Statistics for Tea

Statistics and Research Methods

Last set of blog comments!!

http://psuc0f.wordpress.com/2012/03/25/why-should-we-exercise-its-not-all-about-being-fat/#comment-60

http://psud30stats.wordpress.com/2011/11/24/do-qualitative-research-methods-violate-the-scientific-method/#comment-13

http://whoblogsstatistics.wordpress.com/2012/03/23/have-blogs-been-a-useful-learning-technique/#comment-52

http://psychrno.wordpress.com/2012/03/25/halo-effect/#comment-45

http://danielladzikunoo.wordpress.com/2012/03/25/the-physiology-of-false-memories-2-2/#comment-75

http://cashy3.wordpress.com/2012/03/25/popps-my-view/#comment-95

http://arbe91.wordpress.com/2012/03/25/reliability/#comment-68

 

Thank you! 🙂

Analysis of Variance

ANOVA or analysis of variance is a topic within statistics which I always get confused with, so this week I have decided to dedicate a blog just for ANOVA, so hopefully it’ll help me understand it by carrying out research and possibly help my readers understand it better if they’re confused as well!

From flipping through a selection of statistics books it appears that ANOVA is used for when you have three or more test conditions.

Basically ANOVA seems to do a similar thing to t-tests, in the sense that the sample needs to have come from a normal distribution. There are other assumptions which are important to consider when using ANOVA, which can often be solved using tests in SPSS, examples of these are Levene’s Test and Greenhouse-Geisser. It is important to consider that if a sample is small when using ANOVA the data maybe skewed, which reduces the chances of the data being normally distributed (which you need for ANOVA).

Analysis of variance can be measured in two ways, either between-groups or within-groups. I am going to look at both to create a better explanation of ANOVA.

Firstly, looking at between-groups variance, as with all ANOVA it is looking at the differences of the means. With between-groups variance specifically it literally wants to look at the difference in the means between the groups. Between-groups occurs in different situations, including treatment effects, individual difference and experimental error.

  • Treatment effects–> Is when the differences between means between the groups in big enough to reflect the original manipulation of variables in an experiment.
  • Individual differences–> Even though most experiments use random sampling there still can be the problem of individual differences. This can be explained by considering that every person is different and will react differently to experimentation. Despite random sampling you can sometimes have the situation when individuals which are similar to each other end up in one sample group, therefore the results may become unequal across the sample.
  • Experimental error–> All experiments may have an element of error, from instructions not properly read out to participants, to more major issues like error in design. However any error in an experiment can affect means and variability.

Secondly looking at within-groups variance, this is simply the variation within in the groups. (Its handy that the names for the different types of variance literally mean what they are called!) Within-group variance can occur from individual difference and experimenter error.

  • Individual differences–> Similar to the issue in between-groups, as individuals differ so much, even a random sample may end up not being truly random. Making any scores found unrealistic.
  • Experimental error–> Again explained in between-group variance.

A simple explanation to explain the process of ANOVA would be that it tries to find out whether there are differences between the means of the experimental groups (or variables), and it then looks at the differences between each groups individual means against the total mean of all the groups. Essentially what a t-test does, but for more than two experimental groups. It could be argued that if ANOVA does what t-tests do why don’t we carry out multiple t-tests? The answer to this is as follows*, by using many t-tests to test multiple groups can cause complications and can be difficult to follow, therefore using ANOVA should be a quicker and easier alternative.

The total mean of all the groups can also be referred to as the “grand mean”** this is principally the total value for each group divided by the total sample size. The grand mean is useful because it allows you to carry out further analysis of the data, including the total variation. The total variation shouldn’t be confused with variance as it is different; it allows you to compare within-groups to between-groups and vice-versa.

There are many types of ANOVA, but the two key types I will explore are one-way ANOVA and two-way ANOVA. This is the part where I usually get confused, as in I’m never quite sure which type of ANOVA test I should be running, so hopefully this will help!

A one-way ANOVA is defined*** as a statistical test which can only be carried out by numbers, which evaluates the means of two or more groups, when the groups are all using the same single variable. In other words, a test which shows if the mean for two or more groups is different or not.

A two-way ANOVA is defined****as an ANOVA which uses two variables or treatment types. You can also get ANOVA’s which are greater than two-way which use more than two variables.

I think it is also important to be able to read ANOVA on SPSS, as there are many varieties of ANOVA it is obviously important to be able to read all of them, however to prevent my blog being too long winded and for the sake of my TA(!) I’m just going to look at the SPSS output for one-way ANOVA*****.

Like most tests executed on SPSS, the fist table to appear from an ANOVA is of descriptive statistics, which as most of us know basically tells us the means and standard deviations of our data. Following the assumption that one-way ANOVA is testing similar groups, a table showing homogeneity of variances (assuming the test has been carried out to include this!) usually this would be the previously mentioned Levene’s test, which tells you if homogeneity of variance has occurred or not and the Levene’s significance level.

 Image

And finally what we’ve all been waiting for…and ANOVA table.  This tells us if there was any significance in difference between the group means. Significance is marked as “Sig.” and if it is below 0.05, then a statistical significance has been found. The table also gives us useful information about the difference in between-groups and within-groups, degrees of freedom, means square and f-values. ANOVA tables can also be worked out without using SPSS, but that is a story for another day!

Image

Phewww! A lot of information there on ANOVA’s, and I do actually feel as if I understand them better now! A final note before I conclude, I would like to mention that it is sometimes important to carry out post hoc tests after the initial ANOVA has been carried out, this measures the significance difference. And most of the time it is better to use conservative post hoc tests within science.

Right, so to conclude, ANOVA measures basically what a t-test does, but it is specifically designed to test data with more than two test conditions. 

 

References

*http://www.experiment-resources.com/anova-test.html

**http://people.richland.edu/james/lecture/m170/ch13-1wy.html

***http://www.google.co.uk/webhp?sourceid=chrome-instant&ix=sea&ie=UTF-8&ion=1#hl=en&q=one+way+anova&tbs=dfn:1&tbo=u&sa=X&ei=mpFoT6CvOZOR8gPF_o35CA&ved=0CB8QkQ4&bav=on.2,or.r_gc.r_pw.r_cp.r_qf.,cf.osb&fp=1e80c3eeca38a663&ix=sea&ion=1&biw=1517&bih=752

****http://www.ilri.org/biometrics/tools/Gloosary.htm

*****https://statistics.laerd.com/spss-tutorials/one-way-anova-using-spss-statistics-2.php

(SPSS examples sourced from above website, with second image edited by myself) 

Homework for TA, semester 2, third lot of comments :)

http://psud6e.wordpress.com/2012/03/10/what-is-average/#comment-67

http://psud2d.wordpress.com/2012/03/11/is-there-anything-that-cant-be-measured-by-psychologists/#comment-52

http://itsdefinitelycaela.wordpress.com/2012/03/04/quit-your-jibber-jabber/#comment-52

http://cashy3.wordpress.com/2012/03/11/the-placebo-effect/#comment-78

Correlational Analysis

For this week’s blog I am going to talk about the ever exciting topic of correlational analysis, which is when a measure is produced to demonstrate a relationship between variables. There are lots of topics under the umbrella of correlational analysis; firstly I am going to talk about bivariate correlations. Bivariate correlations are when you are looking at the relationship between two variables, they are correlated if they are “co-related”, as in they relate to one and other. In terms of correlation the two variables usually co-vary, meaning that as one variable does one thing then other variable usually does the other.

It is important when dealing with correlation to remember that correlation does not mean causation, just because A happens with B, B may not happen with A, therefore it is not causal. A way this could be described in real life is, people buy ice cream (Variable A) in nice weather (variable B) as in the rise in temperature has caused people to buy ice cream, however variable A (ice-cream) could not cause variable B (weather). It has been suggested that “significant” results found in correlation are often due to 3rd party variables which are not accounted for in the correlation, making the results unreliable. There are methods used to check as such the correlations, for example using scatter graphs and the statistical test called Pearson’s r.

The purpose of correlation is to find out the direction of the relationship (positive, negative or no relationship) and the strength of the relationship between the two variables. The strength of the relationship can be found out by a statistical test called correlation co-efficient. 

Positive Correlation–> Comes about when you get high scores in one variable (in this case X) and high scores on the other variable (in this case Y). As demonstrated in the image below. However you can also get positive correlation when both variables score low, as long as the results on the two variables co- relate with each other.

Image

Negative Correlation–> This is formed when high score on one variable associates with low scores in the other variable. As shown below, negative correlation tends to show a downwards pattern.

Image

Zero Correlation–> This is when there is no relationship between the two variables and it is obvious from the scattergraph as there is no clear linear pattern “up” or “down” with is familiar with correlated data. 

Image

The strength of the relationship can be determined by a previously mentioned test called “correlation coefficient” or “Pearsons r”. In terms of this test it is suggested that coefficients found from correlations which are closest to 1 (+or-) show a stronger relationship between the two variables, around 0.4/0.5 being moderate strength and weak being 0.3 or under. The correlation coefficient can be worked out either by hand or by the lovely SPSS.

So basically the correlation coefficient is used to tell you how well variables are related. So for a practical use when writing up studies for example, you would report the correlation coefficient (r) if it would be meaningful in your study and related to the rest of your work. All in all using correlation is a good way of seeing the the effect two variables have on each other and the strength to which they are related. 

Homework for TA, semester 2, second lot of blog comments :)

http://arbe91.wordpress.com/2012/02/19/analysis-of-variance-anova/#comment-47

http://danielladzikunoo.wordpress.com/2012/02/18/reading-minds-cheap-trick-or-a-scientific-ability/#comment-58

http://emcg1.wordpress.com/2012/02/18/on-the-psychiatrists-couch-with-correlation-and-causality/#comment-61

http://psuc3c.wordpress.com/2012/02/19/should-psychologists-learn-statistics/#comment-7

http://psucd2.wordpress.com/2012/02/19/placebo-effect/#comment-47

http://statstastic.wordpress.com/2012/02/19/what-makes-a-research-finding-important/#comment-54

http://psucee.wordpress.com/2012/02/19/should-psychology-be-written-for-the-layman-or-should-science-be-exclusively-for-scientists/#comment-39

Errors and Significance

Firstly this week I am going to talk about type I and type II errors, as these are a topic which I sometimes find a little puzzling, so I hope by writing about them in my blog I can confirm my knowledge and pass it on to my readers! Secondly I am going to discuss significance levels used within statistics, why we use them and the pros and cons surrounding them.

A type I error is when we reject the null hypothesis when it is actually true, as in we say there was an effect when there actually wasn’t. When the p-value (α) is 0.05 there is a one in twenty chance that we could make a type I error, as in there is a 5% chance that an experimenter could make the type I error if the null hypothesis is rejected. A way to try and stop the chances of type I error occurring is by repeating the study, as the error could have occurred due to sampling errors. If you find similar results after repeating the study, you can feel more confident of your original results.

A type II error is when we fail to reject the null hypothesis when it was actually false. So saying there isn’t an effect when actually there is, which could be potentially very dangerous to do especially in cases like drug trails etc. Factors that increase the chance of a type II error are situations where the sample size is too small or that the design of the study is poor. Design problems can lead experimenters to have ungeneralizable data which is extremely sensitive to variables.

I think it is interesting to consider why the cut off point for rejecting the null hypothesis is 0.05. If α is set at 0.2 there is a chance of a type I error in one out of five cases. Therefore if one in five times we were to reject the null when it is actually true it causes for a very liberal criterion. With this in mind, it does mean that there is less chance of a type II error. With lower probability of type II error occurring it means we are unlikely to accept the null when it is false.

However this information is different when the α is set at 0.001, in this case there is much less chance of the experimenter making a type I error. This is because there is only a one in one thousand chance that we would reject the null hypothesis when its actually true. Which initially makes 0.001 significance sound great, but all is not dandy, in this situation the chances of making a type II error. This occurs because we reject the null hypothesis less, therefore increasing the chances of not rejecting the null hypothesis when its false. Moreover there are cases where you may want to use this conservative level of significance, for example in drug testing, because if a type I error was made it would be very serious, as drugs can have extremely nasty side effects. (Gets all very confusing doesn’t it!)

So considering these two sides of the story, there need to be a balance between the chances of making a type I or type II error, obviously considering that preferably we don’t want to make either. Therefore  α at 0.05 there is a good balance.

Hope you enjoyed this week’s blogs, and got a better understanding of error and significance levels! 🙂

Homework for my TA, week 2 semester 2 :)

http://psucd2.wordpress.com/2012/02/05/why-milgram-is-ethical/#comment-34

http://itsdefinitelycaela.wordpress.com/2012/02/05/is-it-dishonest-to-remove-outliers-from-data-hwk-semester-2/#comment-26

http://cashy3.wordpress.com/2012/02/04/should-psychology-be-written-for-the-layman-or-should-science-be-exclusively-for-scientists/#comment-46

http://chocolateraisons.wordpress.com/2012/02/05/why-is-the-file-drawer-problem-a-problem/#comment-39

Case Studies

This week I am going to talk about case studies, as it relates to research methods and our module on developmental psychology. Case studies can be used to develop in depth information about an individual. These intensive studies (idiographic approach) are considered as important as the group studies (nomothetic approach) especially in clinical and developmental psychology (Allport, 1961).  The idea of this sort of study design is to obtain a description for one individual, which is then used in a psychological report.

The methods used when carrying out case studies are, observations of the individual during processes of treatments or diagnosis, or even following the lives of individuals with abnormalities which are of psychological interest. Sometimes when the study of an individual does not involve any treatments it is referred to as a case history. In developmental psychology, the investigator gathers rich detailed information about the individual and then tests a developmental hypothesis by analysing the events of the individual’s life. Studies like these are often longitudinal and time consuming, but do allow the investigator to find out more information than you would for a standard treatment and effect style experiment.

In a developmental psychology case studies information is found out about the family background, school/work history and health records of the participant. However there are issues surrounding case studies, for example the studies are not generalisable to the rest of the population, and they even can’t be compared between two individuals as it is likely they will have been observed in different situations and asked different questions. As case studies are not generalisable, any conclusions made from the investigation would not apply to most people.

Even though group based experiments may find a more direct route to the answers puzzling psychologists, they also over look individuals themselves, by putting everyone into one great big average.  Case studies can be used to support results found within group studies, therefore offering more insights into the psychological world. For examples some of the theories by Piaget were largely based on the individual observations of his own children.

One of the main advantages of case studies is that they include such vast detailing, which can’t always be found in alternative types of studies. Case studies also include details which are removed from group experimental designs, such as confounding variables and unexpected events. By including all events witnessed during the case study, new variables may be found to explain results; moreover experimental hypotheses can be formed from new results found from the case studies.

They also give investigators a chance to study phenomenon that isn’t usually studied, which can then be used as constructive criticism against previous theories. Allowing previous theories to expand, and new research questions to be formed. Case studies can also be very persuasive and powerful, as they relate with real life situations, and therefore some people will be more likely to believe them over group studies, as they are found to be more emotive. As a result of triggering more emotions in people, case studies are often remembered more versus traditional experimental procedures.

However there are also limitations to case studies, one of these being that they only simply describe certain events and behaviours rather than attempt to explain why something has occurred. Therefore case studies lack internal validity. They can also be lacking external validity, as the case study only represents one person and one situation, results are likely not to be the same for different individuals, therefore results cannot be transferred out of the barriers of that one situation. It should also be considered that there is a certain level of bias within case studies, as it is at the interpreter’s discression how the results are analysed. Results may also become distorted by the interpreters own view, making them even more bias.

On a final note, case studies have many pros and cons, mainly depending on what sort of information you are hoping to find. Hope you enjoy my first blog of semester 2! 🙂

Comments I’ve made for my TA on week 11 :)

http://psucd2.wordpress.com/2011/12/07/the-importance-of-05-level-of-significance/#comment-23

http://psud2d.wordpress.com/2011/11/25/is-qualitative-research-as-scientific-as-quantitative-methods/#comment-33

http://prpljr.wordpress.com/2011/12/02/ethics-are-they-necessary-or-have-we-outgrown-them-as-a-society/#comment-36

http://emu2468.wordpress.com/2011/11/25/is-it-okay-to-use-computers-to-transcribe-interviews/#comment-41

Observational Methods

Following from my previous blog on qualitative research methods, this week I am going to talk about observational methods, which involves directly recording (observing) a participants behaviour as it occurs. It has been said that observation requires a unconstrained environment to enhance the chances of getting natural behaviour form a participant or groups of participants.
Observational methods are often used to avoid the problems which alternative experimental methods experience, problems such as demand characteristics and the dehumanisation of participants. It could be argued that all psychological experiments involve observation; however there are specific study types which are used for observations. When a study is described as observational in psychology, it means that a researcher is directly watching a person’s normal behaviour.

Observations can be carried out in two different techniques.  Firstly in traditional style experimental designs. An example of this could be the Bandura (1965) study of the children’s aggression levels after watching a video clip of adults behaving aggressively. The children were then observed by Bandura of their behaviour on different types of equipment, including the Bobo doll. This was all observed through a one way mirror, so the children did not know that they were being observed.  Bandura’s study was considered observational, because he observed the children’s behaviour; however it was in a laboratory environment, so it wasn’t strictly 100% natural observation, as the children would not have been conducting behaviour onto said Bobo doll usually!

The second sort of observational technique is that of a field experiment, where the research is carried out in a natural setting.   However there are many ethical issues surrounding such experiments, such as is it ethical to watch people when they are unaware that you are doing so? Moreover by using a field experiment you have less control over extraneous variables, and therefore the study is less replicable, and less reliable.

Observations can be used when the experimenter has no control over any of the independent variables, which is when only the natural behaviour is being observed.

The strengths of observational experiments are that, they produce very natural results. They enable researchers to get results about individuals without having the bias which self-reports have. They also give data which has been collected in genuine natural settings, assuring participants natural behaviours.

However there are a few weaknesses to this type of research technique. If a participant know they are being watched, they are liable to change their behaviour, adding demand characteristics to the research.  It is a time consuming way of collecting research, if compared to other research techniques like questionnaires. The worst issues with observational research is that it is not always easy to establish cause and effect, so even know behaviours maybe observed, there is no way of explained why they occurred.

As a final point, observational research methods are best used when a researcher is looking to find participants natural behaviours.  To help conquer the issues of observational studies, using lab experiments where certain variables are controlled, therefore allowing the researcher to have a certain understanding of cause and effect.

🙂

Create your own website, free on WordPress.com.Create your own website.
Start free