3.2: Qualitative Research
3.2.1 Research Validity
As with any research into education the issue of research validity must be discussed. Schutt (2006) states "We have reached the goal of validity when our statements or conclusions about this [...] reality are correct" (Schutt, 2006, p.19). Therefore research can be considered valid when the statements made about a research subject are determined as correct. This however, raises an issue with social science research whereby data can be interpreted differently by different researchers. Arguing for the credibility for this research methodology, Bryman (2008) describes research as being valid through repetition. If the data is analysed by the same method, then the results should be repeated (ibid). While this could be easily proven with quantitative data, the concept can still relate to qualitative data. If this research is to be determined as valid then the interpretations of the qualitative data must be determined to be empirically correct and this can be done by achieving the same results if studied using the same tool. In order to identify whether research is valid or not, Schutt (2006) and Bryman (2008) provide four characteristics of research validity: Measurement Validity, Generalizability (external Validity), Causal Validity (Internal Validity) and Authenticity (Ecological validity). Measurement validity is concerned with the idea that the measurement tool measures what it has been designed to measure. Generalizability relates to the extent a sample can be determined to be representative of the general population group. Causal validity refers to the truthfulness of statements and authenticity relates to statements being representative of the data collected. Silverman (2006) relates “Credibility” to the validity of research and the reliability of the data collected.
It has previously been argued that science education is currently facing difficulties (Appleton and Kindt, 2002, D.o.E., 2006, Harlen et al. 1995, Murphy et al., 2007, Rocard et al., 2007, Varley et al,. 2008a, 2008b, 2013, Waldron et al. 2009). While the requirement for reform needs to be addressed the relevant educational change factors must be considered for any reform to be successful. Relating to this research, the validity emerges from the analysis of the major factors influencing educational reform. This analysis is developed through the extensive contextual study which partners literature analysis, observations of practice, observations of initial teacher education and the analysis of teaching resources to develop an in-depth and current view of science education in Ireland. Ensuring the reliability of this research is the systematic research methods that are described throughout this thesis. The methodological considerations for each of the methods will be described from Section 3.4 onwards. As the research is developed through so many individual methods, it was important
that each study was developed with its own considerations. This would ensure that any data collected could be developed further combining it in the concatenation process. This therefore verifies the credibility of the research as throughout the research methodology the ideas of validity and reliability are considered.
The development of a Grounded Theory may create issues when considering the validity of the research, as the categorizations that are developed must be scrutinized to ensure that they are appropriate for the research. This is of particular concern to this research whereby the primary data collection is developed through observations. Flick (2002) highlights how grounded theory gives preference to the data collected over the theoretical assumptions in the research. Flick quantifies this statement by highlighting that if a researcher abandons the theoretical assumptions then the research may reduce its relevance in that field (ibid). If research concentrates on the data alone, it also lends itself to criticisms over the data analysis method. Flick (2002) highlights the following issue with relation to data collection:
Decisions on data to be integrated and methods to be used for this are based on the state of the developing theory after analysing the data already at hand at that moment.
(Flick, 2002, p.42)
This means that data analysis methods can only be scrutinised for validity after the data has been collected, an issue which cannot be overcome. Therefore it is up to the researcher to ensure that the analysis method is continually examined for validity rather than the data collection method. Stake (2006) also argues that a ground theory can only be developed through a cyclical analysis of the categorizations that are being considered in the research. By continually analysing the categorizations the research can ensure that the research is applicable to the field and that it maintains its validity. Returning to the approach carried out in this research, it becomes apparent that the research method must consider theories and hypotheses of inquiry in primary science education, do not allow these hypotheses and theories to influence data collection and finally to continually re-examine the categorizations of data collection to maintain their relevance and validity to the field.
As it will be shown in the methodology presented, the primary tool used for the categorization of this research is the 5Es analytical framework which will be discussed in Chapter 5. This framework was developed from the 5Es model, a model which is consistent
with the structure and process of Inquiry in education, as shown in Section 2.2. While it has been shown that the 5E model is an accepted model for the development of inquiry lessons, it was not assumed to be valid for the purposes of this research. In order to ensure its validity, this model was analysed with relation to the needs and processes involved in this research. Indeed, it will be described in the textbook analysis, Section 4.3, that the analytical framework was developed using the 5Es model concurrently to the textbook analysis. It became evident when analysing the textbooks that the definitions used in the 5E model, see Table 2, may lead to ambiguous definitions of the purpose of each activity in textbooks. Therefore the 5E model was re-examined to define its purpose and structure in the form of an analytical framework. This framework was then used to categorize a coding structure that was used to analyse the textbooks. Again it was found that there were ambiguities in the structure and the analytical framework was redefined. This method of cyclically applying the framework and coding to the textbooks refined the framework into a set of specifically determined codes that could then be applied throughout this research. It must be noted that while the coding structure was applicable to the textbooks, it was again re-examined with the analysis of the observations. This lead to the framework and coding structures being further refined, ensuring that they remained valid to the data collection in this research.
When carrying out research quantitative data can reduce many of the issues of research validity. For example causal relations are easier to depict with quantitative data than the interpretation of a researcher's qualitative data. Cohen et al. (2000) argue that a mixed method approach is more suited to studies of human behaviour. Indeed if mixed methods are utilised in research and they identify similar results then the validity of the findings can be confirmed (Silverman, 2006). Therefore this research comprised a mix of qualitative and quantitative methods. Using multiple methods to carry out research is known as triangulation (Schutt, 2006).