Statcheck

<!---searched in commits on statcheck repo: `committer-date:2016-06-16..2016-08-16` --->

Small internal updates
* Import packages instead of Depending on them
* Using `message()` instead of `cat()`

<!---searched in commits on statcheck repo: `committer-date:2015-11-12..2016-06-16` --->

New features
* Make it optional to count p = .000 as an Error

Small updates
* Adapted plot function based on John Sakaluk's code. statcheck can now plot in APA style.
* Removed CopyPaste test; this function checked if the same string of results was reported multiple times in a paper or text and flagged it as a possible copy-paste error. However, this function wasn't very useful and therefore removed.
* Added axis limits to plot function so they won't get cut off when the re are no p-values > .5.

Bug fixes
* Updated regex for chi-square, so that it doesn't match t, F, or r with a subscript
* statcheck sometimes read t-tests in old PDFs as correlations, resulting in correlations >1. This caused an Error in statcheck, but is now ignored.
* In old PDFs "F(1, X) = Y" gets converted by pdftotext to "F(l, X) = Y". If this happens, convert "l" back into a "1". Thanks to Erika Salomon for pointing this out to me.

<!---searched in commits on statcheck repo: `committer-date:2014-07-01..2015-11-12` --->

New features
* Add option to choose whether or not to count p == alpha as significant or not
* Show pop-up window to select files for `checkPDF()` and `checkHTML()`

Small updates
* Improve search for subscripts in html
* Also find chi2 with thousand separators in N
* In the automated one-tailed test, search more specifically of "*one*-sided"", instead of "sided", etc.
* Also plot ns statistics

Bug fixes
* Fixed bugs in determining correct rounding
* If a result is not an error, it can also not be a decision error (this happened in some cases when reported p = .05)

<!---searched in commits on statcheck repo: `committer-date:2012-07-30..2014-07-01` --->

New features
* Added `diagnose()` to guess a probable cause for an error
* Recognize z-tests
* Calculate the APA factor for each article

Small updates
* Small updates to the regular expressions
* Recognize number of decimals to distinguish between p = .04 and p = .040
* Also extract and parse negative test statistics
* Also recognize inexact test statistics
* Also recognize values with thousand separators
* Also detect result reported as ns.
* Also recognize p-values in scientific notation
* Recognize more types of spacing in HTML
* Also accept .htm files instead of only .html
* Search for one-tailed tests in text
* Make it optional to count one-tailed tests as errors or not
* Add an option to assume that all tests are one-sided
* Add warnings for possible different significance levels
* Add the option to also search subdirectories
* Chris Hartgerink added inline documentation

Bug fixes
* Fixed bugs in extracting df for chi-square
* Fixed bug in plotting inexact p-values
* Fixed some bugs in identifying errors
* Better recognition of rounding errors
* Better recognition of p < 0
* Fixed bug so that statcheck doesn't crash if no results are extracted
* Don't read t *space* (df) as a chi-square
* Added Wald tests but removed them again as they were too buggy

Other updates
* Michele became maintainer of the package

<!---searched in commits on statcheck repo: `committer-date:2012-03-27..2012-07-30` --->

New features
* plot.statcheck() to plot statcheck object
* Include search for correlations and chi-square
* Search HTML files
* Search entire directories

Small updates
* Add progress bars

Bug fixes
* Fixed some bugs in summary function (na.rm = TRUE)

Other updates
* Michele Nuijten added as co-author

* First version