Preparing Data Summary Files
MS-V07c
CONCAT is used to create an ASCII (character) file of subject and/or item means and/or errors from one or more experimental data files. Data from these files can be output to the ASCII file in any combination or order that the user may require. As well as concatenating data from a number of different input files, the data may also be modified by the addition of constants or the addition or subtraction of data from one file condition to other file conditions.
Data is output in a series of matrices, each preceded by its own title. Within each matrix, data from different conditions and different files is in different vertical columns, and data from different subjects or items is in different horizontal rows. The output file is suitable for input to program PQRAOV which does a separate analysis of variance on each matrix.
Change the default drive (and directory if necessary) to that containing your DAT file(s). Your output file will be created on the same disk (and in the same directory) as the input file(s). Run the program by typing
CONCAT
at the terminal in response to the prompt. The program responds by typing its name, version number and the current date.
CONCAT will firstly ask you
DO YOU WANT A PRINTED RECORD OF THIS RUN? (N/Y):
If you reply "Y" (yes), all output to the terminal while CONCAT is running will also be reported on the printer. This is the best way to get printed output from CONCAT. If you use control-P you will get very messy output due to the presence of certain control characters in the output to the terminal.
You will be prompted to enter the name of 1 output file. Default extension for the output filename is DAS (DAta Summary). If the output filename you specify is the name of a file which already exists, you will be asked
OVERWRITE EXISTING FILE? (N/Y)
If you reply "Y" (yes), the existing file will be deleted and replaced by the new file to be output by this run of CONCAT. If you respond "N" (no), you will be asked to specify another name for the output file.
You will be prompted to enter up to 8 input filenames. These are the names of the DAT files from which subject and item means are to be copied. After each filename, press Return and you will be prompted for the next. When you have entered all the input filenames, press Return in response to the prompt and the program will move on the file title. Default extension for the input filename(s) is DAT.
CONCAT records all your responses (filenames, data types, file and condition numbers etc.) in a file with the same name as your output file except that the filename extension is CAT. When you enter the output filename, CONCAT will check to see whether a CAT file already exists for that filename. If it does exist, the program will ask you
READ DATA TYPES, FILES AND CONDITIONS FROM A CAT FILE? (N/Y):
If you respond Y (yes), all further input to CONCAT (input filenames, output file title, data types, file and condition numbers and options) will be read from the CAT file and no further prompts will be printed at the terminal. If you respond N (no) or just press Return, or if the CAT file does not already exist, a CAT file will automatically be created using the reponses you enter in this run. This CAT file will then be available for use in future runs of CONCAT - when you have run more subjects, for insatnce.
The program will ask you to enter a title (up to 50 characters in length) for the output file. The text you enter will be appended to the title of every data matrix in the output file. For instance, if you enter the file title
Post hoc analysis of short words only
and the the first two data types you enter are "SR" and "SE", then the first two data matrices in the output file will be preceded by the title lines
SUBJECT RT Post hoc analysis of short words only
and
SUBJ % ERR Post hoc analysis of short words only
This text will be printed at the top of your ANOVA summary table by PQRAOV.
You will now be prompted
ENTER DATA TYPE (OR "EX" TO EXIT)
This means that you must select the type of data to be copied from the DAT file(s) to the DAS file or exit the program. Possible responses to this prompt are:
SR - Write the title SUBJECT RT to the DAS file. The data to be copied into this matrix will be the mean RT for the conditions specified for each subject.
SE - Write the title SUBJ % ERR to the DAS file. The data to be copied into this matrix will be the total number of errors in the conditions specified for each subject.
IR - Write the title ITEM RT to the DAS file. The data to be copied into this matrix will be the mean RT for each item in each condition specified (averaged over subjects).
IE - Write the title ITEM % ERR to the DAS file. The data to be copied into this matrix will be the total number of errors on each item in the conditions specified (totalled over subjects).
<cr> - If you just enter a carriage return, no title will be written to the DAS file and the data type will be the same as for the previous matrix. This matrix will, in effect, be a continuation of the previous one.
EX - Closes the output file and exits from the program. If you exit by using control-C, some or all of your output may be lost.
These commands may be entered in upper or lower case letters.
Each column of the matrix represents a file-condition combination, e.g. condition 1 of input file number 1, condition 2 of file 1, condition 1 of file 2, etc. Input files are referred to by the number determined by the order in which they were entered. The maximum number of columns in a matrix is 50.
Having specified the data type for this matrix, you will now be asked to specify the file-condition combinations to be included. You will be prompted for FILE and CONDITION separately. When you have entered the last file-condition combination, press Return in response to the FILE prompt. CONCAT will then write the matrix you have specified to the output file. The matrix will also be listed at the terminal.
For each file-condition combination, you will also be prompted for an OPTION. If you do not want any options, just press Return and the program will proceed to the next file-condition. Possible responses to the OPTION prompt are:
<cr> No option. Proceed to next file-condition.
+ The data from this file-condition will not create a new column in the output matrix, but will be added (separately for each subject or item) to the data from the previous file-condition. If the previous file condition also had + option, then neither will produce a new column but both will be added to the second-last file-condition., and so on.
- As for + except that the data is subtracted from the previous file-condition.
C A constant will be added to every value for this file- condition. You will be prompted for the value of the constant; this prompt only appears when option C is selected.
If one matrix has already been constructed (e.g. Subject RT), you can repeat the same sequence of file-conditions in the next matrix (e.g. Subject Errors) without having to type it all in again. If you enter -1 in response to the FILE prompt, the entire previous sequence will be repeated. If you enter -n in response to the FILE prompt, the stored sequence from the nth file-condition onwards will be repeated.
Note that if you have already specified some file-condition combinations when you revert to the stored sequence, those file-condition combinations will have been stored. For instance, if you enter 4 file-conditions and then enter -3 in response to the FILE prompt, the third and fourth file-conditions entered will be output again as the fifth and sixth file-conditions of this matrix. The program will type out the actual sequence of file-conditions that is retrieved from the store. The options are stored along with the file-conditions, and will be applied again just as when the stored sequence was originally entered.
The analysis of variance program (PQRAOV) can only handle 3 factors. If your experiment has 4 factors and you want to use PQRAOV, you will have to collapse 1 factor. The way this is done depends on whether or not the factor to be collapsed is a repeated measure.
If the factor to be collapsed is non-repeated (e.g. "Files" in a SUBJECT RT analysis), you can append the set of observations for one level of the factor to the set of observations for the other level(s) of the factor. To do this, first create a matrix containing all the file-conditions for one level of the factor to be collapsed. Then, instead of entering a data type for the next matrix, just press Return, so that no title is written to the file and the matrix is a continuation of the previous one, and enter the file-conditions for the next level of the factor to be collapsed. In this way, data for all levels of the factor to be collapsed can be pooled together in the one matrix. When PQRAOV asks you the number of subjects or items per condition, don't forget to multiply this value by the number of levels of the collapsed factor.
If the factor to be collapsed is repeated, you can add together the observations for all levels of that factor to form a single observation. This is done by adding the appropriate file-conditions together using the + option (see 9.7 above). For error data, adding file-conditions together may produce observations greater than 100 (per cent). Don't worry about this, as it will not affect your analysis of variance.