Analysing data in research based on the constructivism paradigm using ATLAS.ti

June 21, 2020

Written by: Aanuoluwapo Opayinka

There are different paradigms in research and the paradigm on which a research is built would determine the approach and design to be used for the research. The two most popular paradigms are the postpositivism and the constructivism paradigms. My Ph.D. research was built on the constructivism paradigm which according to Creswell (2013) uses the qualitative approach. In the constructivism worldview, individuals develop the interpretation of their experiences subjectively. And since individuals differ, the interpretations of their experiences differ and are in multiples. As a researcher, the need therefore arises to understand the complexity of the participants’ interpretations of their experiences. The qualitative approach was therefore appropriate for conducting the research and ATLAS.ti came in very handy for analysing the vast array of data. My research was on stakeholders’ participation in indigenous vegetable innovation platforms. My respondents were the stakeholders participating in the innovation platforms (IP). The data were therefore collected using focused discussions during IP meetings in different locations. Voice recordings were collected and pictures were taken. The voice recordings were transcribed into text documents.

When it comes to coding, I prefer to use what I call ‘2-step manual coding’. At the first step, the documents are coded manually one after the other without being too critical on code names or number of quotations in each code. This coding gives a general idea of the data collected as a whole. The first set of codes may therefore be very broad and contain many quotations (as shown in Figure 1) or too narrow.

Figure 1: Broad codes from the first coding process

The second coding is done in the code manager by reviewing each code, splitting the broad codes into more specific codes, and merging the very narrow codes as necessary. In the code manager, the codes can be renamed and similar codes can be merged and broad codes can be split. To split codes, the code to be split is highlighted and ‘split code’ is clicked on. In the split code dialogue box, as shown in Figure 2, ‘New Codes’ is clicked on to create new codes.

Figure 2: First split code dialogue box

In creating the new codes in the ‘Create Codes’ dialogue box, the old code name could be retained to serve as a prefix for the new code name as shown in Figure 3. In the case of benefits derived by stakeholders from the innovation platform, retaining the old code as a prefix helps in differentiating benefits derived by farmers (F.benefit) from the benefits derived by other stakeholders, such as marketers (M.benefit).

Figure 3: Creating codes from code splitting

Retaining the old code name as a prefix also allows the codes to be listed consecutively in the code manager, as shown in Figure 4, rather than being scattered in the code listing. Thus, keeping the old code name helps to manage the new codes better, especially if there is a possibility of repeating code names. When ‘create’ is clicked on, the new codes are created and shown in the ‘split code’ dialogue box.

Figure 4: Consecutive code listing

In the ‘split code’ dialogue box, as shown in Figure 5, the quotations are assigned to one or more appropriate and specific code(s). Quotations can still be retained under the old code name as desired since it is still a code. The old code can contain quotations that it best describes or retain all its quotations though they are being split. But when the mutually exclusive option is checked, each quotation can only belong to one code. After assigning the quotations to the appropriate code(s), click on ‘split code’ to complete the splitting.

Figure 5: Second split code dialogue box

Preparing a good network could be quite taxing and require creativity. In a network view, one could show the relationship that exists between different entities in the project. To be able to establish the relationship among entities, code-code links and hyperlinks are used. New code-code links and hyperlinks can also be created, and existing ones created by ATLAS.ti can be edited in the relation manager as shown in Figure 6. This makes both code-code links and hyperlinks malleable, thereby giving users the ability to be creative in showing relationships among entities using links with the properties that portray the desired relationship.

Figure 6: Creating and editing links

It is important to keep networks from being unnecessarily busy and easy to understand in the bid to create a meaningful network. It is not compulsory to use all the quotations (especially if there are repetitions) if the main aim is to make known the unique findings and not frequencies. Also, at times, in achieving an objective, only one network may not be sufficient or it may be too busy and the font size would be too tiny and illegible. To be able to express clear results, a network can be used to highlight the main points as shown in Figure 7 while other networks could be used to explain one or more of the highlighted points per network. The results in Figure 8 show a breakdown of one of the points in Figure 7 (i.e., training).

Figure 7: Highlights of benefits that farmers derived from the innovation platform

In using the links, the direction of the arrowhead is important and should be meaningful. In Figure 8, it can be seen that ‘preservation and processing’, ‘natural insect pest control’ and ‘vegetable and seed production’ were part of ‘training as benefit,’ which the purple arrows pointed to. Also, ‘the truth is that people don’t consume dried vegetable in this area’ explains why ‘…the technology has not been tried out’, hence the arrow pointed to the quotation which is being explained. The same thing applies to the link ‘expands’ which points to the quotation which ‘in this area, this does not make much sense. Consumers have not exhausted the supply of fresh vegetables not to talk of dried vegetables’ expands. Code colours could also be helpful. In Figure 8, the codes that are part of ‘training as benefit’ are of the same colour showing that they are at the same level in the network.

To ensure that texts in networks are legible enough, the networks created should be as concise as possible in the network view. To do this, the links could be as short as possible. When this is done, the font size becomes larger and more legible when the network is fitted to different windows. Also, when the network is imported into the body of the dissertation (e.g., in Microsoft Word), placing it on one full page with landscape orientation makes it even more legible as seen in Figure 8.

Figure 8: A breakdown of one of the benefits (Training) as experienced and expressed by the farmers

It is important to back-up the ATLAS.ti project library. This is very important because a computer can crash without prior notice and the need may arise to reinstall the operating system and reinstall ATLAS.ti. Since the default location of the ATLAS.ti 8 library is not easy to find by a layman, if there is no library backup or project bundles, there might be the need to request technical assistance from the ATLAS.ti technical team in order to be able to retrieve the ATLAS.ti library. It might even be impossible to retrieve the library if the hard disk of the computer has crashed. To be able to back up the ATLAS.ti 8 library in Windows 7, click on the start icon, select all programmes, find the scientific software folder, click on it, then click on ‘backup and restore ATLAS.ti library. Choose ‘backup ATLAS.ti Project Library’ in the ‘ATLAS.ti Backup and Restore’ dialogue box as shown in Figure 9.

Figure 9: ATLAS.ti Backup and Restore dialogue box

As seen in Figure 10, there is an opportunity to choose where to save the backup file. One could even save it in an external drive, which is more advisable.

Figure 10: Dialogue box to choose the desired location for the backup file

In conclusion, ATLAS.ti is a very good tool for analysing data collected during research based on the constructivism worldview. Despite the complexity and diversity in individuals’ experiences and their interpretation of the experiences, ATLAS.ti helps researchers properly analyse everything and gives a very good and organised result that rightly depicts the experiences.



Opayinka, A. (2020). Certified Trainer Case Study: Analysing data in research based on the constructivism paradigm using ATLAS.ti. Retrieved from


About the author:

Aanuoluwapo Opayinka is a lecturer in the field of Agricultural Extension and Rural Development (Obafemi Awolowo University, Ile-Ife, Nigeria). Her interests are in innovation platforms, evaluation of research and development programmes, and gender. She is a Certified ATLAS.ti Professional Junior Trainer.


Share this Article