Struggling with research design? surveyframe is the R package that fixes it: write your analysis plan into the instrument once, and it runs exactly that. Free, on CRAN. Try it: https://t.co/YWBPBE8Hwo
#rstats#OpenScience
Copy the code in this post into RStudio and you get the exact same survey, table, and chart shown here. One survey question, declared in R, start to finish. https://t.co/YWBPBE8Hwo
#rstats#OpenScience#SurveyResearch
Congratulations to the #R-Core-Team on receiving the $1 million @RousseeuwPrize, often regarded as the Nobel Prize of Statistics.
R has become indispensable for analyzing and visualizing biological data.
https://t.co/DRlbURQNjy
#RStats#DataScience#OpenScience
surveyframe 0.3.2 is on CRAN. Write your analysis plan into
the instrument before a single response is collected.
The workflow is validated, reproducible, and citable from
the first question you write.
install.packages("surveyframe")
#rstats#OpenScience#SurveyResearch
AI ads are improving.
Consumers are ignoring them faster.
Some brands are now winning by showing less AI and more human.
The edge is no longer automation.
It’s credibility.
Correlation ≠ causation.
Two maps lining up does not prove meat causes cancer, or that vegetarian diets prevent it. Cancer rates also vary with age, screening, tobacco and alcohol use, pollution, infections, and reporting.
Test it with controlled analysis, not map-matching.
Do you see a higher prevalence of cancer in the states with more meat consumption in India?
Leaving out the Punjab and Himachal Pradesh exception (heavy fertilizer usage), it does seem like that to me.
The more vegetarian food, the lesser the cancer rate.
What do you think?
The brands winning right now? They picked a lane and pissed off half their potential customers.
Your "we're for everyone" messaging is exactly why nobody remembers you.
Good analysis is not about clever tools. It is about knowing when the data is not ready to speak yet. What minimum sample do you accept before acting?
#DecisionMaking
Most business decisions today rely on dashboards. Many of those dashboards hide a basic problem: small samples treated as big facts. A short thread on why this matters for everyday analysis.
#DataAnalysis
Report sample size clearly, show uncertainty rather than just averages, and rerun the test when the data doubles. These steps are dull, but they protect decisions.
#ResearchMethods
1/ Everyone is planning New Year offers. Very few are checking what actually worked last year. New plans fail when memory replaces evidence.
#NewYearPlanning