Meeting Recap and To-Do List

[Jannel Alston]

Hey Phonatics,

So towards the end and after the meeting Stephanie and I played with the data and created some segments that are quantifiably justified YAY! Follow along on the attached excel sheet. What we did was:

1. Used the attitude measurement questions from the questionnaire to create factor loadings [Factor Analysis Sheet]
2. Determined the statistically significant attitudes (with the highest variance among factors) within each factor loading and used them to identify each attribute (Techie, Organized, Self-Expressionist, App Addict) [Factor Analysis Sheet]
3. Gave each of these attitudes a weighted average and organized the data to determine which respondents classified as being high (1) or low (0) in each factor loading [Columns BQ-BT on the Segmentation Sheet]. Since the available numbers for respondents ranged from 1-5, we chose 1-2.5 to be low and 3.5-5 to be high, eliminating the chance of them being "neither high nor low".
4. Cross-referenced all the factors to determine the actual segments, i.e. people who scored similarly as high or low on an attribute across all four attributes [Column BU]. Here we found four segments with >30 sample respondents: Average Joe (0000), Efficient Connector (0011), Organized Techie (1010), and Double-Edged Techie (both organized and socially connected) (1110).
   
5. Next we scaled the number of attributes per segment so they could be compared equally [Columns BV-BY on the Segmentation Sheet] and compiled these numbers together with the INTENT, LIKE, CONSIDER, etc. questions, and what we thought were relevant demographic questions [CorrSegmentation Sheet].
6. Then we did a Correlation Table [CorrTable Sheet]. The second one was just to see which were statistically significant, even if they weren't strong in terms of correlation. At A80 you will find the strong correlations we found in the table (greater than or equal to |.30| ), but look through again to make sure we didn't miss any. 

So now the to-do list. Each of the four segments need:

• ANOVA Analysis
• A more detailed demographic breakdown of each segment (to be used in messaging). May also mean chi-square.
• Messages!

Steph if I got any of this wrong let us know.

Until we meet again, everyone should stake a claim on what analysis they'll run. Everyone should also brainstorm messages.

I believe most people said the earliest they could meet was Wednesday evening? I'll send out a When2Meet Monday or so.

Jannel

[Stephanie A. Chuah]

Thanks Jannel, one correction, we didn't adjust for 'neither high nor low'. Right now it's coded as >3.5 is 1, <3.5 is low. I'll make adjustments tonight to see what we get.

--
 
 

--
Stephanie A. Chuah
The Wharton School, University of Pennsylvania
stephani...@gmail.com |  +1 (215) 919-1574

[Stephanie A. Chuah]

Hi Everyone,

I'm looking at the data now with the adjusted codes-- There are some ideas I'd like to run with some people. Is anyone free tonight for an hour after 9pm to meet in Huntsman real quick?

-Steph

[Maryam Mir]

Yep I'm free at 9pm

--
 
 

[Stephanie A. Chuah]

Cool. All the GSRs are booked, but let's meet at computer lab 375?

I'll be in Huntsman anyway with another group, so let me know if you're running late or something.

-Steph

[Maryam Mir]

Could we meet at 9.30 instead? I just have to finish a paper or if possible at the Starbucks under commons

Sent from my iPhone

--
 
 

[Stephanie A. Chuah]

9.30 it is!

--
 
 

--
 
 

[Stephanie A. Chuah]

Computer lab's pretty full-- I'm on the second floor by a computer near 270.

[Maryam Mir]

What about in the forum?

--
 
 

[Jannel Alston]

What questions are we asking the professor?

--
 
 

[Stephanie A. Chuah]

I'm working with the data-- the plan is for me to ask him questions about where to take our analysis with what we've got so far and just run with it.

-Steph

[Jannel Alston]

I remember saying I would send out a When2Meet for this week.

http://www.when2meet.com/?669781-7NYrM

--