Parents' knowledge and predictions about the age of menarche: experimental evidence from Honduras.

BACKGROUND: Access to accurate, timely and age-appropriate information about menarche is an essential part of menstrual health. Reliable evidence shows that girls primarily obtain information from their mothers and/or other female family members, therefore, it is important to determine parents' knowledge and their predictions about other parents' knowledge of the age of menarche. METHODS: To this end, we performed a pre-registered study with data collected from 360 households in Santa Rosa de Copán, Honduras. We implemented a novel procedure to avoid social desirability bias whereby participants answered two separated questions: i) their knowledge about the age of menarche (self-report) and ii) to predict or guess the modal response of the other participants regarding the same question (modal guess). Participants were paid according to accuracy. Both questions appeared randomly in the survey. RESULTS: Recent studies indicate the age of menarche at 12 years old and 56.11% of the sample gave the same response while 62.78% hit the modal value. We estimated the impact of different sociodemographic variables and found only marginal differences. Interestingly, people with formal education and women tend to respond with lower predictions. CONCLUSION: Parents' knowledge about the age of menarche is high in the study area. The study also found that there was no social desirability bias.

Continuous and binary sets of responses differ in the field.

This paper conducts a pre-registered study aimed to compare binary and continuous set of responses in survey questionnaires. Binary responses consist of two possible opposing response options (Yes/No). Continuous responses are numerical, where respondents can indicate an option on a 0-10 horizontal blind line. We study whether feasible sets of binary and continuous responses yield the same outcome (distribution) and have the same cost (duration in minutes). We collect data from 360 households in Honduras that were randomly assigned to Yes/No questions or given a slider (0-10 visual scale) to mark their responses, therefore, we provide causal evidence. We find that respondents are 13% more likely to respond "Yes" and spend 2.1 min less in the binary setting. Additionally, the results suggest that the type of question matters.