Data Collection

The purpose of research is primarily to answer the research question through the collection, organisation and analysis of data. Before information can be presented and analysed, the data must be carefully collected, managed and organized. The research design and methodology will determine HOW the data is collected.

Data Organisation

One of the ‘invisible’ tasks of research is recording data and placing it in a useable form for analysis. The analysis is dependent on the type of data collected (quantitative, qualitative or clinical). Statistical data analysis organisation is normally in the form of a computer data file or spreadsheets for survey form tools or statistical software packages such as SPSS, SAS, STATA, or EPI Info etc. Qualitative data can also be organised using software such as NVivo or Atlas.ti.

Preparation for data analysis file:

One of the ‘invisible’ tasks of research is recording data and placing it in a useable form for statistical analysis. This is normally in the form of a computer data file or spreadsheets for survey form tools or statistical software packages such as SPSS, SAS, STATA, or EPI Info etc.

The following GENERAL RULES should be applied:

  • Organise information into rows and columns
  • Rows are individual cases (such as patient records or resident responses to a survey.
  • Columns are individual measures such as age, sex, immunization status etc. and reflect all the items you have to measure.
  • If more than one response per case is possible then multiple columns should be used – one for each possible response option.
  • Use one column to create a unique ID for each record. This ID should be encrypted in some way to maintain the confidentiality of the computer records.
  • If the same research subject has multiple files, the multiple files should have the same matching ID.
  • Code the responses – use numbers and not text wherever possible. When using text to code a response, words must be consistently spelled and formatted – even the case must match.
  • When deciding whether to use a number or a category, the rule of thumb is that if the measure is characterized by a continuous number (such as age, blood pressure) then the actual number, not the category should be recorded. The numbers can be converted to categories at a later stage.

Tips

  • Item name in Column headers: avoid special characters and spaces in the name.
  • Item description for labelling in the printout: avoid making this more than 20 characters.
  • Category labels (if applicable): 1 = Male, 2 = Female.
  • Missing value codes are best left blank.

BY YOLISWA NTSEPE (MA, PhD)
ADOLESCENT PROGRAMMES MANAGER

UPDATED NOV 22, 2023

TABLE OF CONTENTS:

REFERENCES & ADDITIONAL RESOURCES: