It seems to me that you need to identify your objectives before choosing an analysis. If it were my problem, my major concern would be whether a purchase was made or not. (How much money was spent could be another consideration but this depends quite largely on what was purchased at the expense of other predictors.) It seems to me that leaves you with a binary criterion variable with two values, purchase vs. no purchase, and multiple predictors, some pertaining to the store and others to the consumer. If your focus were on money expended, I'd go with multiple regression. However, if it's on purchase vs. no purchase, I'd go with logistic regression.

For multiple regression, having more than five or six predictors rarely adds much to the multiple correlation. Accordingly, I'd run separate analyses for store predictors and purchaser predictors, and then run a third analysis that combines the winning predictors from the other two. As regards logistic regression, I have less knowledge and experience about the role of number of predictors, but think I'd recommend the same general strategy.

IYH Dear Youssef Jebbar

IMHO your research calls for a mixed-methods approach, capitalizing on both qualitative and quantitative analysis.

A. Quantitative analysis plan:

1. *Proxemics*: Collect GPS coordinates, calculate straight-line distances, and derive travel times for participants.

2. *Visibility*: Survey consumer awareness of establishments, solicit ratings of visual attractiveness, and record frequency of promotional materials received.

3. *Accessibility*: Document opening hours, availability of entrances, compliance with disability guidelines, and ease of movement within stores.

4. *Transportation*: Tally nearby transit stops, gauge walking convenience, and catalog bus/train frequencies.

5. *Parking*: Count parking spots, monitor turnover, and compare pricing schemes with competitors'.

6. *Local Competition*: Pinpoint adjacent businesses, compute market penetration indices, and contrast product assortments.

7. Socioeconomic & Demographics: Record income, education, occupation, race, ethnicity, household composition, and life cycle status.

Applying descriptive statistics and factorial analysis makes perfect sense to probe correlations among variables.

Protip: Be sure to clean your data, treat missing values appropriately, and validate your measurement instruments.

B. (optional) Bolster w qualitative analysis:

1. Semi-Structured Interviews: Query shoppers on routines, preferences, motivations, and pain points. Dig deeper using open-ended queries supplemented with closed-response items.

2. Observational Studies: Track movements, actions, and interactions in situ to gain firsthand insights into customer behaviors and environmental influences.

3. Focus Group Discussions: Facilitate dialogues among patrons to collectively brainstorm and articulate opinions about shopping locations.

Why should you do this, if time permits? Triangulating data gleaned from these qualitative techniques lets you corroborate quantitative findings, generate fresh hypotheses, and unmask subtleties overlooked by numerical abstractions alone. You never know what people will tell you, it may completely surprise you and lead to new research.

Dear Professor,

I'm not clear what exactly you want to calculate. However, in this type of research, the goal is often prediction. Logistic regression allows you to predict the probability of choosing option 1 over 0, and the predictors can be both quantitative and nominal variables.

If the goal was to organize certain sets of variables, I thought about k-means analysis.

Regards, Natalia.

Find some published papers that have done what you would like to do and then use their methodology given the constraints on resources that you have.

You may want to look at

Store Choice, Store Location and Market Analysis by Neil Wrigley

and you could then do a citation search on that book; e.g.

https://scholar.google.com/scholar?oi=bibs&hl=en&cites=14895551374801630099&as_sdt=5

If you do want to use mixed methods, I suggest that you start with a design that will integrate the two set of results, rather than just reaching separate conclusions (e.g., one from testing hypotheses and the other from generating themes).

It sounds like you already have a substantial set of quantitative hypotheses and variables to measure them, so one likely mixed method approach would be an explanatory sequential design (QUAN --> qual). In this case, the goal of the qualitative follow-up study is to help understand the results form the primary quantitative study.