Published

June 11, 2026

EGAP Logo

EGAP Learning Days 2026 · Abidjan, Côte d’Ivoire

The Effect of Deadline Announcement
on Task Completion Rate

Evidence from EGAP Learning Days Participants

Ouattara N Ben Idriss Godwin UDDON Godson NWANKWO Davis Mawuena Aweso Abdul-Hanan

June 11, 2026 · EGAP Learning Days, Abidjan


SLIDE 2 High-level question

Research Question

Does receiving a deadline reminder increase task completion rates among workshop participants?

Proposed Answer

We hypothesize that participants who receive a deadline reminder will be more likely to complete the assigned task than participants who do not receive a reminder.

2 / 7


SLIDE 3 Research Design — iMIDA

i · Inquiry (abstract) M · Model I · Inquiry (precise) D · Data Strategy A · Answer Strategy

i — How can simple behavioral interventions improve task completion?

Many individuals intend to complete assigned tasks but fail to do so because of forgetfulness, procrastination, or competing priorities. Understanding how to encourage follow-through is important for improving participation and productivity.

M — We propose that deadline reminders increase task completion by making the task more salient and reducing forgetting. Deadline Reminder –> Greater Attention to the Task –> Less Forgetting / Less Procrastination –> Higher Task Completion

I — Does receiving a deadline reminder increase task completion rates among workshop participants?

D — Randomized Controlled Trial (RCT), coin-flip assignment. Treatment Group Receives a deadline reminder before the submission deadline.Control Group Receives no reminder. Outcome: binary completion (1 = submitted; 0 = not).

A — Estimate the causal effect of reminders by comparing completion rates between treatment and control groups. \(ATE= \bar{Y}_T-\bar{Y}_C\)

3 / 7


SLIDE 4 Implementation Report

WHAT We conducted a small randomized controlled trial (RCT) to evaluate whether receiving a deadline reminder increases the likelihood that workshop participants complete an assigned task.

HOW Participants were randomly assigned to one of two groups:

Treatment group: received a deadline reminder before the submission deadline. Control group: did not receive any reminder.

After the deadline, we recorded whether each participant completed the task.

TO WHOM The study was conducted among workshop participants.

WHERE The experiment was implemented within the context of the workshop.

WHEN The intervention was implemented during the workshop period

WHY THIS WAY We used random assignment to ensure that the treatment and control groups were comparable. This design allows us to estimate the causal effect of deadline reminders on task completion while minimizing selection bias.

WHAT WENT BADLY The sample size was relatively small, reducing statistical power.

4 / 7


SLIDE 5 Answering Strategy

Estimator

Difference-in-means (OLS):

Y = β₀ + β₁ · X + ε

  • Y = task completion (0 / 1)
  • X = deadline treatment (0 = control, 1 = treated)
  • β₁ = estimated ATE

Uncertainty Estimation

We report 95% confidence intervals and p-values.

5 / 7


SLIDE 6 Results — Regression Output

lm_robust(Y ~ X, data = dataa) — HC2 Robust Standard Errors
  Model 1
(Intercept) 0.80*
  [ 0.52; 1.08]
X (Deadline) 0.20
  [-0.08; 0.48]
R2 0.11
Adj. R2 0.05
Num. obs. 19
RMSE  
HC2 heteroskedasticity-robust standard errors in parentheses. * 0 outside the confidence interval.

Intercept = completion rate in the control group (no deadline announced).

Coefficient on X = estimated ATE — the average effect of announcing a deadline on completion rate.

6 / 7


SLIDE 7 Interpretation, Threats & Takeaways

Interpretation

Findings

  • Control group completion rate: 80%

  • Estimated treatment effect: +20 percentage points

  • p-value: 0.152

  • 95% CI: [-0.08, 0.48]

c/c :Although participants who received a deadline reminder had a higher task completion rate, the estimated effect is not statistically significant (p = 0.152). Therefore, we fail to reject the null hypothesis that reminders have no effect on task completion.

Key Takeaway

Threats to Inference

  • Low power — DF = 17, too small to detect modest effects
  • Ceiling effect — 80% control rate leaves little room to improve
  • Spillovers — participants may have shared deadline info (SUTVA)
  • Hawthorne effect — observation itself may have raised both groups’ performance

7 / 7


SLIDE 8 Thank You

Thank you for your attention!

A special thank you to our instructors — we truly appreciate your guidance and dedication.

Instructors photo

Vincent · Alyssa · Adikath · Yannick · Brice · Mac · Vin

8 / 8