est.ss {JATSdecoder}

est.ss

Description

Function to estimate studies sample size by maximizing different conservative estimates. Performs four different extraction heuristics for sample sizes mentioned in abstract, text and statistical results.

Usage

est.ss(
  abstract = NULL,
  text = NULL,
  stats = NULL,
  standardStats = NULL,
  quantileDF = 0.9,
  max.only = FALSE,
  max.parts = TRUE
)

Arguments

abstract

an abstract text string.

text

the main text string to process (usually method and result sections). If text has content, arguments "stats" and "standardStats" are deactivated and filled with results by get.stats(text).

stats

statistics extracted with get.stats(x)$stats (only active if no text is submitted).

standardStats

standard statistics extracted with get.stats(x)$standardStats (only active if no text is submitted).

quantileDF

quantile of (df1-1)+(df2+2) to extract.

max.only

Logical. If TRUE only the final estimate will be returned, if FALSE all sub estimates are returned as well.

max.parts

Logical. If FALSE outputs all captured sample sizes in sub inputs.

Details

Sample size extraction from abstract:
- Extracts N= from abstract text and performs position-of-speech search with list of synonyms of sample units

Sample size extraction from text:
- Unifies and extracts textlines with age descriptions, than computes sum of hits as nage - Unifies and extracts all "numeric male-female" patterns than computes sum of first male/female hit - Unifies and extracts textlines with participant description than computes sum of first three hits as ntext

Sample size extraction from statistical results:
- Extracts "N=" in statistical results extracted with allStats() that contain p-value: e.g.: chi(2, N=12)=15.2, p<.05

Sample size extraction by degrees of freedom with result of standardStats(allStats()):
- Extracts df1 and df2 if possible and neither containing a ".", than calculates specified quantile of (df1+1)+(df2+2) (at least 2 group comparison assumed)

Value

Numeric vector with extracted sample sizes by input and estimated sample size.

See Also

study.character for extracting multiple study characteristics at once.

Examples

 a<-"One hundred twelve students participated in our study."
 est.ss(abstract=a)
 x<-"Our sample consists of three hundred twenty five undergraduate students.
     The F-test indicates significant differences in means F(2,102)=3.21, p<.05."
 est.ss(text=x)