Table 7

Return to Signaling by Project and Test Types

Dependent Variable
Project Value (Log)Number of ProjectsNumber of Projects > 0Earnings
(1)(2)(3)(4)
Panel A: Only Including Design Microcredentials
Number of microcredentials−0.0240.0050.0039.864**
(0.071)(0.004)(0.002)(3.698)
Number of microcredentials × design project0.0570.052***0.027***6.208***
(0.044)(0.007)(0.003)(1.364)
Observations3,72641,44041,44041,440
Panel B: Only Including Finance Microcredentials
Number of microcredentials0.277−0.006−0.005*2.322
(0.196)(0.004)(0.003)(2.669)
Number of microcredentials × finance project−0.0560.026***0.018***10.450***
(0.050)(0.004)(0.003)(2.215)
Observations74312,89212,89212,892
Panel C: Only Including Sales and Marketing Microcredentials
Number of microcredentials0.1410. 012*0.00423.136+
(0.136)(0.004)(0.003)(13.439)
Number of microcredentials × sales and marketing project0.1250.035***0.026***−3.566
(0.210)(0.004)(0.003)(15.688)
Observations18,86928,88428,88428,884
Panel D: Only Including Technology Microcredentials
Number of microcredentials0.190****0.0002−0.002**3.271+
(0.052)(0.001)(0.001)(1.957)
Number of microcredentials × technology project−0.049+0.020**0.012***13.376***
(0.025)(0.002)(0.001)(2.204)
Observations10,776165,396165,396165,396
Panel E: Only Including Virtual Assistant Microcredentials
Number of microcredentials0.1800.040***0.025***11.860+
(0.272)(0.007)(0.004)(6.639)
Number of microcredentials × virtual assistant project−0.0460.0120.009+21.582*
(0.260)(0.008)(0.005)(8.856)
Observations98818,58418,58418,584
Panel F: Only Including Writing and Translation Microcredentials
Number of microcredentials0.0420.006*0.003*6.435***
(0.054)(0.003)(0.001)(1.677)
Number of microcredentials × writing and translation project0.0190.040***0.022***7.473***
(0.025)(0.004)(0.002)(1.178)
Observations8,394132,272132,272132,272
  • Notes: Estimates are from the fixed‐effects event‐study specification. In Column 1 the unit of observation is one project. In Columns 2–4, the unit of observation is a 14 day pre‐ or post‐test period. In addition to the variables reported, all models include year dummies, an average rating for completed projects, a dummy variable with the value 1 if the freelancer does not have ratings from past projects, cumulative (arsinh‐transformed) dollars earned on the platform, the cumulative number of completed projects on the platform, measured at the time of project start, and event fixed effects. Standard errors are clustered on the worker level. The significance levels in all specifications are: *** = 0.1%, ** = 1%, * = 5%, and + = 10%.