SCI 201 Chapter 3 - Data Collection (Terms)

Random Error (Variability)

Error in measurement due to random chance or natural variability despite random sampling from the population

Systematic Error

Error due to bias that consistently occurs in the same direction, giving measurements that are too high or too low or otherwise biased

Sampling Error

Error that results from using a single sample statistic to estimate a population parameter

Standard Error of the Estimate

A measure of the sampling error, which is usually the denominator of the test statistic and takes into account sample size

Sampling Variability

The spread or deviation that occurs from one sample to the other when taking several samples from the same population

Instrument Error

A source of error in data collection in which instruments may give consistent systematic errors due to inaccuracy or imprecision

Researcher (Observer) Error

A component of observational error in which subjectivity can be involved in the process of collecting data

Precision

A term used in natural sciences to describe the closeness of repeated measurements to the same quantity, which implies how low the level of random error is

Accuracy

A term used in natural sciences to describe the closeness of measurements of a quantity to the true value of said quantity, which implies how low the level of systematic error is

Reliability

A term used in social sciences to describe the consistency or dependability in a measurement, which is the opposite to random error

Validity

A term used in social sciences to describe the trueness of a study, where we measure how accurately a conclusion “fits” the reality or what it actually aims at studying

Face Validity

A type of validity measurement measuring whether the variables and methods in the study appear scientifically acceptable at face value

Content Validity

A type of validity measurement measuring whether the content of the variables measured captures the content in the aim of the study

Statistical Validity

A type of validity measurement measuring whether the most appropriate hypothesis test or statistical procedure is selected based on experimental design

Pilot Study (Preliminary Survey)

A small trial run or survey conducted before an actual experiment in order to test the suitability of the research design or sampling strategy

Questionnaire

A method of data collection used in social sciences in which a list of questions is filled out by the respondent

Interview

A method of data collection used in social sciences in which a list of questions is asked verbally by an interviewer, then respondent answers are recorded

Mail-Out Questionnaires

A type of questionnaire that is sent out at random to people’s residences or workplaces

Internet Surveys

A type of questionnaire that is available to respondents online, either sent via email or by uploading to a website

Face-to-Face Interviews

A type of interview in which interviewers verbally question the respondent in person

Social Desirability Bias

Bias that occurs in interviews or unanonymous questionnaires where the respondent is reluctant to answer questions honestly in fear of releasing sensitive information or having answers that make them look bad to the interviewer

Interviewer Bias

Bias that occurs in interviews where the respondent is influenced by the way an interviewer presents a question, or the interviewer’s record of the respondent’s answer is biased

Telephone Interviews

A type of interview in which interviewers verbally question the respondent over the telephone, often done by random digit dialing

Structured Questionnaires/Interview Schedules

Questionnaires or interviews with only closed-ended questions that have a fixed number of mutually exclusive answers on a nominal scale or an ordinal scale

Unstructured Questionnaires/Interview Schedules

Questionnaires or interviews with open-ended questions with a numerous possibility of different answers that are not limited by the interviewer

Leading Question

A type of bias-inducing question in which information is introduced at the start of the question that implies a “correct” answer to the respondent

Loaded Questions

A type of bias-inducing question in which wording or information in the question makes the question emotionally loaded, implying a “correct” answer to the respondent

Contingency Questions

Additional questions asked if an answer to a previous question makes it necessary or advantageous to gain more knowledge

Focus Group Discussions (Focus Group Interviews)

Discussions of 2-3 hours held in groups of 6-10 specifically selected people that focuses on a particular topic relevant to the study, potentially leading to discussion of related topics

Key Informants

Individuals specifically selected for an interview or questionnaire because of their deep knowledge or experience in the particular topic under study