Different Approaches to Entrepreneurial Decision-Making: A Field Experiment in Italy

Entrepreneurs will receive 8 sessions of online training. These sessions will include interactive lectures and coaching by qualified mentors/instructors each working with a subset of the startup sample. Treated and control startups will receive the same amount of training on entrepreneurial decision making (on topics like business model canvas, customers' interviews, minimum viable products/services, concierge/prototype, etc.). One set of treated startups will be taught how to make entrepreneurial decisions according to a scientific approach. Another set of treated startups will be taught how to make entrepreneurial decisions according to an effectuation approach. The third set of entrepreneurs will receive equivalent training without any specific approach. A final set of entrepreneurs will not receive any training.

2020-10-16

2021-02-13

1. Revenue: our main dependent variable is the amount of revenue (euros)
2. Dropout: this is a binary variable that takes value 0 until the firm drops out (they abandon the program and cease the startup), 1 in the time period in which the firm drops out, and it is a missing value thereafter. In order to avoid attrition biases, we will check that the entrepreneurs that leave the program will actually cease any entrepreneurial activity.
3. Pivot: this is the cumulative number of times in which a startup makes a major change to its business model. We defined a change to be major by analyzing whether the entrepreneur moved from the original idea to another idea that changed the core value proposition of the product or service.

The research team will also examine potential mechanisms behind these outcomes by collecting variables related to how and when entrepreneurs make decisions. The key mechanism being tested is precision and improvement in the precision of predictions.

The research team is also examining the interaction of the treatments with a number of covariates of interest, including:
a) Gender
b) Psychological traits
c) Perception of competition and information sharing
d) Management practices adopted by the entrepreneurial teams
e) Well-being
f) Accumulation of knowledge
g) Passion
h) Communication
i) Knowledge sharing among team members
j) Influence of external mentors
k) Role of experience and background in adopting different approaches

The RCT focuses on nascent entrepreneurs (defined as those in the process of starting a new business). We do not restrict participation to startups operating in one or a few industries. We advertise the program through digital channels as a general course overseeing all the important aspects of new venture creation. The course is free of charge in order to ensure the participation of firms with limited financial resources. Startups receive training sessions remotely through an online platform. The same amount of training is offered to all three groups in order to ensure comparability. Multiple instructors/coaches will be trained to deliver the training using standard materials, purposely designed for this study. Each instructor/coach teaches three classes (one subgroup of startups treated with the scientific approach, one subgroup of startups treated with an effectuation approach, and one subgroup of control startups). The research team designs, coordinates and oversees all the activities, ensuring that the learning modules and coaching activities are appropriately carried out by the instructors/coaches.

Before starting the training program we ask all the applicant startups to send a description of their business idea and the vitae of the founders. On the basis of this information, we categorized them across stages of development. We plan to select roughly 500 entrepreneurs running early-stage start-ups that complete a baseline survey and a brief call to assess the levels of scientific decision-making and effectual decision-making before the intervention takes place. All participants that complete the baseline data collection process are admitted to the training program and are randomly assigned to one of the conditions. In order to avoid contamination, treated and control startups are taught/coached in different time slots of the same day (morning and afternoon) or on different days of the week. All communications to the three groups of startups are distinct and kept separate to avoid contamination. Individual-level, team level, and performance data are gathered periodically through telephone interviews by a team of 30 research assistants following a purposely defined research protocol. The data collection protocol includes a mix of closed-ended and open-ended questions covering startups’ performance, learning and behavioral aspects (level of adoption of a scientific approach to entrepreneurial decision making and effectuation). All entrepreneurs receive regular periodic phone calls by a research assistant.

Pure parallel randomization, done in office by a computer using STATA. After being randomly assigned into one of the four groups (pure control, control, effectuation or scientific), multiple subgroups of startups (of about 35 startups) will be created and randomly matched with the instructors/coaches so that each instructor/coach teaches three subgroups - of approximately 35 startups. Randomization checks: mean values comparison and t-tests across groups.

Start-up

No

Not clustered

500 startups X 12 (points in time) = 6000 observations

500 startups (125 in the scientific group, 125 in the effectuation group, 125 in the control group, 125 in the pure control group)

Minimum detectable effect size for main outcomes (accounting for sample design and clustering) Power calculations were conducted with GPower, assuming a very small effect (0.1) F tests - ANOVA: Repeated measures, within-between interaction Analysis: Post hoc: Compute achieved power Input: Effect size f = 0.1 α err prob = 0.05 Total sample size = 500 Number of groups = 2 Number of measurements = 4 Corr among rep measures = 0.5 Nonsphericity correction ε = 1 Output: Noncentrality parameter λ = 40.0000000 Critical F = 2.6108586 Numerator df = 3.0000000 Denominator df = 1494 Power (1-β err prob) = 0.9999133