Learner questionnaires

3.8.1.1 Learners’ Understanding of Science and Scientific Inquiry

For the LUSSI, scoring closed learner responses was done as follows: strongly agree = 1,

agree = 2, not decided = 3, disagree = 4 and strongly disagree = 5 (Liang, et al., 2006).

Items representing naive conception of science (negative Likert items) were scored in reverse. These items are: 1B, IC, 2D, 3B, 3C, 4A, 4D, 5C, 5D, 6B, and 6C (Appendix A). A low total (minimum score = 24) score meant agreement with naive views of science and scientific knowledge and a high total score (maximum score = 120) meant views that were informed. Liang et al. (2006; 2008) and Miller et al. (2010) have used this scoring technique. The results reported in the current study do not separate certainty of scientific knowledge and the nature of scientific inquiry (which includes nature of scientific enterprise and scientists according to the definition already given). Computation of scores was only done along the lines of naïve and informed views.

Learner open-ended responses were scored using the SUSSI rubric provided by Liang et al. (2009) (personal communication, see Appendix A2), categorizing responses as informed (3), transitional (2), naïve (1), or not classifiable (0) as developed by Liang et al. (2009) (See Appendix A1) with the scores in brackets. The LUSSI open-ended scoring rubric is based on the multidimensional framework. The six (6) dimensions were then conceptualized as a continuum ranging from positivist/empiricist to constructivist/relativist perspectives (see, eg., Bell, 2006; Crowther, et al., 2005; Lederman & Lederman, 2005; Lederman, 2007a; McComas, 1996). Learner responses were classified as not classifiable, naive views, transitional views or naive views. This involved continuous readings of

141

responses and meaning-making so as to determine the relevant category. Positivist/empiricist views were labelled as naive or inadequate views, whereas the constructivist/relativist views were labelled as informed or adequate. This was done with the understanding that being an empiricist does not mean one is naïve as is depicted in most NOS and NOSI literature.

The researcher (author) and a post-doctoral research fellow registered in the same centre and university with the researcher first analyzed the data from LUSSI open-ended responses separately following Liang et al.’s (2009) SUSSI open ended scoring rubric. Discussions were then held concerning the categorizations and classification of the NOSI views. This was done for validating the coding purposes and checking inter-rater reliability. Once the categorization was agreed upon, the same codes were utilized in probes’ questionnaire and interview analysis. Atlas.ti version 6.2 (a computer software program, mostly used in qualitative data analysis) was utilized for analysis. Use of the same codes (the meaning applied to a quotation for analysis) in Atlas.ti version 6.2 enabled labelling of phrases to be consistent and frequencies determined. All quotations belonging to each code were given as a print out. Tables and figures were then generated from the codes which enabled learners’ responses to be classified into different categories.

3.8.1.2 Probes

To analyze data from probes’ open-ended responses, the researcher read through sets of transcripts making preliminary notes regarding patterns that emerged from individual participants. This followed a model used by Ibrahim, et al. (2009) and Campbell, Lubben, Buffler and Allie (2005) since the structure of the probes used were similar. However, the coding was not exactly the same as done by Campbell, Lubben, Buffler and Allie (2005) in their monograph on teaching scientific measurement at the University of Cape Town. For this study, a hybrid model was produced after fusing the Ibrahim, Buffler and Lubben’s (2009) coding model with that of Liang et al.’s (2009) rubric for scoring SUSSI open-ended responses. This was done purposefully and intentionally because probes were used to corroborate LUSSI open-ended responses. Thus, some aspects of grounded theory

142

analytical procedures (Strauss & Corbin, 1998) especially interpretive analysis (see, e.g. Gall, et al., 2003) were used to inductively analyze the participants’ probes’ open-ended responses. The data collected from the sets of open-ended probes’ responses were coded using the hybrid model. These procedures involved (1) the simultaneous collection and analysis of probes data and (2) comparative methods of analysis whereby participants’ responses were compared among one another and within each participant, and (3) the integration of a theoretical framework that guided the study. The multidimensional

framework was used to categorize NOSI conceptions of the participants when all responses

from the other instruments were considered for each participant.

For the hybrid model, dimensions from the LUSSI instrument were used to develop the coding patterns. The transcribed probes data were read looking for patterns, relationships and other themes within the dimensions. Entries were coded according to patterning identified while keeping a record of what entries went with which element of the patterns. In other words, the data was read and then chunked based on common language. The frequencies of responses for each probe were tallied and clusters of responses showing similar types of reasoning were identified, for example, LT0, LT1, LT2, LT3, LT4, LT5 and so forth for the probe category responses on laws and theories in science. The accompanying ascending numbers represent a continuum of positivist/empiricist views labelled as inadequate to constructivist/relativist views labelled as adequate. Underlying reasoning was then identified for each learner by writing the category codes for each probe response and this enabled understandings of the nature of scientific inquiry to be determined for each learner’s response. Examples of illustrative codes for the probe on ways scientists validate new knowledge from the Probes questionnaire include: WSVNK0 for no response; WSVNK1 for not able to code reason given; WSVNK2 for scientists do not require accurate record keeping, peer review and replicability; WSVNK3 for scientists require accurate record keeping, peer review and replicability with no further elaboration or examples given; WSVNK4 for scientists review and ask questions about the results of others so as to validate discoveries with reasoning; and WSVNK5 for communication and peer review impact what and how science progresses.

143 3.8.1.3 Learner Perception for Classroom Inquiry

The response alternatives on the five-point bipolar Likert scale ranging from (1) never occurred (2) seldom, (3) sometimes, (4) often to (5) almost always were allocated scores from 1, 2, 3, 4 to 5, respectively. As was the case with the analysis of lesson observations, scoring was done in reverse for statements representing non-inquiry or closed-inquiry laboratory. Reverse scoring was done for the following items: A1, A2, A3, A4, B2, B3, B4, C1, C3, C4, D1, D2, D3, D4, E1, E2, E3 and E4. Total scores were obtained for all the 20 items in the instrument and for each of the five sections. Open-ended inquiry is represented by high scores (maximum = 100) and laboratory work which is verificationistic, expository, or closed inquiry is reflected by low scores (minimum = 20). The rankings given to the statements by each class were used to categorize the type of laboratory they experienced into verificationistic, structured, guided inquiry or open-ended inquiry. As an example, learners exposed to more open-ended laboratory work were expected to rank statement C3 ‘‘I actively participate in investigations as they are conducted’’ very highly, and those in low-inquiry laboratories to rank statement B1 ‘‘I am given step-by- step instructions as I conduct investigations’’ very highly.