Research Methods I Pt. 2, Articles

Question 1

Surveys

Answer 1

Series of questions based on a specific construct at hand.

Question 2

Response set

Answer 2

Tendency to respond to questions from particular perspective rather than providing answers related to questions, affect usefulness of data obtained from self-reports.

Question 3

Survey Questions

Answer 3

Focused on attitudes and beliefs, facts and demographics, and behaviours

Question 4

Creating a Survey Question

Answer 4

Difficulties arise with unfamiliar technical, vague, terms keep it simple.

Question 5

Double-Barrelled Questions

Answer 5

Ask two things at once, ask two questions not one

Question 6

Loaded Questions

Answer 6

Written in manner to make participants answer in a certain fashion.

Question 7

Negative Wording

Answer 7

Negative words are used to elicit agreement and result in inaccurate answers.

Question 8

Yea and Naysaing

Answer 8

Response set where person agrees or disagrees with whatever you say.

Question 9

Context-effect

Answer 9

Have options in two response set in different location. A and D switch position.

Question 10

Confidentiality vs. Anonymity

Answer 10

Anonymity is used for sensitive answers, accurate responses, confidential means you know their name but will not disclose answers publicly.

Question 11

Close-ended Question

Answer 11

Limited responses, alternative answers exist, give well-defined dimensions, rating scales are common.

Question 12

Variation on Rating Scales

Answer 12

graphic rating scale has a mark along continuous 100mm line, semantic differential scale measures concepts on dimensions, evaluations, activity, and potency.

Question 13

Nonverbal scale for children

Answer 13

Give ratings despite inability to understand harder concepts, alternative responses and their meanings are decided upon by participant.

Question 14

Finalize Questionnaire

Answer 14

Make it appear attractive and professional with neatly typed sentences free of spelling errors.

Question 15

Refining Questions

Answer 15

Pilot test the survey and have participants think aloud while answering questions, get them to tell you their interpretation of the question.

Question 16

Administrating Surveys

Answer 16

two ways to administer: questionnaires, and interview format

Question 17

Questionnaires

Answer 17

Written format, less costly, complete anonymity, require comprehension and reading ability, problem of motivation as well.

Question 18

Personal Administration

Answer 18

Often give questionnaires to groups of individuals, class, meeting, orientation, appointment, captive audience likely to complete questionnaire, present to answer questions if necessary.

Question 19

Mail Surveys

Answer 19

Inexpensive way by sending to homes/business, drawback due to potential for low response rates, can be forgotten, distraction may occur, bored and throw in trash, no one present to answer questions.

Question 20

Internet Surveys

Answer 20

Sampling problem, listed in search engines and discover people interested in providing info on topic. Sample can be chosen from database and invite is sent. One problem is whether results are same as traditional method one might use. Not extensive, indicate data is comparable though, ambiguity about characteristics of individual, people may mispresent their individual. While unlikely is still possibility.

Question 21

Other Tech

Answer 21

Pagers, cell phones, other wireless communication devices allow people to be contacted at various times and provide immediate feedback.

Question 22

Interviews

Answer 22

The interaction has many implications of its own, likely to answer questions for real person compared to mailed questionnaire, skilled in convincing participation, higher response rates, rapport is established to answer all questions and complete survey. Can verify question is but interviewer bias arises with show approval or disapproval. Screening prevents this and limits the biases.

Question 23

Face-to-face interviews

Answer 23

Require meeting to conduct interview, expensive, time-consuming and typically used when sample size is small and clear benefits to interaction.

Question 24

telephone interviews

Answer 24

large-scale surveys are done by telephone, inexpensive compared to face-to-face, data collected quickly thanks to working on same survey at once, lower cost of telephone survey with labor and data analyst costs. CATI system has questions prompted on screen then entered into computer.

Question 25

Focus Group Interviews

Answer 25

Group of 6-10 individuals for 2-3 hours, any topic explored in group, selected members have knowledge or interest in topic, spend time and incur costs traveling to focus group location with monetary or gift incentive to participate. Open-ended questions are given typically, skilled interviewer to work with groups, facilitate communication and deal with problems arising, recorded and transcribe meeting, analysed records for themes and consensus and disagreement. Two to three groups to ensure topic is not specific to one group, yet time-consuming and costly, with lots of information so don't do very many groups on any one topic.

Question 26

Panel Study

Answer 26

Method to study changes over time, two-wave panel study has people surveyed at two or more points, three-wave has three points, good for establishing connection of variables at one point in time and another.

Question 27

Confidence Intervals

Answer 27

An interval between two points where the you are confident the true population values lies, best estimate is sample value but only a sample and not the entire population so may be wrong. Formally known as sampling error.

Question 28

Sample Size

Answer 28

Larger sample sizes reduce size of confidence intervals, determined by other factors but likely to yield data reflecting true population value.

Question 29

Simple Random Sampling

Answer 29

Every member of population has equal likelihood of being selected for sample.

Question 30

Stratified Random Sampling

Answer 30

Population divided into subgroups and random sampling techniques select sample members from each stratum, dimensions used to divide into subgroups but relevant to problem of study. Built-in assurance of accurate reflection on numerical composition of subgroups.

Question 31

Cluster Sampling

Answer 31

Identify clusters of individuals and sample from clusters, requires sample from larger to smaller clusters, multistage approach, advantage here is not sampling from lists of individuals to obtain random sample of individuals.

Question 32

Haphazard Sampling

Answer 32

Convenience sampling has a sample of students selected in a convenient manner, introduce biases into sample, no accurate representation of population of interest, limit ability to use sample data to estimate true population values.

Question 33

Purposive Sampling

Answer 33

Purpose being to obtain sample of people meeting predetermined criterion, limits sample to certain group but not probability.

Question 34

Quota Sampling

Answer 34

Sample chosen which reflects numerical composition of subgroups in population, similar to stratified sampling, you ensure you have a certain percentage of people, problem is no restrictions on how we choose individuals.

Question 35

Sampling frame

Answer 35

Actual population from which random sample is drawn, rarely coincides with population of interest, biases are introduced, need to consider how the frame matches the population of interest, while biases can be minor they can be consequential.

Question 36

Response Rate

Answer 36

Percentage of people in sample who completed survey, indicates how much bias is in final sample, those who respond differ from those who do not in many ways, general rule is the lower the response rate, the greater likelihood of biases distorting findings limiting ability to generalize findings. With all the survey methods need follow-up reminders, second mailing of questionnaires, incentive may be necessary, like cash, gift, gift card, raffle ticket. then convince the purposes are of the survey are important and the valuable contribution.

Question 37

Matched-Groups design

Answer 37

Matched sets of subjects are randomly distributed, assess sample on characteristics believed to exert influence on dependent measure and group subjects who characteristics match.

Question 38

Logic of Matched-Groups Design

Answer 38

Effect of characteristic you are interested in is distributed evenly, and contributes little to difference between group means, effect of error variance minimized, likely that effects of independent variable are detected.

Question 39

Pros and Cons of Matched-group design

Answer 39

Control subject variables which would obscure independent variable effect under investigation, can increase sensitivity to effect of independent variable, potent advantage to detecting effects otherwise missed. May demonstrate given effect with fewer subjects, saving money and time. Risk concerns matched characteristics not affecting dependent variable under conditions of study, require modified inferential statistics in unmatched, randomized design, less powerful so no discriminating effects from uncontrolled, extraneous variables.