SCIENCE SELF-EFFICACY INVENTORY (Middle School Students)
Description
This dataset contains the anonymised responses of 100 middle school students (Grades 6 to 8) to the Science Self-Efficacy Inventory, a 36-item instrument measuring science self-efficacy among middle school learners in Tripura, India. Data were collected at two schools using a paper-based, self-administered questionnaire. The inventory measures six dimensions, each represented by six items: D1 Academic Task Self-Efficacy (items 1 to 6), D2 Motivation and Persistence (7 to 12), D3 Emotional Regulation and Science Anxiety (13 to 18), D4 Social Support (19 to 24), D5 Science Identity and Aspiration (25 to 30), and D6 Contextual and Socio-cultural Relevance (31 to 36). Every item uses a five-point Likert format (1 = Strongly Disagree to 5 = Strongly Agree) and no item is reverse-scored. A dimension score is the sum or mean of its six items; the total score is the sum of all 36 items, ranging from 36 to 180, higher scores indicating higher science self-efficacy. The data file holds 100 rows and 46 columns: the 36 item responses plus participant code, anonymised school code, gender, grade, medium of instruction, school board, favourite subject, preferred language and career aspiration (as reported and recoded). The sample comprises 71 boys and 29 girls; 40 in Grade 6, 28 in Grade 7 and 32 in Grade 8; 93 studying through English medium and 7 through Bengali; 85 affiliated to the Tripura Board and 15 to CBSE. The dataset is complete apart from one missing value (item I18); blank cells denote non-response and should be declared missing, not zero. Item variables are documented as Likert rather than ordinal, following established practice in educational and psychological measurement, which treats summated Likert data as approximately continuous; means and other parametric statistics may therefore be computed on item and scale scores. Psychometric evidence is given in the accompanying report. Internal consistency is good (Cronbach's alpha = .891; McDonald's omega = .895), and no item deletion would raise alpha above .893. Item-total correlations range from .186 to .629, with 33 of 36 items meeting the .30 criterion. Each dimension correlates strongly with the total (.737 to .884), while inter-dimension correlations (.441 to .699) show the dimensions to be related but distinct. Comparing the upper and lower 27 per cent groups, 35 of the 36 items discriminate significantly. Participation was voluntary. Consent and assent were obtained, participants could skip any item or withdraw at any time without penalty, and were told that their anonymous responses would be deposited in a public repository for reuse. No direct or indirect identifiers are present: student and school names were permanently removed before deposit, participants identified only by a sequential code and schools by a non-reversible code. Four files are provided: the data sheet, the variable codebook, the blank response sheet, and the descriptive and psychometric analysis report.
Files
Steps to reproduce
1. Instrument. The Science Self-Efficacy Inventory contains 36 statements distributed equally across six dimensions (six items each): D1 Academic Task Self-Efficacy, D2 Motivation and Persistence, D3 Emotional Regulation and Science Anxiety, D4 Social Support, D5 Science Identity and Aspiration, D6 Contextual and Socio-cultural Relevance. The blank instrument is included as the response sheet file. 2. Administration. The inventory was administered on paper in intact classrooms at two schools to students of Grades 6 to 8. Participants first read the information and assent section, then completed Part A (background), Part B (interests and career aspiration) and Part C (the 36 statements). No names or school names were recorded on the form. 3. Scoring. Each statement is scored 1 = Strongly Disagree, 2 = Disagree, 3 = Neutral, 4 = Agree, 5 = Strongly Agree. No item is reverse-scored. Dimension score = sum of its six items (range 6 to 30). Total score = sum of all 36 items (range 36 to 180). 4. Data preparation. Responses were entered into the data sheet, one row per participant. Student names and school names were deleted; participants were assigned sequential codes and schools non-reversible codes. Career aspiration was retained verbatim (spelling standardised) and recoded into the inventory categories. Non-responses were left blank and declared as missing values, not zero. The dataset contains one missing value (item I18). 5. Software. All analyses were run in jamovi (Version 2.7), built on R (Version 4.5). 6. Descriptive analysis. Using Exploration > Descriptives, request N, missing, mean, standard error, median, mode, standard deviation, minimum, maximum, skewness, kurtosis and the Shapiro-Wilk test for the total score and the six dimension scores. 7. Reliability. Using Factor > Reliability Analysis, enter all 36 items and request Cronbach's alpha, McDonald's omega, and the alpha and omega values obtained if each item is dropped. 8. Item-total correlation. Using Regression > Correlation Matrix, correlate each item with the total score. Items are evaluated against the criterion of r = .30. Note that these coefficients are uncorrected, since each item forms part of the total. 9. Dimension-total correlation. Using Regression > Correlation Matrix, correlate the six dimension scores with one another and with the total score. 10. Discrimination power. Rank participants on the total score. Form an upper group from the highest scoring 27 per cent (n = 27) and a lower group from the lowest scoring 27 per cent (n = 27), and create a grouping variable. Using T-Tests > Independent Samples T-Test, compare the two groups on each of the 36 items, requesting group descriptives and Levene's test. Compute the discrimination index for each item as (mean of upper group minus mean of lower group) divided by 4. Full results and their interpretation are given in the analysis report file.
Institutions
- ICFAI University, TripuraUttar Pradesh, Mohanpur