The following principles must be kept in mind when collecting quantitative data: |
- The quality of data is inter alia determined by the quality of data collection instruments. The following types of questions should be designed in specific ways to ensure that the data will be captured in a way that will allow the data to be analysed properly:
-
- Pre-coded
- Open-ended
- Multiple response questions
- Addressing multiple issues in one question
- Filter (skip) questions
- Some questions that are regularly used, such as income, highest educational qualification, require that these be formulated in a consistent way across studies. In many such cases standard coding frames such as those provided by StatsSA, must be used.
- Data collection instruments used by previously conducted HSRC surveys can be used as examples of best practice.
- Each record must have a unique identifier, i.e. a record number. This is especially important when a data set forms part of a bigger collection where unique identifiers are used to link records across data sets.
- In many cases, collecting biographical information about respondents that includes the spatial scales at which information should be collected, compiled, analyzed, is necessary.
- When data is omitted in a cell of a data matrix, it is referred to as a system missing value. These are not regarded as valid values in statistical analysis procedures. System missing values do not allow for meaning to be attributed to why the values are missing. Instead user-defined missing values are used in this regard. Read more about how these values are defined and are used.
- Numeric continuous variables must be captured in its raw form instead of per categories, if possible. These values can then be categorised at a later stage in various ways to aid analysis. Categorical variables must be captured as numeric codes (values) with descriptions (value labels) and not as string/ character variables.
The following principles must be kept in mind when collecting qualitative data: |
Audio/ video recordings
Principles to adhere to when transcribing interviews or focus group discussions include:
Noise refers to an unwanted signal that gets recorded to a medium. Noise can take different forms, the most common of which is background noise. Background noise is usually caused by recording in a noisy or untreated environment.
- Material used for recording
Audio cables: it is one of the most important parts of the audio recording process and an important part of sound recording basics.
Microphone: when sounds are too low, it does not get picked up by the microphone. If sounds are too loud it also doesn't get picked up, and the phenomenon called audio clipping occurs. You need to know what kind of microphones to use to record your audio otherwise you are not going to quality of recording you require.
Before personally identifiable data can be processed, the data subject needs to give informed consent. The respondent must know about the nature of the data being collected, the reasons for processing it and what will be done with it. Data subjects should not be misled, lied to or tricked. Participation is voluntary and they can withdraw at any time.
Transcribing
Principles to adhere to when transcribing interviews or focus group discussions include:
- Accuracy. Only type the words that are spoken in the audio file. Phrases or words you don't understand should not be omitted. Instead, they should be tagged appropriately according to the institution’s guidelines.
- Use proper English. Use proper English - capitalization, punctuation and spelling. Do not write phonetics or netspeak such as “u” for “you".
- Do not paraphrase. Do not correct the speaker's grammar nor rearrange words. Also, do not cut words that you think are off-topic or irrelevant. Any words not spoken should not be included. Type the actual words spoken.
- Do not add additional information. Do not add additional information such as page numbers, job numbers, titles, or your comments in your submission. Such information can be added in separate fields below the transcript.
- Verbatim work should be truly verbatim. When transcribing verbatim work, include every utterance and sound exactly as you hear it. Unless directed in the work's “Notes” section, all filler words should be included. Also, transcribe stutters as accurately as possible.
- Do not use ellipses (…) in transcribing oral history recordings because they give the impression that material has been left out.
- Nonverbal sounds or events should be recorded in parenthesis and italicized such as (laughs), (cries), or (knocks at the door).
- Brackets [ ] are to indicate notes and words not present on the recording and added to the transcript. Example: I never expected to find myself in a swamp in LA [Los Angeles].
For more examples and principles click here