Wednesday, April 11th
Good work in class this week. I'm looking forward to reading your research assignments and should have grades posted by Friday evening. Here is a review of chapter 5.
Every score is made up of random error and measurement error
Potential threats to Internal Validity
How the research is conducted
Multiple Administration
Internal Consistency
Measurement Validity
Content validity (face, criterion, construct)
Good treatment
Environmental influences
Effects due to research participants
Hawthorne Effect
Selection (self selection, regression effect, ceiling and floor effect)
Mortality
Evaluation apprehension
Effects due to researchers
Research personal attribute effect
Researcher unintentional expectancy effect
Researcher Observational Bias (observer Drift, observer bias, halo effect)
Potential threats to external validity
Sampling
Random sample (simple, systematic, stratified, cluster)
Nonrandom sample (convenience, volunteer, purposive, quota, network sample)
Ecological validity
Replicate findings
Constructing Scales
Question for thought:
Think of a research question where a nonrandom sample would be needed. Describe the study, the nonrandom sample, and why this sampling would be needed.
Every score is made up of random error and measurement error
Potential threats to Internal Validity
How the research is conducted
Multiple Administration
Internal Consistency
Measurement Validity
Content validity (face, criterion, construct)
Good treatment
Environmental influences
Effects due to research participants
Hawthorne Effect
Selection (self selection, regression effect, ceiling and floor effect)
Mortality
Evaluation apprehension
Effects due to researchers
Research personal attribute effect
Researcher unintentional expectancy effect
Researcher Observational Bias (observer Drift, observer bias, halo effect)
Potential threats to external validity
Sampling
Random sample (simple, systematic, stratified, cluster)
Nonrandom sample (convenience, volunteer, purposive, quota, network sample)
Ecological validity
Replicate findings
Constructing Scales
Question for thought:
Think of a research question where a nonrandom sample would be needed. Describe the study, the nonrandom sample, and why this sampling would be needed.
posted by Dr. Jeffrey Bolt at 8:08 PM
12 Comments:
I would think that a study about a certain group of people would need a nonrandom sample. For Instance, a study looking at the opinions of African American baseball players on why they have such limited representation in the majors. The question couldn't be asked to all players because it would not be relevant to all players.
Drew Wilson
I think it would be interesting to see whether cetain opinions of students who attended both grade school and high school at Catholic institutions changed after a year at a large, public university. It would be a nonrandom, puposive sample because only those who went to Catholic school for twelve years and then moved on to a university with 20,000 or more students could be studied.
Amanda Gabriele
I would use a non-random sample in pharmaceutical studies that require definite results for comparison. For instance, studying a drug that is designed for best results in a population of people with bipolar or manic depressive disorder need not be tested in all those with mental illness. A random sampling including those with clinical depression, personality disorders or other illness would be unnecessary with a need as defined as a specific pharmaceutical for a certain disorder.
Steph Repp
I would use a nonrandom sample to carry out a study asking mothers, single or married, their opinions about "the working mother", "the working woman", or "the woman that can do it all" and if this is even possible. This study would be classified as a nonrandom purposive sample, because the women in this study all have the same characteristic, they are all mothers.
Ashley Mayle
I think an instance where you would need to use a nonrandom sample would be trying to see if dorm life was enjoyable for college students. You would be unable to ask all college students because not all of them have lived in the dorms. You would only be able to ask the college students that had experienced dorm life.
Machelle Montgomery
I think that it would be intersting to do a study on what college students think about radio host Don Imus being fired. This would be a nonrandom convenience sample because we couldn't ask all college students because not all college students would respond to the survey. It would also be convenience because not all college students might have heard of Don Imus or know what's going on with Don Imus.
Richard Toth
I would ask Kent State student athletes to answer questions regaurding their thoughts on preferential treatment. This is a nonrandom sample because we would only be asking the athletes to get the opinion from students that may or may not experience it first hand.
Megan Ebert
I would conduct a study with the goal to find out the various reasons students at Kent State take longer than 4 full years to graduate. I would administer a short questionaire to Kent State students and have the students respond with their reasons. It would be a nonrandom, purposive sample because only students who have been enrolled more than four years at Kent would be studied. A nonrandom sample would be good here because it would be difficult to sample all the students who take more than four years. Also, I believe that we would find many of the same select few reasons for taking longer than four years (money, grades, work, kids, etc).
-James Sherlock
I think it would be interesting to do a nonrandom sample on college students that work full time. Do they think they have a harder time in school because of the interferece of work? It would be nonrandom because the question could only be applied to students who worked full time during the semesters and summer.
Tiffani Clark
I would use a non-random sample on all students involved in greek life, when I ask, Is it difficult to focus on being involved in your fraternity/sorority and still maintain a high GPA? This would be nonrandom because the question would only be applied to students involved in greek life.
Leah Verbosky
i would use a non random sample of students having 7:45 am classes. I would conduct the study to determine whether they wake up on time for class. this would be different than the results if the students samples were students who had either night classes, commuter students who didnt live close to their classes or students who had class after noon.
Coleman Caster
the idea of a nonrandom sample could be used for a case like why people in a town were getting sick you couldnt ask the whole population you would to just need to ask those people in the town.
Justin Quinlan
Post a Comment
<< Home