2.9

# Data analysis

To gain insight from the data you have collected, you need to carry out a detailed analysis. This important step should be closely aligned with your research design and should be appropriately planned at the outset of your research.

## Quantitative data analysis

Quantitative data analysis is a systematic method through which numerical data are collected, or observed data are transformed into numerical data and then processed, in order to answer specific research questions or solve problems. It helps unearth the evidence to either debunk or buttress a hypothesis (the hypothesis provides a plausible answer to a research question).

Quantitative data analysis utilises statistics. To analyse quantitative data, you will have to assign codes to your respondents’ answers. Through this process you will identify the valid responses and eliminate errors.

Analysis can be conducted in a descriptive manner (exploratory analysis) or an inferential manner (confirmatory data analysis). Descriptive analysis investigates, describes and summarises variables in a given data set. It includes mean, mode and median calculations, referred to as measures of central tendency. Inferential analysis draws conclusions about the larger population based on the results obtained from a sample drawn from larger population.

## Qualitative data analysis

There is no single accepted procedure for analysing qualitative data. However, there are some processes that have been tested to be systematically effective.

Analysing qualitative data, such as that from an interview, will involve a number of key stages. For example:

1. Create transcripts of the interviews, including the pauses and tone of voice
2. Sift and sort transcripts by rereading the transcript slowly and carefully, making comments in margins (informal codes)
3. Search for emergent categories, themes and sub-themes by coding the text (formal codes), allow these codes to be continually refined
4. Mark up (eg, colour) or cut up your transcripts according to these categories and themes
5. Build a theory, the entire process should follow accepted procedures (eg, ‘semiotics’ or ‘discourse analysis’) and relate to your own interpretation of the interview transcript as text, for example:

• The literal (ie, focus on what is said)
• The interpretative (ie, focus on meanings, power relations etc that extend beyond the interview situation)
• The reflexive (ie, focus on relationship between researcher and researched)

## Data analysis software

The most common software packages for data analysis in construction research include Excel, Statistical Package for Social Science (SPSS) and NVIVO.