Upload
emily-pitts
View
218
Download
4
Tags:
Embed Size (px)
Citation preview
KITCHEN CONFIDENTIAL?NORMS FOR THE USE OF TRANSFERRED KNOWLEDGE
IN GOURMET CUISINE
Giada Di StefanoAndrew King
Gianmario Verona
Our questions
• When legal property rights are not available, transferring knowledge is perilous. – When will knowledge holders reveal it to peers?– Do norms of knowledge use really allow transfer?– How do norms of knowledge use interact with
unilateral means of protection?
KITEs, Milan, March 22, 2012
Framing 1: Institutional Economics
• Exchange of ideas is perilous if protection mechanisms are unavailable.– Arrow’s Information Paradox
• Protection is accomplished by– Government legal protection (e.g. patent laws)– Unilateral protection (firm and knowledge attributes)
• New idea– Social norms can also provide protection for
transferred knowledge.
KITEs, Milan, March 22, 2012
MIT March 28, 2011
Framing 2: Nature of Norms
• Understanding in the Management Literature– Norms are “strong” when they engulf the context and
“weak” when the do not (Snyder and Ickes, 1985)
– “…I don't think you can have it both ways. Either strong knowledge sharing norms exist among chefs and they hold across contexts, or they are weak and subject to "substitution" effects with other IP related variables.” – Associate Editor on Paper 1.
KITEs, Milan, March 22, 2012
Framing 2: Nature of Norms
• Norms are universally binding.
• People think strategically about the application of norms, but norms retain power.
• Norms are just the terminology used for reflective strategic analysis of incentives.
KITEs, Milan, March 22, 2012
Empirical Setting
Gourmet Cuisine (i.e. fine dining)– Scarce applicability of formal IPRs:
– “How could you pay copyrights if you can just misplace a leaf on the plate and copyright would not be infringed anymore?”
– Creativity and innovation are considered important:– “Creativity comes first; then comes the customer” (F. Adrià)
• Existence of a system of social norms at play.
KITEs, Milan, March 22, 2012
Our Understanding
KITEs, Milan, March 22, 2012
H1 H2H3
Hypotheses• Hypothesis 1: Contextual transfer conditions influence the knowledge
holder’s expectation that a knowledge recipient will conform to norms for using transferred knowledge.
• Hypothesis 2: Accounting for differing contextual transfer conditions, the knowledge holder’s expectation that a knowledge recipient will conform to norms for using transferred knowledge will increase the likelihood that the knowledge holder will transfer knowledge to the recipient.
• Hypothesis 3: Access of the knowledge holder to unilateral protection mechanisms moderates the relationship between the knowledge holder’s expectation that a knowledge recipient will conform to norms for using transferred knowledge and the likelihood that the knowledge holder will transfer knowledge to the recipient.
KITEs, Milan, March 22, 2012
Interviews• Sharing
– I am not very good at pasta. This is my Achilles heel. I cannot make pasta very well. I call my friend, who makes the best pasta I know in town. And, you know, she is, "What do you mean you cannot make pasta? I have shown you like so many times." And she will get out her bag of tricks and tell them to me.
• Norms– I think that any of us, even when he takes the recipe of another, does it for taking
inspiration out of a detail, rather than replicating the entire recipe. He will never replicate the same recipe. That is, he will capture some ideas out of that recipe and will use them to do other things.
– If a chef I do not know calls me and asks me for a recipe, I would tell him: please, put my name on the menu. If he does not, then I would not give any more recipes to him in the future.
– There is one recipe that I do not give away. I have never given that away. I inherited that recipe. I feel like I am carrying that recipe, it does not really belong to me but I am carrying it.
• Sanctioning– You never want to burn any bridges in the industry because you might need something.– You might say to somebody who has worked for you for three years is thinking about
going to work for that guy. You might say, I would not recommend you go there. KITEs, Milan, March 22, 2012
Method• Qualitative exploration: detailed interviews:
Formal interviews: 8 Michelin-starred chefs in Milan. Informal interviews: 6 Michelin-starred chefs around Italy.Additional interviews: 9 top chefs in the U.S.
• Quantitative analysis: scenario-based experiment (Florey and Harrison, 2000; Gomez et al., 2000)– Trade-off between precision and fidelity.– Two scenarios per respondent.– Based on: (1) Interviews with top chefs; (2) Second round of
interviews for face validity; and (3) Pretest of 224 restaurants.– Response Rate: 21.12% (534 responses). Mainly male (82%), owners
(78%), 46 years old on average (23-80), better rated and more expensive.
KITEs, Milan, March 22, 2012
• Two scenarios per respondent: OLS regression with dummy for each respondent.
• Coefficient estimations are based on the differences created by the two scenarios.
• Classic “mediation” structure.
KITEs, Milan, March 22, 2012
MethodModel Specification
Identification StrategyEq. 1 Eq. 2i subjects, j scenarios, and k knowledge types
Differencing relative to subject (e.g. Eq. 3
Eq. 4
Differencing relative to subject and knowledge (e.g.:Eq. 5
Eq. 6
KITEs, Milan, March 22, 2012
MethodScenario-based experiment
KITEs, Milan, March 22, 2012
Measures
KITEs, Milan, March 22, 2012
Variable Measure OperationalizationDependent Variable
Knowledge Transfer
Likelihood that, if asked, the respondent would provide the requested knowledge to the chef in the scenario
7-point scale, where 1 is very unlikely and 7 is very likely.
Intermediate Variable
Normative Expectation
Likelihood that the chef in the scenario would conform to the three social norms regulating the use of transferred knowledge in the industry (i.e. “don’t copy”, “cite the source”, “don’t pass on”).
7-point scale, where 1 is very unlikely and 7 is very likely. Response for three norms combined (α=.73)
Independent Variables
KITEs, Milan, March 22, 2012
Context: Degree of Competition
Proximity The restaurant is geographically very X compared to the respondent.
ManipulatedX:Close = 1, Far = -1
PositioningThe restaurant has a cuisine style and an ambience that is very X to of the respondent.
ManipulatedX:Similar = 1, Different = -1
Context: Prestige of Recipient
Prestige The restaurant has X ratings and reviews.ManipulatedX:Excellent = 1, Average = -1
Context: Identifiable OriginSignature Transferred knowledge is a signature recipe. True = 1, False = -1
Protection: Knowledge Attributes
Technique Transferred knowledge is a cooking technique. True = 1, False = -1
Ephemerality Percentage of dishes in the current menu that were listed in the same month last year.
5-point scale: <25%, 25%, 50%, 75%, > 75%.
Protection: Complementary AssetsSuppliers Reliance on unique suppliers. True = 1, False = -1
Setting Unique location, setting, building, or atmosphere.* True = 1, False = -1
Italy: Michelin Guide 20092,529 restaurants: 2,254 non-starred (yellow), 236 one star (green), 34 two stars (red), and 5
three stars (black)
KITEs, Milan, March 22, 2012
Expectation of Norms
KITEs, Milan, March 22, 2012
Model 1 Model 2 coef se coef seContextual Transfer Conditions
Proximity -0.079*** 0.024 -0.071*** 0.023Positioning 0.007 0.025 -0.006 0.023Prestige 0.220*** 0.024 0.202*** 0.023Signature 0.042* 0.022 0.046** 0.022
Control Variables Technique -0.019 0.022 -0.015 0.022Review 0.047* 0.024 0.033 0.023Experience -0.002 0.024 -0.016 0.023Owner -0.092* 0.056Male -0.045 0.057Chain -0.113 0.096Tenure -0.010** 0.004Stars 0.149*** 0.057
_cons 3.726*** 0.021 4.079*** 0.155Model Stats F=15.512*** Chi2 =123.213*** R2 (ω) 0.041 0.041Hausman Test Not Passed
Likelihood of Transfer
KITEs, Milan, March 22, 2012
Model 1.3 Model 2.3 Model 1.5 Model 2.5 Model 2.1coef SE coef SE coef SE coef SE coef SE
Social Protect. MechanismsNormative Expectation (NE) 0.136*** 0.034 0.408*** 0.117 0.155*** 0.045 0.564*** 0.159 0.403*** 0.086Contextual Transfer Cond.Degree of Competition
Proximity -0.217*** 0.045 -0.220*** 0.045 -0.214*** 0.045 -0.217*** 0.045 -0.208*** 0.031Positioning -0.194** 0.090 -0.192** 0.089 -0.199** 0.090 -0.196** 0.089 -0.113* 0.063
Prestige of RecipientPrestige 0.142*** 0.048 0.140*** 0.047 0.138*** 0.048 0.137*** 0.047 0.112*** 0.031Identifiable OriginSignature -0.413*** 0.030 -0.416*** 0.030 -0.418*** 0.029Unilateral Protect. Mech.Knowledge Attributes
Technique 0.244*** 0.026 0.292*** 0.087 0.324*** 0.079Ephemerality 0.397*** 0.095
Complementary AssetsSuppliers -0.279** 0.119Setting -0.010 0.211
InteractionsTechnique * NE -0.013 0.022 0.011 0.030 -0.022 0.020Ephemerality * NE -0.062** 0.029 -0.084** 0.039 -0.065*** 0.021Suppliers * NE 0.033 0.036 -0.006 0.044 0.061** 0.027Setting * NE 0.097* 0.053 0.143* 0.078 0.039 0.044N 3,036 3,036 3,036 3,036 3,036F 43.908*** 29.803*** 10.149*** 6.697*** 689.960*** (Chi2)R2 (ω) 0.201 0.205 0.074 0.085 0.203
Conclusions
KITEs, Milan, March 22, 2012
Advantages and Limitations
• Mixed-method: richness and objectivity• Experimental treatment: improved test• Scenario-based experiment: precision and
fidelity
• We did not treat normative expectations.• We measured intended rather than real action• A single industry.
KITEs, Milan, March 22, 2012