Does negative marking lead to gendered exam taking behavior: experimental evidence from India

How to cite

Sharma, Swati; Sengupta, Nandana; Mukherjee, Sumitava. 2023. "Does negative marking lead to gendered exam taking behavior: experimental evidence from India." AEA RCT Registry. AEARCTR-0012042. https://doi.org/10.1257/rct.12042-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Swati Sharma* , Nandana Sengupta , Sumitava Mukherjee

* Primary investigator

Registry ID
AEARCTR-0012042
DOI
https://doi.org/10.1257/rct.12042-2.0
First registered
Last updated
Status
In development
Trial dates
2023-09-23 – 2024-05-05
Countries
Private
JEL codes
D90
Secondary IDs
NA
Registry record
http://www.socialscienceregistry.org/trials/12042

Abstract

Negative marking in high stake competitive examinations is often used as a tool to identify true ability or academic rigor of young aspirants across the globe. Literature from global North/developed countries shows that MCQ-negative marking-based screening process places women at a disadvantageous position because they skip more questions/guess lesser questions either because of differences in the degree of risk or loss aversion or confidence, controlling for ability/subject knowledge. Evidence from global South/developing countries is almost absent, moreover providing mixed results. In our lab experiment set up, we randomize marking scheme with half of subjects falling into no penalty for wrong answers marking scheme (control group) and others in marking scheme having 50% penalty for wrong answers (treatment group). Our sample consists of students pursuing graduation from one of the Institutes of Eminence from India as these institutes have a long history of skewed gender ratio in application and take up of technical STEM courses.

Intervention

This is a lab experiment where students would be invited to participate in an online quiz (effort based task). Half of the participants would be randomized in the control group (no penalty for wrong answers) and other half in the treatment group (minus half points from possible score of correct answer per incorrect question). Their payoff would be determined by their performance in the quiz.

Experimental design

It is a simple experimental design. Half of the participants will be randomly assigned into control group (no penalty for wrong answers) and other half in the treatment group (minus 50% of what could have been scored with correct answer per wrong answer). Subjects' pay off depends on the number of correct answers -penalty*number of wrong answers + fixed honorarium for participation.

Design details

Randomization method
Randomization done in office by computer. First sample will be stratified into male and female and 50% of females and 50% of males will be allocated to the treatment (other 50% to the control), thus achieving a total of 50% of sample in the treatment group.
Randomization unit
Individual
Clusters
NA
Observations
150 - 300 [per institute] students pursuing technical graduation course from 3 STEM institutes in India
Arms
65 -150 in No negative marking scheme (Control) and 65-150 in Negative marking scheme (Treatment) [per institute]
Primary outcomes
1. Total score, 2. Total payoff, 3. Total number of questions attempted, 4. Total number of questions with correct answers, 5. Total number of questions skipped, 6. Total number of questions with incorrect answers, 6. At question level – whether a question was answered correctly/skipped/incorrectly answered

Cited by

No citation data available for this registration yet.

Keywords