Data Collection (Social)¶
The method used to collect data will depend on; the theory, research design, research question and what you’re tying to measure. The collection techniques can be:
- Secondary analysis (existing statistics)
- Focus group
- Secondary analysis
Focus on causal relations and testing hypotheses. They are based on the principle of simulating and manipulating conditions and are characterised by a few key components:
- Independent and dependent variables
- Pre and post-testing
- Experimental and control groups
Is considered to be “field research”, and can be; quantitative or qualitative, obtrusive or unobtrusive, participant or non-participant.
It involves significant raining and skill to undertake and often os used in combination with other data collection techniques. Due to the nature of the presence of the observer, there may be an observer bias.
This types of research takes data that has already been collected, such as from a government dataset and applies statistics to re-analyse them. As a result ethics has already been dealt with, and there’s no labour involved in collecting the data, though a certain level of skepticism needs to be employed by the researcher, as they do not know the specifics in which the data was collected.
Quite often this methodology looks for answers to questions that were not answered in the original study.
Are a way of getting information directly from a participant. They can utilise questionnaires, interviews and focus groups
Can be delivered by multiple different means, each with their own advantages and disadvanteges:
|Face to Face||- Highest response rate
- Can have much longer questionnaires
- Can observe non-verbal communication
|- Highest cost
- Can have an interviewer bias (sense of being judged)
- Can be difficult to get answers to difficult questions
|Telephone||- Large sampling frame (95% of the population)
- High response rates
|- High cost
- Difficult to use open ended questions
|- Cheapest method
- Can be anonymous
- Limits the amount of interviewer bias
|- Has a low response rate due to the effort required
- There is no guarantee that the respondent will answer themselves
|Web||- Fast and inexpensive
- Flexible survey design
- Convenient for the respondent
|- Coverage is limited to those with internet/computer literacy
- Data privacy can be a complicated issue
- There is no guarantee that the respondent will answer themselves
Types of questions¶
- Closed ended questions - respondents must choose from a set of fixed answers
- Partially open questions - an “other” option is provided with a short field for entry
- Open-ended questions - a text box is provided so that the respondent can answer in their own words
Principles of good question writing¶
- Avoid jargon, slang and abbreviations
- Avoid ambiguity, confusion and vagueness (e.g. “doctor” - is a dentist a doctor? “last month” - calendar month? 30 days? )
- Avoid emotional language and prestige bias
- Avoid double barrelled questions (asks multiple questions with only one answer)
- Avoid leading questions (questions that prompt an answer)
- Avoid asking questions that are beyond the respondent’s capabilities (e.g. recording information from too long ago)
- Avoid false premises (questions based on a misunderstanding or incorrect information)
- Avoid asking about future intentions
- Avoid double negatives
- Avoid overlapping or unbalanced response categories (e.g. “what’s your age?” a) 10-20 b) 20-30 c) 30-40)
Are an in-depth “conversation” type relationship between two people. They typically involve two people coming together to try and create meaning about a particular topic. The participant is the “expert” and the interviewer is the “student”.
Before an interview, the interviewer needs to answer a few questions:
- What kind of relationship do they want to develop with the respondent?
- How intimate should this relationship be?
- How will they gain access to potential interviewees?
- What kinds of information are they willing to disclose about themselves?
- Is the interview process a method for gaining information from the participant, or is it more collaborative in trying to find a common meaning?
An interview guide can help to focus the interview by listing the main topics and typically the wording of questions. It might also include some ideas of follow up questions and prompts to be able to extract more information, though it is not a script.
The questions that an interviewer might ask could involve questions about their:
- Experience or behaviours
- Opinions and values
- Factual knowledge
- Sensory experiences
- Personal background
The interview should be structured so as to put “softer” questions first and to make sure that questions don’t overlap and become redundant.
Just like in survey questions, there are some rules for asking interview questions:
- Avoid dichotomies - you want to keep questions open
- Avoid leading questions - you want the participant to answer of their own accord
- Ask both general and specific questions (broad and targeted)
- Try not to ask “why”, as it can lead to defensive answers
Before the interview, it’s important to prepare including to:
- Test the interview on someone else (pre-test)
- Establish a location (social settings can be more favourable or could be too open)
- Construct a “facte sheet”
- Decide what to bring
- Decide what to wear (do you want to be professional, or casual?)
- Make sure to warm up before the interview itself
During the interview, it’s important to keep on track, think on your feet, observe non-verbal communication and to record the interview
These are the most formal type of interview and are rigidly controlled. They’re typically involved in telephone interviews, market research and political polling, since they always follow a strict script. The questions must be asked exactly as written and the questions cannot be rephrased if the respondent misunderstands. The sequence is therefore preset
Allow for the participants to express their ideas and opinions in their own words whith the interviewer moving beyond the question to understand the participant’s point of view. While it begins with a basic structure, the interviewer can re-structure the interview based on the participant’s responses. These take a lot more skill to carry out, though they can be useful for being able to construct a theory.
These are typically conducted in the field, accompanied by an observational study and don’t really have a pre-defined structure. The responses are more organic and natural and the experience is more conversational
Are essentially group interviews of less than 10 participants, and can be structured or unstructured, though they are interested in the interactions between the participants. They incorporate all the same skills and methodologies of interviews, with the researcher acting as a facilitator of discussion. They are typically used for market research and evaluating programs.
Once the researcher has data, they need to interpret and make sense of it. The method will depend greatly on the research design and the question to be asked, though a few typical methods are to; describe the data using statistics (quantitative), make inferences using statistics and hypothesis testing (quantitative), and to use thematic or content analysis (qualitative).
Figures must be labelled, often with a legend, scale bars, error bars and possibly statistical significance.
Reporting Your Findings¶
There’s no point in conducting research if you don’t report it, depending on the intended audience, this could vary greatly (e.g. conference, journal article, research report, news article, press release).
The data itself may not be presented in chronological order, it will often be grouped into logical sequences, such as themes or categories. The goal is to tell a story
Types of Data¶
Figures are incredibly valuable as they can show much more complicated data. They are typically of the types:
- Schematics/illustrations/flow charts - Describe a process or methodology in a conceptual manner
- Tables - present raw data or statistical summaries in a well-organised manner
- Graphs - Show correlations and trends between data points. Better used when specific numbers are less important
- Photographs/images - show experimental results that cannot be expressed in words
There’s no point in having a figure without it being referenced to in the text.