Internal Consistency
Reliability for the Teachers Attitudes Toward Information Technology (TAT) Questionnaire
Technical Report 97.1
Texas Center for Educational
Technology
Telecommunications and Informatics Laboratory
Gerald Knezek, Matthews Chair for Research in Education
Rhonda Christensen, Research Associate
July 20, 1997
Overview
The Teachers Attitudes Toward Information Technology Questionnaire (TAT) was developed during the second phase of the 1995-97 Matthews Chair for Research in Education Project at the University of North Texas. During Phase I (1995-96) activities focused on developing a more parsimonious instrument to assess the areas covered by 14 previously validated computer attitude scales. The result of that effort was the Teachers Attitudes Toward Computers questionnaire (TAC) (Christensen & Knezek, 1996, 1997). During 1996-97 construction of the TAT was begun to address areas not covered by the TAC. These were primarily what have come to be know in Europe as the New Information Technologies (NIT) -- multimedia, electronic mail, and the World-Wide Web (WWW). In addition to the NITs, other areas, such as the use of information technology to improve teacher productivity, were also included, and two well-validated subscales from the TAC (semantic perception of computers and classroom learning via Email) were included for comparison purposes. Refinement of the TAT is ongoing. In this document internal consistency reliabilities are reported for the first large-scale pilot test of the instrument.
Subjects
The subjects in this study were 147 teachers from six schools in the a large urban public school district in northern Texas. TATs were completed during May of 1997. This was the last month of the 1996-97 school year in Texas.
Data Acquisition, Preparation, and Analysis
Data were gathered for 10 separate indices from the teachers responding. Nine of these ten subscales were composed of semantic differential items (see Appendix). Semantic items were hand coded with a number from 1-7, representing the particular space the respondent marked between the adjective pairs, then keypunched by the University of North Texas Computation Center data entry staff. Cronbach's Alpha was produced for each subscale using SPSS on a Macintosh computer.
Internal Consistency Reliabilities
As shown in Table 1, internal consistency reliabilities for the ten TAT subscales ranged from a low of .91 to a high of .98 (These reliabilities are based on responses from 74 subjects who furnished complete data). According to the guidelines provided by DeVellis (1991), this is excellent reliability for a research instrument. Additional strength is added to this conclusion because the two subscales carried over from the TAC exhibited internal consistency reliabilities quite comparable to the .93 (Kay's Semantic) and .95 (D'Souza's Classroom Email) found for these indices in the previous study of TAC reliabilities (Christensen & Knezek, 1997). The earlier study involved a much larger and diverse sample.
Scale # cases
# items
Alpha
Kay's Semantic (CAM) 74
10
.91
Email (teacher) 74
10
.93
Email (student) 74
10
.95
WWW (teacher) 74
10
.95
WWW (student) 74
10
.96
Multimedia (teacher) 74
10
.96
Multimedia (student) 74
10
.98
Productivity (teacher) 74
10
.96
Productivity (student) 74
10
.96
D'Souza's Email 74
11
.95
Table 1. TAT Reliabilities for 1996 public school teacher data
Future Work on Validity
Content validity for the TAT is believed to be quite high due to the way the instrument was constructed. Subscales were selected precisely because various scholars and practitioners in the field had identified these areas as important but not measured by previously-existing questionnaires. Construct (factor) validity needs to be checked through confirmatory factor analysis to determine if some of the ten current subscales are not in fact duplicates of another. Criterion-related validity, such as the ability of the instrument to separate (discriminate between) groups thought to be different on one or more attributes, or the ability to measure pre-post change in a single group, needs to be established as well. Future research is planned in these areas.
References
Christensen, R. & Knezek, G. (1996). Constructing the Teachers' Attitudes toward Computers (TAC) Questionnaire. Paper presented to the Southwest Educational Research Association Annual Conference, New Orleans, Louisiana.
Christensen, R. & Knezek, G. (1997). Internal consistency reliabilities for 14 computer attitude scales. in Willis, J. et al. (Eds.) Technology in Teacher Education Annual, 1997, 877-880.
DeVellis, R. F. (1991). Scale Development. Newbury Park, NJ: Sage Publications.