TalkBank Software

TalkBank Programs

Currently there are two publicly available tools for the creation and analysis of TalkBank data. They are:

CLAN is currently the most general tool available for transcription, coding, and analysis. You can also download source code from github.com for OSX CLAN or Windows CLAN .

Phon is design to facilitate the analysis of phonological and phonetic data transcribed in CHAT. It is fully compatible and interoperable with CLAN.

XML Utilities

TalkBank also provides utilities for maintaining and serving XML versions of the database.

Chatter is a cross-platform Java application that performs strict validation of CHAT transcripts and also can convert CHAT to XML conforming to our XML schema.

It may be helpful to review the XML Schema for TalkBank data, along with documentation of the Schema.

Interoperable Programs

Praat is a program for phonological and phonetic analysis that complements the features of Phon. CHAT files can be exported to and imported from Praat TextGrid files by using the PRAAT2CHAT command in CLAN, along with an attribs.cut file that declares the names of the various Praat tiers. The praat2chat folder in the /examples folder in the CLAN distribution provides examples and an explanation, as does the CLAN manual and the tutorial screencast. The Phon program, which is fully interoperable with CLAN, also provides complete integration with Praat.

ELAN is an open source, full function, stable Partitur (sound score) editor that is excellent for analysis of gesture from video and conversational overlap. Using the CHAT2ELAN and ELAN2CHAT programs in CLAN, it is possible to reformat data for complete interoperability between ELAN and CLAN together. The elan2chat folder in the /examples folder in the CLAN distribution provides examples and an explanation, as does the CLAN manual and the tutorial screencast.

DataVyu is linked to the Databrary system. We are currently working on developing a closer integration of Databrary and TalkBank.

QDA Coding Programs

Neither CLAN nor ELAN currently provide extensive methods for coding blocks of data, including textual data, in accord with qualitative data analysis (QDA) methodology. CLAN does provide methods such as headers, gems, comments, and postcodes that can be used for some aspects of QDA. However, it does not implement a general model. Popular programs for fuller QDA are:

Transana provides a subset of the features in CLAN with a nicer user interface and additional facilities for user-defined coding. Although development of Transana was supported by the TalkBank Project, it has not yet been possible for Transana developers to implement importing and exporting of TalkBank CHAT data, although we continue to work on this level of compatibility. Transana is available for a modest cost for both Mac and Windows.

NVivo7 is probably the most widely used QDA program. However, NVivo7 cannot process video and cannot import from or export to either XML or CHAT. Fortunately, the core features of NVivo7 are also available in Transana. Users are encouraged to test out NVivo7 using the 30-day trial version. $495 for a single-user Windows-only educational license.

ATLAS.ti is similar in many ways to NVivo, although it does provide support for XML input and output. Like NVivo, ATLAS.ti is designed for the analysis of text and cannot be linked to video. A demo version is available. $892 for a single-user Windows-only educational license.

The Observer is probably the most powerful program available for behavior analysis. It is primarily designed for researchers who want to code and mark behaviors without also creating a transcript. We recommend CLAN, ELAN, and Transana for users who want to create transcripts. However, some researchers may find this powerful tool useful. Pricing is complex, but you can watch a demo short tour from the Noldus website.

Interact is a major competitor to The Observer, providing many similar features.

HyperRESEARCH provides and easy to use interface for the qualitative analysis of text, images, audio and video data. Audio and video sources do not need to be transcribed for behaviors to be coded for analysis. A companion tool, HyperTRANSCRIBE, is designed specifically for rapid transcription of audio and video materials if desired. Both HyperRESEARCH and HyperTRANSCRIBE are available for both Mac and Windows for $370 or at a student price of $200.

Relevant programs that appear to no longer be available or no longer under development include VPRISM, LessonLab, and Annotater.