TALLY 3.0 (IBM-based)

Reviewed by Robert Trotter, II

Anthropology Department, Northern Arizona University

Originally posted on Anthro-l, 1992
TALLY 3.0 is a text oriented ethnographic data manager and a content analysis tool. Its primary purpose is to allow researchers to interactively code segments of text for future reference, for searches, for cut and paste into articles, and for several types of ethnographic, content, and thematic analysis operations. TALLY 3.0 is particularly valuable for the management and examination of field notes, open ended questionnaires, interview transcripts, or any form of machine readable text, including secondary data scanned into computer files through optical character reading (OCR) technology. However, be aware that TALLY 3.0 has trouble handling files that are larger than one megabite in size.

TALLY works from simple menus and contains excellent help files that will pop-up from anywhere in the program with contextual help information. If the user is familiar with the commands used in the program, these can be used in lieu of the menus. This combination of menus and commands makes the program both easy to learn and fast to use.

TALLY utilizes ASCII text files, which can be of any size, based on available storage space in your computer. The files do not need special preparation or formatting. The program allows researchers to interactively code text at varying theoretical levels, and then manipulate the data interactively. The minimum level of coding is the single word, which makes TALLY useful for a number of cognitive anthropology and linguistic anthropology purposes. The maximum level of coding is the whole file, with any level of segmentation possible in between. Codes can be nested or grouped, so that multiple codes can be recorded for the same text segment. This allows for complex queries and search options that support the use of boolean operators and sequence sensitivity for coded segments. The codes are stored in DIF format for users who wish to perform further statistical analyses.

The coding system is simple to use. You either create a code file on your word processor, or you can use TALLY's mnemonics editor. You can invent up to 256 codes in a single coding file, and can link multiple coded text files together for analysis, if necessary. We use several code files for each text file, to allow us to take advantage of analyzing the text at varying levels or from multiple theoretical constructions. A code may be any combination of letters and numbers, up to a total of eight characters. Each code has a forty character description line associated with it, to assist in identifying the proper code and to reduce intercoder variability. When coding text, you can pop- up the code list from inside the text coding section of the program, search for the proper code, and with the stroke of a key you can enter it into the coded file at any selected position. The program allows for rapid global searches for single codes. And, it allows for global recodes, or targeted recodes. TALLY's coding operation is a great time saver over all of the other ethnographic data base managers that we have tested.

TALLY helps organize and manipulate ethnographic data for the types of analysis that formerly were accomplished (slowly) by physical and mental juxtaposition of text segments. One of the most onerous used to be cut-and-paste operations. With TALLY, when you find a particularly exemplary quotation, you can easily cut and paste that segment into a separate file, which can later be imported directly into an article using any word processor. You can interactively add comments to cut-and-paste segments. You can also decide whether or not to save the contextual information that is available for that segment (text file name, code file name, and a contextual statement) for either future reference or analysis.

In addition to the ethnographic data management functions, TALLY has a built-in content analysis program that we are finding very useful. You can request reports of the number and cumulative percentages of codes for any file, or any set of files. TALLY counts the number of times each code was used in the file, and notes the percentage of that code to all of the codes you have created. For example, we have coded focus group verbatim transcripts on alcohol, drugs and AIDS risks for teenagers, then compared the males and females for the number of times that they mentioned certain kinds of risks. As would be expected, they differ. TALLY allows us to identify and analyze those differences enormously faster than would be possible by simply doing a hand search of the text and counting the codes ourselves. It frees the ethnographer to spend time on analysis rather than scut work.

The report section of the program also allows you to perform a global cut and paste operation which pulls out all of the coded segments of the file and either prints them out as hard copy or saves them to an ASCII file. You can request that all segmented information be included, or only the first line of each segment, and you can determine how the nested codes will be handled (inclusively or exclusively). This operation orders each coded segment so all segments identified by a single code are printed or saved together. This is a significant time saver over having to search back and forth through the files, and it can be helpful in a number of different analytical functions that can be performed on ethnographic data.

The report section of TALLY has another nice feature. The Matrix Report function can be used to analyze open ended questions on questionnaires, or open ended questions asked of a series of informants in ethnographic interviews. The program creates a profile of each informant, by counting whether each code in your code set is present or absent for a particular question. It takes each of these individual code profiles and creates a person-by-code matrix which is saved as an ascii file. That file can be converted into a distance matrix by any statistical package. The distance matrix is then available for statistical analyses for underlying organizational principles. We have used it for cluster analysis and multiple dimensional scaling of both the individuals and the coded variables. This is a handy confirmatory device that can be used in conjunction with a wide variety of ethnographic explorations.

In addition to its other functions, TALLY 3.0 allows you to print out hard copy of all of the important files, such as the code file (providing you with an instant code book), the text file, and the coded text file (with or without line numbers). These can be used as backup, or to assist in the analysis of your data. It also has utilities that allow you to easily escape to DOS, then get back in the program, and has several utilities (renaming files, copying files, erasing files, etc.) that can be accomplished from inside of TALLY, without having to exit the program.

The author of TALLY 3.0 is Jeffrey W. Bowyer. It is now available from Wm. C. Brown Publishers in Dubuque, Iowa, for $49.95, including the user manual.