Second Update: Coffeehouse Communities

It’s been a while since my last update, so this one will serve mainly as a post to summarize what I’ve been working on for the past four weeks, and explain where I’ll be moving from here. I spent a few weeks working on creating the survey itself, which in the ended measured the satisfaction and Sense of Community a respondent had for their most-frequented coffeehouse, as well as a question set to measure the New Urbanist environment potentially surrounding the coffee shop. Through both convenience and random sampling both through social media, contacts, and Amazon Mechanical Turk, the survey returned 508 completed responses. An average time of 10min 1sec was taken to complete the survey.

I worked with Professor Michael Luchs to develop new ways to combat non-sampling respondent error in my distributed survey. First we randomized the question sets for every user when the open the Qualtrics survey, then we additionally embedded an attention check question into a long question set so as to use that answer to ensure that the respondent was taking the survey seriously. Amazon mTurk also uses rewards survey takers with a small payment as paid for through the survey requester, and in order for these respondents to receive their nominal fee, they must input a code found at the end of my survey. This code, through the help of my professor, was coded to have a timestamp embedded within a series of numbers. That way, I was able to make sure that the timestamp matched the time that the survey was complete (to stop potential abuses of the completed survey code). Very cool tricks of Qualtrics that I’m excited to carry into my future marketing research projects!

The bulk of the survey (ignoring questions needed for basic information and GIS data) dealt with satisfaction, Sense of Community, and New Urbanism. I used a modified version of a SoC Index developed by Professors McMillan and Chavez in order to gain insight into consumer perception of Sense of Community, and the questions were asked in a likert scale format in a 14 point question set. The New Urbanism set was 19 questions, and was based off of Patricia Patterson’s study of New Urbanism in residential communities and neighborhoods. Finally, I developed some general satisfaction questions and the classic Net Promoter Score question. Screenshots of a few example questions can be found here:

Screen Shot 2014-08-07 at 2.42.44 PM Screen Shot 2014-08-07 at 2.42.59 PM Screen Shot 2014-08-07 at 2.43.43 PM Screen Shot 2014-08-07 at 2.43.56 PM


The past few weeks were also spent understanding the basics of GIS – a program that now seems unlikely to provide me much insight into my current project, but I’m definitely looking forward to using in the future. I learned the basics of downloading, cropping, and creating layers and shapefiles, as well and how to address match the coffeehouses I was given the address of onto a US census map. That map can show me the population and residency information surrounding each address, such that I can tell whether each shop falls within certain requirements of New Urbanism. ┬áSince discovering the index created by Patricia Patterson, I can was able to have respondents self-report the area surrounding the shop – which has it’s drawbacks, but also conducts analysis through a more thorough set of questions.

Now that the survey responses have been collected, I am working to create a Variable Table to make understanding how the various questions fit into each variable very easy, and then I will be analyzing the data in SPSS to gain an answer to my questions originally posed. I’ll have an update next week on how that analysis goes!