Advanced

Thinking in probabilities: how natural frequencies facilitate understanding among Swedish medical students

von Schreeb, Sebastian LU (2017) PSYK11 20171
Department of Psychology
Abstract
The ability to draw diagnostic inferences from medical test results is important for future doctors. Earlier research has indicated that this ability may generally be inadequate. However, it can be improved by presenting statistics in a new format, so called natural frequencies, which is adapted to the way the human mind process probabilities. If the effect of this format is robust and generalizable, it may be useful in medical statistics education. The present study includes a sample of medical students in their final year of study. It is the first article in Sweden to investigate the effect of presenting information as natural frequencies. Furthermore, the study explores how individual differences in attitudes toward statistics may... (More)
The ability to draw diagnostic inferences from medical test results is important for future doctors. Earlier research has indicated that this ability may generally be inadequate. However, it can be improved by presenting statistics in a new format, so called natural frequencies, which is adapted to the way the human mind process probabilities. If the effect of this format is robust and generalizable, it may be useful in medical statistics education. The present study includes a sample of medical students in their final year of study. It is the first article in Sweden to investigate the effect of presenting information as natural frequencies. Furthermore, the study explores how individual differences in attitudes toward statistics may influence a potential advantage of natural frequencies, which has previously not been studied. In order to measure how attitudes predict the effect of natural frequencies, the Survey of Attitudes Toward Statistics was used. The study used an intra individual crossover design. Participants solved one task in which statistical information was presented in a natural frequencies format and a similar task in which information was presented using percentages, which is the common format for presenting probabilities. Results showed that 74 % of participants made a correct inference when statistical information was presented in natural frequencies, compared to 54 % when the task involved percentages. Only the Affect component of the attitude scale appeared a possible predictor of the advantageous effect of natural frequencies. The two findings are discussed in relation to other research on natural frequencies, which may have importance for future medical statistics education. (Less)
Abstract (Swedish)
Det är viktigt för blivande läkare att kunna göra korrekta riskbedömningar utifrån patienters testresultat. Tidigare forskning indikerar att den förmågan kan vara bristfällig, men att den går att förbättra genom ett nytt format att presentera statistik på. Det nya formatet kallas naturliga antal och är potentiellt användbart inom undervisning i medicinsk statistik. Den här studien är den första i Sverige att undersöka effekten av det nya formatet. Studien är även den första att undersöka om individuella skillnader i attityd mot statistik kan förutsäga effekten av naturliga antal. I utförandet användes en intra-individuell överkorsningsdesign, där deltagarna fick lösa en uppgift presenterad med naturliga antal samt en liknande uppgift där... (More)
Det är viktigt för blivande läkare att kunna göra korrekta riskbedömningar utifrån patienters testresultat. Tidigare forskning indikerar att den förmågan kan vara bristfällig, men att den går att förbättra genom ett nytt format att presentera statistik på. Det nya formatet kallas naturliga antal och är potentiellt användbart inom undervisning i medicinsk statistik. Den här studien är den första i Sverige att undersöka effekten av det nya formatet. Studien är även den första att undersöka om individuella skillnader i attityd mot statistik kan förutsäga effekten av naturliga antal. I utförandet användes en intra-individuell överkorsningsdesign, där deltagarna fick lösa en uppgift presenterad med naturliga antal samt en liknande uppgift där statistiken presenterades med procent - vilket är det allmänt utbredda formatet att presentera sannolikheter på. För att mäta deltagarnas attityder användes Survey of Attitudes Towards Statistcs. Resultaten visade att 74% av deltagarna gjorde en korrekt slutledning när uppgiften presenterades med naturliga antal, jämfört med 54% då uppgiften presenterades med procent. Endast komponenten affekt visades vara en potentiell prediktor för den fördelaktiga effekten av naturliga antal. Fynden diskuteras i relation till möjligheter för utveckling av medicinsk statistikundervisning. (Less)
Please use this url to cite or link to this publication:
author
von Schreeb, Sebastian LU
supervisor
organization
course
PSYK11 20171
year
type
M2 - Bachelor Degree
subject
keywords
Natural frequencies, Risk, Attitudes Toward Statistics, Sensitivity, Specificity, Bayesian reasoning, Medical education
language
English
id
8925053
date added to LUP
2017-09-08 15:16:31
date last changed
2017-09-08 15:16:31
@misc{8925053,
  abstract     = {The ability to draw diagnostic inferences from medical test results is important for future doctors. Earlier research has indicated that this ability may generally be inadequate. However, it can be improved by presenting statistics in a new format, so called natural frequencies, which is adapted to the way the human mind process probabilities. If the effect of this format is robust and generalizable, it may be useful in medical statistics education. The present study includes a sample of medical students in their final year of study. It is the first article in Sweden to investigate the effect of presenting information as natural frequencies. Furthermore, the study explores how individual differences in attitudes toward statistics may influence a potential advantage of natural frequencies, which has previously not been studied. In order to measure how attitudes predict the effect of natural frequencies, the Survey of Attitudes Toward Statistics was used. The study used an intra individual crossover design. Participants solved one task in which statistical information was presented in a natural frequencies format and a similar task in which information was presented using percentages, which is the common format for presenting probabilities. Results showed that 74 % of participants made a correct inference when statistical information was presented in natural frequencies, compared to 54 % when the task involved percentages. Only the Affect component of the attitude scale appeared a possible predictor of the advantageous effect of natural frequencies. The two findings are discussed in relation to other research on natural frequencies, which may have importance for future medical statistics education.},
  author       = {von Schreeb, Sebastian},
  keyword      = {Natural frequencies,Risk,Attitudes Toward Statistics,Sensitivity,Specificity,Bayesian reasoning,Medical education},
  language     = {eng},
  note         = {Student Paper},
  title        = {Thinking in probabilities: how natural frequencies facilitate understanding among Swedish medical students},
  year         = {2017},
}