Third Update: Coffeehouse Communities

Unfortunately, this post is coming later than I would have liked or intended – however the end of the summer brought challenges that I did not foresee. However, despite surgery, a lost SPSS license, and W&M orientation, my research was able to get back on track at the start of the school year! I spent the last week and a half, in between life and classes, analyzing the data gathered from the 500+ survey responses that I received at the beginning of August. The first thing that I did was create a Variable Table that organized my data into Variable, Question Text, Data Type (nominal/ordinal/scale), and Response Options (text, 1-7, checkboxes, etc.). This allowed subsequent analysis to move much faster, and gave me a tool to use to help explain the survey and analysis to my peers.

Using IBM’s SPSS student pack, I was able to take my survey data and analyze it so as to answer questions that pertained to my research question: do scale effects and new urbanism affect sense of community in coffeehouses, and how do these variables play into satisfaction/Net Promoter Score and other variables that marketers and business owners look for from their customers?

After cleaning the data and removing outliers, the first test that I ran was a Cronbach Alpha test to measure the internal consistency of the indexes created and used in the survey: New Urbanism (NU_Index – 19 items), Sense of Community (SoC_Index – 15 items), and satisfaction (SATIS_Index – 5 items). I was worried about the scores I would received, especially because NU_Index contained 19 separate questions all designed to measure the same thing, but could easily be misinterpreted by respondents into answers that would throw off the consistency. This however, was not the case. SATIS received an alpha score of .837, NU of .738, and SoC of .896 – all of which were above the .70 I needed to be able to use these indexes in my study.

I wanted to determine first how the respondent percept of a “chain” versus “unique” coffeehouse affected the SoC created for the consumer. Looking first at the descriptive statistics, the means of Sense of Community clearly differed whether a respondent answered “unique”, “handful”, “unsure”, or “chain” in regard to the scale of their chosen coffeehouse. Through a One Way Analysis of Variance (ANOVA) with a statistically significant (p<.001) F-value of  16.041, I determined with a post-hoc Sheffe test that the “unique” SoC mean differed from the “chain” SoC mean in a statistically significant way (p<.001) with a mean difference of .46238. Therefore, a consumer was more likely to have a higher feeling of sense of community if they perceived their coffeehouse to be unique in and of itself.

I also wanted to look at the correlation that may have existed between Sense of Community and the Net Promoter Score – meaning the likelihood that a consumer would then recommend that coffeehouse to someone else. What I found was a strong correlation (with p<.001) that yielded a pearson correlation of .551. This led me to conclude that there is a positive relationship between SoC and the likelihood that a consumer will promote their chosen “third place”. This could validate the marketers at Starbucks and other chains that work to create a neighborhood feel in their shops at every corner.

Finally (though I ran other tests out of personal interest and for the sake of my research), the other test that I wanted to share was the linear regression analysis that I ran using the SoC, NU, and SATIS variables. I wanted to see how satisfaction with the locale, decor, customers, and staff as well as the level of New Urbanism environment present outside the coffeehouse acted as functions of the Sense of Community created for the respondents. With statistical significance (p<.001), Y(SoC) = .065 + .242(NU) + .414(SATIS). Both New Urbanism and Satisfaction positively facilitate the created Sense of Community in these society third places, though satisfaction has a larger effect.

There are certainly more conclusions drawn from this data, but these examples illustrate the strength of the data set thus far in helping to support my hypotheses on SoC, NU, satisfaction, NPS, and other relevant concepts to the marketers and sociologists that would care about this research. My final post will be summarizing my thoughts and findings throughout the research period, and my struggles to make it to where I have. Stay tuned!