Warren Kindzierski

Large Language Models, LLMs, and Gen AI Technology are being used to improve the efficiency of research. LLMs are trained on vast amounts of data, including research papers, so they can provide a window into a research topic. Here a User queries ChatGPT, a LLM, on the nature of a Data Dredge.

The Southern California Children’s Health Study (CHS) was started in 1992 and continues to this time. Currently, it contains three cohorts recruited across 16 southern Californian communities and included in a follow study in adulthood:

• fourth graders (aged ~9-10 years) in 1992-1993 (cohort “C”)
• fourth graders (aged ~9-10 years) in 1995-1996 (cohort “D”)
• kindergarteners and first graders (aged ~5-7 years) in 2002-2003 (cohort “E”)

Beginning in their respective baseline year and continuing until high school graduation, annual or biannual questionnaires collected a range of information, including demographic characteristics, housing characteristics, prevalence and incidence of bronchitic symptoms, report of doctor-diagnosed asthma, family history of asthma, history of smoking in the household, and residential history.

Reanalysis of several meta-analysis papers dealing with the effects of lockdowns

Many Systematic Review and Meta-analysis, SRMA, studies based on observational studies are appearing in the literature. There is a need to be able to evaluate the reliability of a SRMA. Starting with the risk ratios and confidence limits, we compute a p-value for each study and examine the distribution of the p-values using a p-value plot. In this paper we examine four SRMAs dealing with the effects of COVID lockdowns on humans.

Many claims made by researchers often fail to replicate when tested rigorously. In short, these claims can be wrong. There is a need for students (and citizens) to understand how incorrect claims can come about in research by using questionable research practices. Ask a lot of questions of a data set and make a claim if any p-value is less than 0.05 is p-hacking. HARKing is making up a claim/narrative after looking at the data set. Statistical experts know about p-hacking and HARKing, but they appear to be largely silent. Some researchers know too but they ignore the problem. We present a hands-on demo about rolling ten-sided dice multiple times to show how incorrect claims come about. Several individuals executed simulations of p-values with ten-sided dice and show how easily a small p-value can come about by chance for a modest number of questions (rolls of the dice). Notably, small p-values found by one individual were not replicated by other individuals. These simple simulations allow students (and citizens) to better judge the reliability of a science claim when multiple questions are asked of a data set.