The post Interpreting Your Data Analysis: How to Determine Statistical Significance appeared first on Statistics Solutions.
]]>With p values, t values, F values, correlation coefficients, and a bunch of other numbers staring at you, it is easy to get discouraged. However, the basic question you need to answer, do I or do I not have statistical significance, can be answered looking at one simple number: the p value.
Before you can determine if you have rejected or failed to reject your null hypothesis, you must designate the maximum probability of falsely rejecting the null hypothesis that you are willing to accept in your analysis. This is referred to as the alpha level and is typically set at .05 in social science research. An alpha level of .05 means that you are willing to accept up to a 5% chance of rejecting the null hypothesis when the null hypothesis is actually true. Depending on your field of study and the nature of your analysis, you may choose to decrease or increase the alpha level to make the decision point more or less stringent.
Once you conduct your analysis, you will get a p value, also called a significance (Sig.) value. Your statistical software package will return this number to you once you conduct your analysis. This number reflects the probability of obtaining results as extreme as what you obtained in your sample if the null hypothesis was true.
Let’s give this concept some legs with an example. Our research question asks: are there differences in the number of women hired in higher education institutions by region? The null hypothesis we are testing is: there are no differences in the number of women hired by region. We conducted a oneway ANOVA in order to compare the number of women hired in each region. The alpha level for this analysis is .05. We conducted the analysis in SPSS and got the following output:
ANOVA 

Sum of Squares 
df 
Mean Square 
F 
Sig. 

Between Groups 
182800.683 
2 
91400.342 
1.132 
.375 
Within Groups 
565205.417 
7 
80743.631 

Total 
748006.100 
9 
We have a p value (Sig.) of .375. This number exceeds our alpha level of .05. Therefore, we fail to reject the null hypothesis. We did not find statistically significant differences in the number of women hired in higher education institutions by region. If the p value had been less than .05, we would have rejected the null hypothesis.
The post Interpreting Your Data Analysis: How to Determine Statistical Significance appeared first on Statistics Solutions.
]]>The post Is There a Purpose in NonPurposeful Sampling? Sampling Strategies for Quantitative and Qualitative Research appeared first on Statistics Solutions.
]]>One of the first steps to deciding on a sampling strategy is to know whether your study is going to be qualitative (mainly using interviews or observations) or quantitative (using numerical data from surveys, census data, etc.). Qualitative research is usually intended to hone in on a specific focus, while quantitative research is usually intended to be generalizable to a broader sense. This distinction is usually what draws the line between purposeful and random sampling. Based on intent of your research, you can look at both forms of sampling like this:
Specific focus: If you are interested in one tidepool from the ocean, you can take a single vial of the water and be pretty sure that whatever you find will be applicable to that pool. However, saying that what you find will hold true for the entire ocean would be a long shot. The purpose of this tight focus would be to learn lots of information about a very specific thing.
General focus: Using the same analogy, a researcher would not try to learn about global water temperatures by taking a vial from one tide pool. The best way to do this would be to take to the sea and randomly collect vials from many different areas to get a general idea of what the average temperature is. The randomness here helps to even out confounding things like water currents, random temperature spikes, distance from the equator, and the like.
Later on, we will talk a little more about the different methods within the overall schools of random and purposeful sampling. Stay tuned!
In 50 Word or Less: If you are conducting a qualitative study, you probably want to use purposeful sampling. If your study is quantitative, random sampling is usually best. However, this is just the first step in the decisionmaking process, so it does not always hold true.
The post Is There a Purpose in NonPurposeful Sampling? Sampling Strategies for Quantitative and Qualitative Research appeared first on Statistics Solutions.
]]>The post What is the Difference Between Population and Sample? appeared first on Statistics Solutions.
]]>First, your sample is the group of individuals who actually participate in your study. These are the individuals who you end up interviewing (e.g., in a qualitative study) or who actually complete your survey (e.g., in a quantitative study). People who could have been participants in your study but did not actually participate are not considered part of your sample. For example, say you emailed study invitations to 200 people on a listserv and 100 of them end up participating in your study (i.e., complete your survey or your experiment). Your sample is the 100 individuals who participated in your study. The 100 individuals who received invitations but did not participate would not be considered part of your sample; rather, they are part of what is often called the sampling frame. Your sampling frame is the group of individuals who could possibly be in your study, which in the above example would be the 200 individuals on the email listserv.
On the other hand, your population is the broader group of people to whom you intend to generalize the results of your study. Your sample will always be a subset of your population. Your exact population will depend on the scope of your study. For instance, say your research question asks if there is an association between emotional intelligence and job satisfaction in nurses. In this case, your population might be nurses in the United States. However, if the scope of your study is more narrow (e.g., if your study deals with a local problem or a specific specialty/industry), then your population would be more specific, such as “nurses in the state of Florida” or “licensed practical nurses in the United States.” Importantly, your population should only include people to whom your results will apply. For example, if you do not have good reason to believe that your results will apply to all nurses in the United States, then your population will need to be more specific. If you are stuck on defining your population, think about how you would fill in the blank in the following sentence: “The results of my study will apply to _____.” Your answer will help determine how you define your population.
To summarize: your sample is the group of individuals who participate in your study, and your population is the broader group of people to whom your results will apply. As an analogy, you can think of your sample as an aquarium and your population as the ocean. Your sample is small portion of a vaster ocean that you are attempting to understand. Properly distinguishing between these two concepts will aid you as you navigate the methodological details of your dissertation.
The post What is the Difference Between Population and Sample? appeared first on Statistics Solutions.
]]>The post Tackling Literature Review Organization appeared first on Statistics Solutions.
]]>The best way to go about organizing your literature review is to think about in terms of an inverted triangle. The broad base at which you are starting is your topic in its most general sense. This generally includes the basic history and the most important seminal works regarding your topic. For example, if you are studying perceptions of leadership styles in the banking system, this is where you are going to write about the basics of leadership styles, devoting headings to transformational, transactional, and laissezfaire leadership styles. As you work your way through your literature, you want to become more refined, narrowing the scope of the articles you are reviewing to the point where you are setting up your study as the next obvious step in the progression of the previous literature. That is really what the literature review is all about: honing your literature to the fine point that is your study. As you do this, remember that headings are your friend, as they help you to narrow the focus of your chapter.
Below is a visual guide to assist you with organizing your literature review
The post Tackling Literature Review Organization appeared first on Statistics Solutions.
]]>The post Misconceptions about Confidence Intervals appeared first on Statistics Solutions.
]]>Reporting confidence intervals is all well and good, but it is important to understand what confidence intervals are if you do include them in your results. Interestingly, confidence intervals are among the most commonly misunderstood concepts in statistics. These misconceptions have even been demonstrated by scientific research. For example, Hoekstra and colleagues conducted a study in 2014 showing that individuals with varying levels of statistics expertise (from students to expert researchers) frequently endorsed false statements about confidence intervals. This suggests that misconceptions are common not only among statistics novices, but experienced researchers as well! Some of the most common misconceptions about confidence intervals are:
So what exactly is a confidence interval? A confidence interval is an estimate of the possible values of a population mean; the key word here being estimate. Just as with any statistic estimated from a sample, the upper and lower bounds of the confidence interval will vary from sample to sample. For a given population, the 95% confidence interval from one random sample might be between 2 and 5, but for another random sample it might be between 1 and 4. Some of the intervals calculated from these random samples will contain the true population mean, and some will not. A 95% level of confidence means that 95% of the confidence intervals calculated from these random samples will contain the true population mean. In other words, if you conducted your study 100 times you would produce 100 different confidence intervals. We would expect that 95 out of those 100 confidence intervals will contain the true population mean.
To conclude, confidence intervals can be a bit difficult to wrap your head around, whether you are a beginner in statistics or an expert. But being aware of the misconceptions and avoiding them in your interpretation will help you (and your readers) develop an accurate understanding of your results.
Reference:
Hoekstra, R., Morey, R. D., Rouder, J. N., & Wagenmakers, E. J. (2014). Robust misinterpretation of confidence intervals. Psychonomic Bulletin & Review, 21(5), 11571164.
The post Misconceptions about Confidence Intervals appeared first on Statistics Solutions.
]]>The post Communicating a Positive Relationship with Your Chair appeared first on Statistics Solutions.
]]>It would have been easy for me to become defensive. However, not having gone through the experience of writing a thesis before, I was simply guilty of being ignorant of the process. So, I learned. I changed my approach and we moved forward. This was a valuable lesson that later allowed me to establish a positive working relationship with my dissertation chair, something that made the already challenging process of writing a dissertation a little easier.
Establishing a professional, respectful working relationship with your chair is a crucial part of keeping the dissertation process moving forward. Regular communication with your chair should reflect and reinforce this relationship. Ineffective communication can lead to tension and misunderstanding—or worse, silence—all of which can impede progress. Here are some tips for understanding how to communicate positively and effectively with your chair.
Be professional. Although you are not yet at your chair’s professional level, you are an advanced studentresearcher entering into the professional conversations of your field. In a sense, you and your chair are future colleagues. Do your homework, read the literature. Own what you know, admit what you do not know. Speak confidently, but not boastfully. Do not be afraid to ask questions about what you are still learning.
Be respectful. Just like you, chairs are human beings. As in my example above, chairs know when they are being treated shabbily. Although chairs should maintain a degree of professionalism, they may take umbrage with nonprofessional and disrespectful communication, such as unreasonable requests, emotional outbursts, and defensive language. Being professional and being respectful go hand in hand.
Be consistent. Above all, stay in regular contact with your chair. Although most chairs will encourage regular communication, the onus is on you as an advanced learner to keep in touch with your chair. A major reason why people get stalled in the dissertation process is falling out of contact with their chairs. To maintain progress, it is important to set and keep regular meeting times and work deadlines. Keep your chair in the loop about your ideas and your progress, as well as any changes that may affect your work or your work schedule.
Writing a dissertation involves a high degree of collaboration between learner and mentor over an extended period. For such collaboration, regular and effective communication is key. Establishing and maintaining professional and respectful communication with your chair are crucial to keep the dissertation process moving forward.
The post Communicating a Positive Relationship with Your Chair appeared first on Statistics Solutions.
]]>The post How to Update Your Sources appeared first on Statistics Solutions.
]]>The first thing we would recommend is checking to see how many of your sources have published updates or new editions of the same research. The easiest way to do this is to go to Google Scholar and type in the name of the source or the author(s) and the publication year. If you are able to find your source this way, there will be a link at the bottom of the Google Scholar hit that says “All _ versions.” If you are lucky, there will be the exact same article published within the past five years there. If there is no current publication of your source listed there, you can try typing in the authors’ names in the Google Scholar search bar. It is possible they have published more works in the same field, and may have a more current publication that relates very closely to the one you have used. That would be a suitable replacement for your source.
If you need to replace an outdated book, you can go to Amazon’s website and type in the title of the book in the search bar there. Amazon always has the newest editions of books published, and we have had a lot of luck finding newer editions of books that clients have cited this way.
Finally, and most frustratingly, what do you do if there is no newer publication or edition to the one you have used? We recommend going back to Google Scholar and typing in the information for your source again. Right next to the button that reads “All _ versions” is one that says “Related articles.” If you click on that, Google Scholar will bring up a list of articles with similar keywords to the ones in your original article. You can also click on “Cited by _” to see which researchers have cited that article in their publication. Either of these methods can, with some additional notes and adjustments to your literature review, provide an updated source to replace anything that is too old for use in your dissertation.
The post How to Update Your Sources appeared first on Statistics Solutions.
]]>The post The Importance of Assumption Testing appeared first on Statistics Solutions.
]]>In thinking about it another way, suppose you visit your favorite butcher to buy deli meat. You purchase a container of presliced meats because you know the butcher and he slices your deli meat just the way you like it. Without inspecting the package, you confidently purchase your meat, and take it home to prepare a sandwich. Upon opening the package, you see that the meat is sliced thinner than you prefer and immediately conclude that your butcher is losing his touch. You purchased your meat and made your conclusion regarding your butcher’s skills based upon the assumption that he was the one that sliced your meat. However, your conclusion is incorrect if your assumption is incorrect. Your assumption would be incorrect if the butcher had not prepared the meat, but instead it was prepared by his apprentice.
The statistical assumptions of your study are based upon the type of analysis you plan to conduct. I recommend Field’s (2013) text regarding statistical analyses. Field provides explanations of the different statistical analyses, the assumptions of the analyses, and the tests for the analyses that are very comprehensible for any level of researcher.
Assumption testing helps you to ensure that you are not drawing false conclusions from your analysis. As you develop your results chapter you will be tasked with numerous steps to develop a solid, scholarly dissertation. Testing your assumptions is just one of those steps. Hopefully this helps underscore the value of assumption testing and helps you proceed smoothly through your dissertation journey.
Field, A. (2013). Discovering statistics using IBM SPSS statistics. Sage.
The post The Importance of Assumption Testing appeared first on Statistics Solutions.
]]>The post Snapshot or Marathon: Why to Use CrossSectional or Longitudinal Research appeared first on Statistics Solutions.
]]>Crosssectional research is a kind of research that is focused on relationships between variables, but does not need to take into account the fact that any variables might fluctuate over time. This kind of research is usually used in correlational studies, but can also take the form of a comparative study. The importance of crosssectional research is in taking timewise fluctuations out of the data, taking a crosssection of the data, and looking at relationships or differences that exist at that moment.
An example of research that takes this form would be the examination of student test scores in relation to teachers’ perceptions of the subject matter. In a study like this, the goal of taking a crosssection would be to measure student scores, which could fluctuate over time, and also to measure teachers’ perceptions of the subject matter, which could also change from one day to the next. By taking a freezeframe of the data, the researcher is able to check whether high student scores also correspond with positive perceptions of the subject matter (or vice versa) without worrying about the fact that both may be slightly different next time they are analyzed. If the two are assumed to correlate, even if they were examined at a freeze frame of a second time, the same direct or inverse relationship should exist – if a teachers’ perceptions have become more positive, student scores should still correspond with those perceptions at any specific time. Though changes over time might be interesting, the crosssectional approach captures both variables at the same point in time for a more accurate comparison.
Longitudinal research is used when a researcher is interested in longlasting effects, the influence of a treatment or intervention, or trends over time. The importance of this kind of research lies in an ability to examine how things change; it is this ability that lets the researcher make causal interpretations. Though it is not always classified as longitudinal, a good true experimental design often has a longitudinal component, meaning it tracks at least one prescore and one postscore, though more subsequent measurements can give the analysis more detail into longterm effects. A second kind of study, time series studies, are the very definition of longitudinal research, and make the main goal an assessment of trends over time, sometimes going so far as assessing whether two trends correspond to one another.
An example of longitudinal research might take the form of a time series study where researchers assess global temperatures over time. By taking the same measurement over and over, researchers might have the goal of checking for a repeating trend (think seasons), and whether this trend has held steady, or begun to deviate from what is measured as typical. A different form of longitudinal research is an experimental study. A strong experimental study makes use of the multiple time points to check both a treatment and a control group for changes after some kind of intervention. This is the gold standard among many researchers, mainly because of the causal inferences that experimental studies can make.
With any type of longitudinal research, the most important thing to keep in mind is how you will track your measurements. For this kind of research to really hold up to its potential, you need to know how each pre and post score matches up. This is easy when you are measuring things like global temperatures – it all matches up to the Earth. However, for intervention studies, it can be a little more tedious. The repeated measurements need to be matched to the participants that produced the measurements. An easy way to make sure everyone is accurately tracked is to assign them an ID, but do not make the mistake of giving any duplicate IDs!
The post Snapshot or Marathon: Why to Use CrossSectional or Longitudinal Research appeared first on Statistics Solutions.
]]>The post Jump appeared first on Statistics Solutions.
]]>“A bit of advice given to a young Native American at the time of his initiation: as you do the way of life, you will see a great chasm. Jump. It is not as wide as you think.” –Joseph Campbell
Happy New Year to you all! The new year is filled with lessons and successes from 2016, and with plans, resolutions, and goalsetting for 2017. In this letter, I want to inspire you all to risk. So often we want to stay safe and yet we wonder why we haven’t changed year to year.
Risk. Risking by its very nature means doing something that is unfamiliar, challenging, something that gets your heart pumping. Viscott said of risking that it entails, “giving up false beliefs, compromised allegiances, misdirected investments, superficial attachments, and destructive habits.” It could mean being more vulnerable with others, pushing oneself physically, or even the risk of greater autonomy. Risking is to risk being your best self.
Risking also models growth for your children. Children learn that selfesteem and preparedness is necessary to risking. Being able to risk is a sign that you are living your right life. Living the right life gives you wiggle room to try new things. A person risking is able to do and say things that need to be done and said.
“When you walk to the edge of all the light you have and take the first step into the darkness of the unknown, you must believe that one of two things will happen. There will be something solid for you to stand on or you will be taught to fly.” –Patrick Overton
And so, I invite you all to have a wonderful 2017, and to risking for a better self!
Best,
James
The post Jump appeared first on Statistics Solutions.
]]>