2. Collecting and using data

The types of data collected for the research are primarily determined by the research questions. The data can consist of, for example, images, interviews, statistical data, interview recordings or numerical measurements. Furthermore, the data types influence how the data can be produced and processed. For example, interview data involve the processing of personal data, so it is crucial to be especially careful about the devices and software used for collecting and handling the data. In turn, the size of the data generated by measuring devices may place special demands on storage space and sharing solutions. It is therefore necessary to consider in advance which file formats or programs are suitable for which research data and research setting. The choices are also influenced by discipline-specific practices. 

In the context of data management, the quality of data is first and foremost related to the data not being accidentally altered, damaged, or destroyed. The researcher needs to identify any processing operations or situations where the data content could be inadvertently altered. These include, for example, conversions from one file format to another and interview transcription. Ensuring data quality is therefore an exercise in risk management. 

Reflecting on the quality of data collection is intrinsically linked to research methods, but also partly to data management. Quality management measures in the data collection phase include the calibration of instruments, taking multiple measurements and using standardised methods and protocols. Accurate documentation of data collection is also an indicator of quality. 

Research data can be classified according to whether the data are specifically collected for the research (e.g. samples, surveys, measurements, imaging), pre-existing data collected from original sources and reused (e.g. text corpora, image banks, biobanks, registers), or produced in the research (e.g. data produced through the analysis of collected data, software). 

The origin of the data holds significance for what can be done with the data and who decides on the matter. If the researcher collects the data from the outset, they themselves can define the conditions of use within the limits of the law and the contracts concluded. In case the researcher makes use of pre-existing data, the conditions of use must be considered.  The terms of use for open access data are usually defined by a licence (e.g. CC licence). The terms of use are further discussed in chapter 5. 

When collecting the material, existing data can also be reused. Data reuse means the use of existing data for new purposes such as research, teaching, studying, or commercial purposes. The existing data can be searched from data repositories and search services (see, e.g. the sections Research data servicesand Reuse of open research data in the UEF Open science website). 

Reusing data helps in saving time and money because you don’t have to do everything related to data collection all the way from the beginning. This option is worth considering if there is useful and applicable data available. 

If you are planning to reuse pre-existing data, this research data must be cited just like books and scientific articles. The data reference should consist of following elements: 

• creator 
• title 
• host organization 
• publication time and/or date 
• persistent identifier.
  

Useful additional elements are: 

• version 
• resource type 
• license 
• ORCID (persistent identifier for researchers) 
• possible embargo information 
• Repository.
  

Data repositories and archives usually have guidelines for data citation. In addition, publishers can have their own guidelines on how to refer to data in journals. 

Although commonly associated only with research involving personal data, research ethics is relevant to all scientific research. Of course, if the research material contains personal data or information, e.g. on endangered animals, national defence or trade secrets (sensitive/confidential information), research ethics and legislation must be borne in mind from the very beginning of the research and its design. 

It is important that the researcher identifies any ethical, contractual, and juridical aspects and restrictions related to the research data and addresses them throughout the data management process. It is primarily the responsibility of each researcher to implement good scientific practice in their own work, but they are not alone. The responsibility is carried by the entire scientific community, including research teams and their responsible researchers, as well as research organisations and their management.  

In the context of research ethics, we often talk about research integrity. It is based on the principles of reliability, integrity, respect, and responsibility. Research integrity guides our actions at the different stages of research. In Finland, the Finnish National Board on Research Integrity (TENK) provides guidance on matters concerning research ethics in its guidelines for research integrity, which state the following about good research practices in data management: All partners 

• agree in advance about the ownership of the research data and about the rights to its use, its processing, storage and possible reuse 
• if necessary, revisit the agreements later during the course of the research, for potential amendments 
• comply with current data protection legislation and obligations related to non-disclosure, confidentiality and secrecy 
• promote to openness and further use of the data to the extent possible. 

Personal data can be used for scientific research when the use is appropriate, planned and justified, and when there is a legal basis for processing the personal data. When collecting data on individuals for research purposes, all ethical and legal guidelines for the research in question must be followed. This must be borne in mind throughout the life cycle of the research, from design to implementation and post-research storage. As a general rule, application of the General Data Protection Regulation (GDPR), should always be considered, unless you are absolutely sure that there is no human-related information in the research data. 

Personal data is any information relating to an identified or identifiable natural person: 

• direct identifiers that alone are enough to identify a person (e.g. such as full name, social security number, email address containing the personal name, biometric identifiers such as facial image, voice patterns, fingerprints, iris, handwritten signature) 
• strong indirect identifiers with which a person can be identified with reasonable effort (e.g. postal address, phone number, IP-address, student number, insurance number, bank account number, exact annual income, vehicle registration number, unusual job title, rare disease) 
• indirect identifiers, i.e. information which on their own are not enough to identify someone but when linked with other available information, could be used for deducing the identity of the person (e.g. sex, age, principal abode, profession, workplace, education, school, dates such as date of birth, death, accident).
  

When processing personal data, it is important to identify information belonging to certain special category of personal data, known in everyday language as sensitive personal data. These include e.g. religious beliefs, health information, and trade union membership. It is also important to remember the difference between pseudonymous and anonymous research data, which are discussed in the next section (Processing personal data). 

In accordance with research ethical principles, the research subjects should be informed about what kind of data will be collected, how it is handled during the research and what will happen to the data after the research. It is therefore essential to plan the management of data during and after the research right from the outset. Research involving human subjects may also require prior ethical review. As a general rule, personal data should be minimized when collecting research data. This means that only such personal data necessary for the research purpose should be collected. 

Research organisations will instruct on the practices for handling personal data as required by the GDPR. For example, UEF’s Heimo services provide guidelines and template forms for the processing of personal data in scientific research (requires UEF credentials). The templates will help you in preparing the necessary documents. These might include: 

• privacy statement  
• participant information 
• ethical consent to participate 
• consent to process personal data 
• an assessment of data protection risks or data protection impact assessment (check your own organisation’s instructions, processes, and templates)
  

In data management, special attention must be paid to secure storing, processing, and transferring of data, so that only those who have justified reasons to access the data can access it. It is good to plan carefully how the research participants are informed about data processing and data management, for example, how date security is taken care of, how long the data is preserved, where will the data be shared for reuse and how.  

For research purposes data can be processed with identifiers, or as pseudonymous. Pseudonymization means removing or replacing identifiers with pseudonyms or codes, which are kept separately. Pseudonymized data is still personal data. 

Anonymization means removing all identifiers permanently. Several techniques and tools can be used in anonymisation.  Anonymisation is necessary if the data is shared openly or stored for reuse purposes. It should be remembered that anonymisation as a procedure involves the processing of personal data, but the anonymised research data no longer contain personal data. At this point, it can be treated as any other data that does not contain personal information. However, it is important to remember that it may not be possible to produce completely anonymised data. Read more about personal data and anonymization from FSD guidelines. 

For further information on processing personal data, see, e.g. Data protection (UEF) in Heimo services (requires UEF login). 

Laine, Heidi (ed.) 2018. Tracing data: Data citation roadmap for Finland Finnish Committee for Research Data (FCRD).

Finnish advisory board on research integrity (TENK)

UEF Research Ethics

The Office of the Data Protection Ombudsman