How Researchers Gain Your Trust

The two different types of research are, of course, quantitative (Ninja) and qualitative (Pirate). As a researcher, there are certain things to keep in mind to insure that their research gains trust and is credible. Since quantitative is very calculated and numerically based research, Ninja's have to gain your trust differently than the flexible, person based, qualitative Pirates. The following chart shows some of those differences and how these two show their credibility.

QUANTITATIVE	QUALITATIVE
Internal Validity: Ninjas form a hypothesis about a relationship between an independent variable and a dependent variable and then figure out a way to test the relationship. The testing of the relationship is to see if the independent variable affects the dependent, and the results that the researchers attain tells about the internal validity. It is also helpful to gain trust by discussing the test and the materials used and how the tests were performed so the readers can tell if the tests were even relevant to what the hypothesis was suggesting.	Credibility: There are multiple techniques for Pirates (naturalists) to consider when trying to make their research credible. -Include activities such as prolonged engagement and persistent observation that add to their data and the credibility. Triangulation (using different methods, theories, etc.) is another way to incorporate more activities that can show credibility with your data. -Have fellow researchers or other peers that are not involved in the study at hand in order to get their view of the research. Feedback from these peers helps the researcher work on further credibility. -A technique also used is called “negative case analysis” where the researcher continuously adjusts the hypothesis based on new data. -Pirates should also find previous data to refer to and compare their own information gained. This is called referential adequacy. -Member checking is when the researcher gives the information they gleaned back to the participants to get their confirmation.
External Validity: Ninjas have a goal of generalizing the information from their research to the general public, and in order for this to occur their have to be random samples used and all demographic data must be recorded for participants.	Transferability: Pirates cannot generalize to the public the way ninjas can, however they can provide detailed information pertaining to the context of the study  and form hypotheses based on their research for the general public.
Reliability: This is when a ninja repeats a test over and over again, and if the results are the same or similar then it helps validate the information gained. The more a study is replicated and the results come out the same, the more reliable it is.	Dependability: You cannot always repeat the study techniques used by Pirates, therefore they try to establish dependability to ensure credibility (instead of reliability to validity.) Some techniques such as “overlap methods” and “stepwise replication” can be used in this area. Also, the research can get someone to audit the study.
Objectivity: This is just validity found through having multiple researchers agreeing separately on a phenomenon.	Confirmability: This is found by having an audit completed on the study that confirms the trustworthiness of the study. This is a fairly lengthy process used instead of objectivity.

Article Review Two

The Mental Health of Homeless School-Age Children

This is a quantitative research study that sought to find out information on the mental of health of homeless children as well as whether gender or ethnicity played a role in that mental health and if the children's self reports were aligned with the parent's reports.

Participants: Convenience sample of children from age 8 to 12 in homeless families. The homeless status was determined by whether the mother perceived themselves to be homeless and if the family was staying at a place in which they had not intended to stay at for more than 45 days. The total amount of children participating was 46 from 46 different families.

Data Collection: The participants are all from a midwestern metropolitan city. The homeless children and mothers were found within agencies that provided shelter and services to homeless families, soup kitchens, and a community settlement house. Also the researchers walked around the downtown area to find homeless families that may not use agencies or the other services provided. The researchers then interviewed the participants who agreed to the complete the study in as private a setting as possible, typically a closed off room in the place they were residing. The mother and child completed the interviews including a mental health survey separately.

Data Analysis: The information gained was analyzed and then the results from the two surveys were statistically compared to see if there was a significant difference between the two. There was also an analysis to see if there was a significant difference between the mother's perspective of the child's mental health and the child's perspective.

Conclusion: According to the researchers, the results show that the group met criteria for additional depression evaluation. The results also showed that there were no statistically significant differences between African American and Caucasian nor male and female. In regards to the information the mothers provided, the results showed that 26% of the children studied should have additional evaluation for other mental health issues.

This study did not have a large enough sample for it to be generalized to the population. Also, the fact that the sample was convenient means that the results could be skewed. These two reasons are the main ones that would lead me to not have faith in the results of the study. It could be seen as helpful in regards to showing that homeless children may need evaluation for mental health issues, but there was no comparison to children who are not considered homeless so the difference may not be significant. Therefore, the information in this study does not seem worth reading.

Menke, E. M. (1998). The mental health of homeless school-age children. Journal of child and adolescent psychiatric nursing, 11(3), Retrieved from http://onlinelibrary.wiley.com.proxy.library.vcu.edu/store/10.1111/j.1744-6171.1998.tb00020.x/asset/j.1744-6171.1998.tb00020.x.pdf?v=1&t=h8nw4cil&s=4844f9ff39d2ded2b41894e91f65faed92308ed3&systemMessage=Wiley Online Library will be disrupted on 27 October from 10:00-12:00 BST (05:00-07:00 EDT) for essential maintenance

Article Review One

Psychosocial Characteristics of Homeless Children and Children With Homes

This study was a quantitative study that set out to compare children of homeless families to those that were poor but had homes, both of which were single mother families. 

Participants: 49 families with 86 children living within six family shelters in Boston, Massachusetts between April 1985 and July 1985. The mothers within these families were a mean age of 28 years old and had a mean number of 2.4 children. It was stated that 33% of the mothers were white and Hispanics were highly underrepresented.

For the control, the poor families that had housing were selected based on their willingness to participate once asked by researchers who came to their door. The houses that were tried had been chosen based on the neighborhoods poverty rate in Boston. There were 820 houses selected and only 81 families participated with a total of 134 children. Mean age of these mothers was 29 and mean number of children was 2.5. The mothers of these families were 32% white.

The age of the children for both variables was a mean of around 6 years old.

Data Collection: The data was collected within the homes and shelters of the participants and little privacy was considered. One researcher would interview the mother while another researcher would interview the children. The interviews included behavioral checklists as well as demographic information collection. Interviewing time varied based on interruptions, the child's ability to focus, the setting, as well as the need for the researcher to assist the participant.

Data Analysis: A statistical analysis was performed on the information gained using the behavioral checklists to find differences that were statistically significant (P<.05). Also, since some of the mothers had more than one child and that information may skew the data, the researchers also picked one child from each family and repeated the analysis.

Conclusion: The results showed that the homeless children generally scored worse on the behavioral checklists compared to the children with homes, but the differences were not significant except for with one of the five checklists used. The one significant finding was that preschool students that are homeless have more developmental lags in comparison to preschoolers with homes.

When reviewing this study, I find that I would not put much faith into the results being given. The way the researchers went about finding the sample for the control was very convenient because it only included those people that answered their doors and agreed to complete the study. A convenience sample is never very trustworthy when it comes to studies. Also, the sample size was fairly small therefore can not be generalized to the homeless population as a whole.

The research article did not discuss the validity of of the checklists used in the interviews. Also the interviews were not completed in a setting that was conducive to best results. It was stated in the article that the checklists given in the beginning of the interviews are probably more valid compared to those given later in the interview when the children's focus was waning. 

Overall I believe this study could have been done in a better way.

Bassuk, E. L., & Rosenberg, L. (1990). Psychosocial characteristics of homeless children and children with homes. Pediatrics, (85), 257. Retrieved from http://pediatrics.aappublications.org/content/85/3/257

Things that could make my research suck.

Things that could threaten internal validity.

Research Question: Is homelessness associated with higher rates of introverted behavioral issues in grade school students?

History: This could have a fairly large effect on my research question considering the behavioral issues could be due to something that has happened previously in the student's life other than being a homeless or transient student. To some students, being homeless or transient may not be something that they have a very hard time dealing with, whereas there are other possible situations that could cause more harm, like a divorce or death of a parent. Every person's reaction to a situation is different.

Selection: The selection is based solely on information within the student's records at the school that the counselor doing the research would be working in. This could be leaving people out because of a lack of records. Also, since the selection is limited to the school that the counselor is working in, then the sample will be very small and may not produce much information.

Maturation: Would I only be doing the research on students that are currently homeless/transient? Or would I be looking at the students that have been in the past as well? If I look at student's that were previously homeless, then their behavior may be different because of the time difference. Also, it may depend on how long a student has been homeless, because those that have dealt with it for a long time may be affected differently than those who are newly homeless.

Pretesting: This would not be an issue with my research because I would not be doing any pretesting, because there is no way to test the students before they were homeless. Unless I do a behavioral test and then a certain amount of time later do another one in order to see if there is a change with the amount of time being homeless, but I do not think that is the direction I would go.

Instrumentation: This could be an issue with the validity of my research if the survey given regarding the behavior is not reliable. Also, it could be a threat if the survey is not particularly related to the variables that you want to study. It is important to try to match the instruments to the variables.

Treatment Replication: When it comes to my research question, replication within the same school may not be very useful because chances are the information would not change much, unless there has been some sort of intervention set in place to try to help the students with the behavioral issues. However, having other counselors replicate the study at their schools would be very helpful in order to get a higher sample size.

Subject Attrition: Students dropping out of the study could affect the internal validity, however, with my research question, there is only a one part study, therefore the participate once and that is it. Either they participate or they do not, so there is no room to drop out and skew data.

Statistical Regression: This is a threat to validity when you administer and pretest and post-test. It is the chance that the participants who score low on the pretest will score higher on the post-test whether the intervention worked with them or not. This would not be an issue with my participants since there will not be a set up like this one.

Diffusion of Treatment: This is definitely a possible issue with the study that I am proposing. If the students  that are homeless are aware of each other, they may discuss what the study is about and how they feel about being homeless or transient. This is unlikely, however, because it is not a topic commonly brought up between students. It is possible, but unlikely.

Experimenter Effects: Being a counselor in a school means that the students in your research have most likely interacted with you before. This interaction, or the way the counselor is dressed, or the demeanor of the counselor when administering the survey, may have an effect on how the student answers questions. 

Subject Effects: There is always a chance that the student will answer differently because they know they are being researched. This is where "subjects effects" could cause the validity of the research to go down. Some students do not feel comfortable being honest, or they will answer the way they think the researcher wants them to.

There's a lot that could cause problems.

"Fastest-Growing" XKCD

 In honor of

People who drown love ice cream O.O

In every statistics class I have ever taken I have heard the story of the correlation between amount of ice cream sold and deaths by drowning. Supposedly a study was done that found that both rates increase at the same time! Well we all know what that means... Ice cream must cause people to drown!

...or maybe people who drown love ice cream?

...or could there possibly be an unknown third variable!?

... or maybe, just maybe, it could have been by chance!?!

Who will ever be able to solve this problem!?!?

And that is where you find the downfall to comparative and correlation designs.

When comparing two variables, like ice cream and drowning, you may find that their rates both increase at the same time but maybe it is because there are other variables in play. (McMillan, 2012) In this case, it could be the fact that more people eat ice cream when its hot, which is also a great time for everyone to go swimming. Another example could be that, according to a study by Paul Kirschner, the more a person uses facebook the lower their GPA will be. (Kirschner & Karpinski, 2009) Other variables that may cause these to seem related could possibly be that a person who is on facebook all the time does not attribute enough time to studying, and then the lack of studying could relate to the lower GPA. So many other possibilities!

Another problem with correlation studies is that the direction is unknown. (McMillan, 2012) People who use facebook a lot may have lower GPA's because of it, or maybe people with low GPA's happen to really like facebook. One of the girls in the movie Mean Girls hit it spot on:

Ouch! 

Then there is the issue of maybe it is all just coincidence. (McMillan, 2012) What if there is no actual relationship other than the two variables are changing at the same time. What if kids who use facebook a lot just happen to have lower GPA's and there is no real relationship between the two. One could probably safely assume that the amount of ice cream sold does not affect how many people die by drowning. Unless ice cream makes you so dense that you cannot float, or it has some sort of magical quality to it that just makes your body incapable of floating. Those are bad thoughts though, because ice cream is too delicious to cause any harm other than a few extra pounds of fat........ which might even help you float O.O

I should stop there, and leave you with another nifty comic.

Just remember, Correlation does not equal causation!

McMillan, J.H. (2012). Educational research: Fundamentals for the consumer (6th ed.). Boston: Pearson.

Kirschner, P.A., & Karpinski, A.C. (2009). Facebook and academic performance. Retrieved from: http://lnx-hrl-075v.web.pwo.ou.nl/bitstream/1820/2880/1/Facebook_and_Academic%20Performance.pdf