points of contention? What are they?” I wanted to look for patterns, anything we could latch onto and say “YES! That’s it! Let’s dig deeper there!” My cousin, who just graduated from high school, was set on a different approach. She felt that it was crucial to know just one thing.

“Is our family normal? Why or why not?”

It took me a while to figure out why this one question seemed so crucial. Maybe she was looking for someone to confirm her suspicions that our family is not normal. Maybe she wanted everyone to tell her that we are, in fact, normal and that she didn’t need to worry. I mostly just put it off to the fact that she hadn’t had the training in asking questions that I have.

A couple of weeks later, I started working at Corona as an intern for the summer. By then, I’d more or less forgotten about the incident with my cousin. My life became filled with interviews and survey data. It wasn’t until this week that I had a chance to step back again and think about my cousin’s question. Apparently all of this first-hand exposure to research has been teaching me something, because I suddenly understood why she had been so focused on the question of normal. She knew what we were really trying to get at through this documentary.

“Am I normal?”

Through my coursework and experience interviewing people, I’ve always focused on the process. It never occurred to me that all the data analysis (whether from surveys or interviews or focus groups) I’ve done is really just a quest to find “normal”. Of course, each data set is slightly different, focusing on finding normalcies among slightly different populations. But no matter what kind of data I am looking at, people love to emphasize their own ‘normalness’.

In my academic research on German Turks, many interviews will start with people telling me that they are “not really the kind of person I’m looking for”. Their story is different from the ones I’m used to hearing because they don’t come from Eastern Anatolia. Or because they “did not have a choice when they moved to Germany”. Or because they “didn’t even move back to Turkey on purpose”.  And yet, in each interview I hear more or less the same thing. The details are different, of course. But there are notable patterns in what people emphasize when I ask what it means to be a German Turk. These people who present themselves as unique or abnormal almost always end up giving me details about how they are just like all the other German Turks. I hear “Anyone who has had experience in Germany will tell you that…” in almost every interview, including the ones that begin with the phrase “I’m not really the kind of person you’re looking for”. Those normal habits of abnormal people are exactly what I’m looking for when I’m building my senior thesis. Those habits are the ones that matter.

My exposure to qualitative research at Corona has shown me similar results, albeit without the introductory phrase of “I’m not the right person”. Research here is often conducted to find out people’s opinions. I’ve noticed that most people believe that their own ideas are also the most widely-held opinions, the ‘normal’ opinions. Answers to an interview question sometimes begin with pauses implying how ‘obvious’ the answer is going to be. In the next interview someone may pause for the same reason, then completely contradict the last person’s answer. What I’ve learned from this is that everyone is normal. Everyone knows the ins and outs of their industry. Everyone knows what everyone else thinks. At least until you talk to everyone else.

Even quantitative survey data allows room for people to remark upon their normal lives. People will circle survey questions and comment “why are you even asking this?”, either forgetting or oblivious to other people’s perceptions of normal. In open ended responses people will write that they do “the normal things” or go somewhere “right next to where you get off the train”. That version of normal may or may not be the same as mine. Many of these people are filling out the surveys quickly, and do not remember that they’ve never met me. Why would any one person’s experience be anything BUT normal?

I will spoil the end of the unfinished family documentary.  Everyone in my family responded that we are normal. Their responses to my open-ended, non-judgmental, broad, neutral questions varied pretty greatly. And yet, there were patterns. In my experience, following those patterns would lead to a pretty representative profile of a ‘normal’ family with the same education, socio-economic status, geographic location, and generational breakdown as my own. On her quest to find normal, my cousin may be beginning to realize that everything and everyone is normal. On my quest to become a better researcher, I’m beginning to realize that what I’m looking for is the answer to the question my cousin was so set on asking.

“Are you normal? Why or why not?”


The more things change…

If you have been reading our blog you probably have noticed that our summer here at Corona has been one of professional development and learning. Granted, we like to think every season is, but maybe it’s the heat that makes staying indoors and reading or conferencing all the more attractive.

Conferences especially are supposed to show off the latest and greatest, and the MRA’s Insights and Strategies Conference certainly had that, from Trend Hunter presenting, to Adobe and other “future of market research” keynotes.

What has been particularly interesting to me, though, is the “back to basics” theme of many conference sessions, books, and webinars as of late. Take for example these takeaways from some recent conference sessions I attended:

  • Staying on top of demographic trends. Census data isn’t always sexy, but it’s reliable and paints the best picture of America and how it’s changing.  Taking a step back to gut check how you think your market is changing to how it actually is changing is invaluable.
  • Phone research isn’t dead. Our clients, I think, are sometimes surprised when we propose a telephone survey over online.  Yes, telephone (including cell phone) still exists, and in fact can be the best sampling means out there depending on your objective (and price competitive too).
  • Know what you’re trying to accomplish before starting the research. Seems obvious, but worth repeating. This was interestingly both the topic at the conference as well as Corona’s most recent book club (see Mollie’s post here). Knowing how you’ll use the research can inform what type of information you need to learn which will then inform the methodology.

Indeed, the more things change the more they stay the same. If only we all got summer vacation again too.

Hey honey, wanna help me with a big project?

Facilitating success over 15 years of love and consulting.

In my mind, the question with milestones is this, which one is significant enough to celebrate? First year, 10, 20 or 50 – no brainer. What about 15? Sometimes the mid-decade mark is an important one. That’s the case for me this year as I celebrate 15 years with my company, Corona Insights. Oh, lots has changed over those years. I’m no longer officing out of the spare bedroom with the double bed that was great for holding files and for taking naps. Today I can catnap in my comfy office chair. Oh, and then there is air conditioning instead of a swamp cooler. Plus I have more colleagues than the one who is forever first for me, Kevin Raines.

We don’t do payroll in-house anymore off the 1st Bank checkbook. Now we have a payroll service and a super talented office manager to handle all of that stuff. Like many a start-up there was no salary for the first six months.

Those first six months of 2000 were a bit like the austerity period when Kevin and I were saving to become homeowners in 1994. No frills, no fluff. Our old house would just have to wait a bit as we saved for future remodels, especially since Raines Manor also served as Corona world headquarters. Who knew that the habits we’d built as first-time home buyers, and later remodelers, would serve us well as we had to chart the financial course for our company?

I’ve had the joy and Now, Future, Pastpleasure of building my consulting practice from scratch since 2000. I’ve also had the joy and pleasure of helping to build a small business too. Of course, if I’m honest, there has been some pain too, maybe quite a bit depending on the day, month or year. 2010, in the depths of the recession, many a painful moment. Certain challenging consulting gigs required many hours of free therapy from fellow consultants and colleagues. But I get ahead of myself once again.

Over the next several blogs I will reflect on my early years as a Jane of all trades, back in the day when I staffed research projects. (Some of you will recall the observational seat belt research we conducted on Denver street corners.) Then I’ll focus on the clients and projects that served as the crucible for my consulting practice. Finally, I’ll touch on what keeps me fresh and motivated today.

In the meantime it is interesting to note how much 2015 is like 2000. The nonprofit strategic consulting market is as fragmented today as it was then, possibly even more so. It remains largely populated by individual practitioners and smaller firms, many of which are still going today, although a few of my consultant peers have left the field to pursue other dreams and new folks have come into the market.

Business is still largely derived from repeat and referral sources. Friends and colleagues mattered then and matter now. What also matters is deep passion and love for my craft and the people who have helped me master it. There are too many of you to mention by name – and I thank you from the bottom of my heart.

Like they did in 2000, 2005 and 2010, prospective clients often begin their call or email with “I’d like to speak with you about facilitating our strategic planning process.”

Fortunately, some things never change.

I remain deeply indebted to my beloved who said, “Hey, I won a big project. Do you wanna help?” Who knew love and consulting would go so well together?


Adventures in Book Club: A Research Story

As it finally starts to feel like summer, one of my favorite activities is sitting outside on the patio reading a good book.  My fellow Coronerds must agree with this sentiment as we recently kicked off summer with our very own Corona Book Club, fearlessly led by Senior Associate, Matt Bruce.  As we started discussing this quarter’s book, I got to thinking that the ingredients for a successful book club really mirror the ingredients for a successful research project.

Coronerd Book Club

Those of you who may have participated in a book club on your own personal time may have noticed that some book clubs thrive, while others flop.  I would argue that success starts with clearly defining a purpose.  Is the purpose of the book club to socialize, to learn new writing techniques or to learn about a specific topic?  If all members of the club don’t agree on a purpose, it probably won’t be successful.   Imagine the chaos that may ensue if half of the members haven’t read the book because they came just to socialize, while the other half planned to have an in-depth discussion about literary techniques.  If you don’t have a clear purpose for the book club, you won’t be able to agree on a book or have a meaningful discussion.

The same is true for research projects.  We’ve talked a lot in the past about beginning with the end in mind—that is, being clear about why the research is being conducted in the first place, and how it will be used once complete.  When our clients are clear about the purpose behind the research project, and all stakeholders are on the same page, we are better able to guide the research by creating research goals, asking the right questions and using appropriate analysis techniques so that the research is actionable.

Wondering about the purpose of Corona’s book club?  It’s to continually learn about ways to improve our processes to make our work more useful to our clients, of course!

Tax Burden by County in Colorado

We saw an article in the Durango Herald the other day noting that La Plata County has the fourth-lowest tax burden in the state, and it made us curious.  We found the web site where the figures were posted.

The study used assumptions about a typical household, and calculated its cost in state income tax, property tax, sales tax, and fuel tax.  Because Colorado has a flat-rate income tax, that’s a wash throughout the state.

If one examines only the other three costs, the lowest tax for this hypothetical household is in Las Animas County at $1,337.  The next lowest rates are found in Gilpin County ($1,370), San Miguel County ($1,467), La Plata County ($1,676), and Gunnison County ($1,777).  The median among the 58 counties measured was $2,292.  At the bottom, the most heavily taxed Colorado residents live in Adams County, where the same household will pay $3,635, almost 60 percent above the median and nearly triple the Las Animas rate.  The difference from the median in Adams County is almost exclusively due to higher property taxes.

Tracking H2O

As a skiing and a river rafting enthusiast, I’m interested in snow levels of our nearby Rocky Mountains and of water levels in our creeks and rivers.  Early summer is therefore an exciting time to monitor snowpack and river levels because changes are dramatic. Luckily, two websites provide current data on snowpack and river levels that illustrate data in interesting ways.

To understand snowpack levels, I surf over to the National Resource Conservation Service (NRCS) SNOTEL website, which tracks snowpack across eight watersheds in Colorado. This website is easy to use and the data is updated about every four to six days.

Since I live in the South Platte River Basin, I frequently check out our time series snowpack summary, which displays snowpack estimates for the current year as well as the previous three years. As you can see below, the average and median (called normal) snowpack levels (red lines) are close to identical.  Looking at these lines, you can see that May 1st is a typical “tipping point” that leads to a downward melting trend.

We bucked the trend in 2015.  This year (dark blue line) was pretty close to normal until March when we went through a six-week dry spell. Storms and cool weather in April and May helped pull the snowpack upwards until we peaked on May 24, which is about a month after our normal peak day.  Although we are now sliding into snowmelt season, you can see that our current snowpack, as a percent, is 309% greater than a normal year.

South Platte River Bason Time Series Snowpack Summary Graph

After looking at our snowpack levels, I frequently cruise over to the United States Geological Survey (USGS) Water Information System website, which tracks real-time water levels in rivers and creeks nationwide.  Their top-level map depicts river and creek water levels as a percentile of that day over a 30-year average.  Just as any good data-driven exhibit should do, their national map makes it easy to see major trends and patterns. Rivers in the Pacific Coast states are very low, with many running below the 10th percentile.  The Great Plains states have experienced a very wet spring and early summer, and rivers here are generally running very high compared to normal.  The Appalachia and Atlantic Coast rivers are normal or low.

River Levels Map

Zooming into the Denver area, we see our river levels are very high, with most above the 90th percentile.  By clicking on a dot (which represent water level gauges), you can get information about a specific sites.  For example, the chart below shows water levels of the South Platte River in Englewood, CO, a suburb of Denver. You can see two steep spikes on June 5th, one in the early morning and one in the evening, that were caused by strong storms. Evidence of the first storm was obvious to the residents in Southeast Denver, who woke up that morning to the sound of front-end loaders moving four feet of hail off their streets.

South Platte River Levels_June

While the June 5th spikes are related to localized rainstorms, the melting snowpack is also influencing water levels.  If we extend the graph from seven days to 60 days, we see that water levels have consistently climbed since early April, due to spring rainstorms and melting snowpack.  The spikes of June 5th, which looked so dramatic in the top graph, become just one of many steep spikes the river has seen in the past couple of months.

South Platte River Levels_Spring

So who cares about all this anyway?  Well, you might care if you operate a tourism business in Colorado.  For example, organizers of running races scheduled to take place in the high mountains changed their routes due to snow covered trails, and two popular driving roads, Trail Ridge Road in Rocky Mountain National Park and Mt. Evans Road in the Arapahoe-Roosevelt National Forest, were both closed over Memorial Day weekend (their traditional opening day).  Besides the recreation and tourism industry, other professionals such as utility managers, city planners, emergency respondents, water companies, and farmers all depend on snowpack and river levels.  While Coloradoans will continue to experience wild weather, these two websites can help us all understand and react to it.


Corona Summer Camp 2015: AAPOR

Ah, summer camp. For those of us who were generally allergic to the outdoors as kids, summer camp did not necessarily mean cabins, building fires, and outdoor recreation. In my case, summer camp usually meant summer orchestra, which was the best. Not only did I get to play music for hours every day, I also got to spend time with a bunch of other orchestra kids who would not even bat an eye if I wanted to discuss which Suzuki book they were on or the pros and cons of catgut strings.

In many ways, AAPOR feels similar to those years of summer orchestra—conference attendees discuss sampling issues and response rates as if they were discussing the weather. This year over one thousand people met at AAPOR to discuss all the intricacies of public opinion research. Beth and I had a really difficult time choosing which talks to go to because they all sounded so relevant and interesting.

This conference in particular gives us a good sense of the changing landscape of survey research and of the big issues in the field. One big issue was the mode to use for survey research. Phone surveys have been the gold standard for opinion research for a while, but with the increasing rate of cell phone use (especially cell phone only households) and the decreasing response rates, it has become harder and more expensive to maintain phone survey quality. There was at least one presentation about a survey that has transitioned over to cell phone only. Other talks discussed the quality of online panel surveys or using multiple survey modes (e.g., a paper survey with phone, an online survey with paper, etc.) to increase response rates.

Another big issue was incorporating all the available data into public opinion research. New research has started looking at whether social media and other big datasets can be analyzed as another way of measuring public opinion.  A final big issue had to do with transparency in survey research. Although there is a ton of survey data floating around, it is not always clear who is following correct and ethical survey methodology standards and who is not. Moreover, a lot of survey data reports do not include enough information to even judge what standards are being used. AAPOR’s Transparency Initiative encourages organizations to include with public release of data certain survey methodology information that allows people to judge the quality of the data.

Beth and Kate w Nate Silver at AAPOR 2015On top of all the great discussions of survey methodologyy and of the future of public opinion research, Beth and I also got to meet (briefly) Nate Silver, who received an award from AAPOR this year. (Someone may have texted this photo to her mom.) After the awards reception, AAPOR held a casino night to raise money for student awards. Silver participated in the poker tournament and kindly posed for many photos. He also has posted his own thoughts  about some issues raised at the conference.