Monday, November 30, 2020

The First Month #NoRegrets

This day last month was my last official working day at the National College of Ireland - I can't believe how fast the time has flown, and how easily I am fitting into retirement. Of course, all my plans for travel have had to be postponed for another day, and the Lockdown has reduced options for things to do. 

The best part of retirement so far is that I can now visit my parents Joe and Phil during the week as well as at the weekend. I have been stopped several times by the very polite GardaĆ­ on the road to Carnew in Co Wicklow asking me the purpose of my journey. I usually have some dinner for my parents on the passenger seat to show them, and they let me continue. We have plenty of chat and we have not yet exhausted all topics.

Have I missed my work at NCI? Yes. I do miss teaching, but I am 100% satisfied at my decision to retire. I am so glad that I am not stuck at my computer working from home all day long - this must be hard for everyone. While I feel that on-line classes are here to stay, there's nothing like the classroom experience on a good teaching day. 

At first, in the days after my retirement day there were so many on-line messages congratulating me on my retirement - I was overwhelmed! There were over 800 "Reactions" and 100 comments to NCI's lovely message on LinkedIn. I loved getting messages from past students and from people I had worked with in the past. This has now all dried up - the party is definitely over!

As posted elsewhere on this blog, I have been making some new YouTube videos to keep me busy. And plans for a series of videos on R Programming are well advanced - I hope to start releasing these videos in mid-January. I have also done a small piece of work for another third-level College, which I found really satisfying - it was good to feel useful again.

So far - retirement is good!

Monday, November 23, 2020

Lecturer in Computing Vacancies at NCI

National College of Ireland has just advertised for FIVE permanent positions for Lecturers in Computing - see NCI Vacancies here. The areas of specialty sought are: "data science, cybersecurity, financial technologies, computing, business computing, cloud computing, (and) artificial intelligence". If you are an aspiring Lecturer, NCI is a great place to work and you should consider applying. Please read the "Qualifications & Experience Required" - there are a lot. The salary is quite low starting off, and it will take a few years to rise up the salary scales. You won't be doing this job to get rich.

NCI - A Great Place to Work.
Image source: The Irish Times.

One thing that does strike me about the job advertisement is that I would not be qualified to apply! I do not have a "PhD in Computing, Informatics, Computer Science, Data Science, Cybersecurity, Financial Technologies, Business Computing/technologies or related areas". Nor do I have an "Established track record of research achievement as evidenced by scientific publications in peer-reviewed journals". I have absolutely no "Evidence of research funding". It's sobering to think that I would not have qualified to apply for my old job.

Nevertheless, applicants should not be put off by this. It's rare that a candidate ticks all the boxes on any job description. Passion and commitment to education are the primary characteristics needed. Good luck to all who apply!

Thursday, November 19, 2020

Return to Whole Blood Donation #GiveBlood

I have been donating blood to the Irish Blood Transfusion Service since I was about 18 years old, but for the last several years I had been doing so at the Blood Platelet Clinic in St James' Hospital. Platelet donation is very different from whole blood donation - it can be done much more frequently. However, earlier this year I was retired from the Platelet Panel due to several consecutive low platelet count levels. So, after 155 donations, today was my first day back at the whole blood clinic in Stillorgan. Lots of Covid-19 restrictions of course, but nevertheless a pleasant experience.

If you can, give blood.

Friday, November 13, 2020

Four New "How To..." YouTube Videos

Things to do during retirement? Make videos!

Today I posted four new statistics videos to YouTube - I have never done this in one single day before. These are about using SPSS for non-parametric tests (Mann-Whitney, Wilcoxon Signed Rank, and Kruskal-Wallis), and one for how to determine if a data set is normal or not. While I used to cover these exact same tests in class, I never made the effort to make videos about how to do them in SPSS. 

I have now posted 11 new videos in the past two weeks. They will take a while to attract new views, but as of today they have received almost 1,000 views between them already. I'm not expecting too much, but they do fill some gaps in what I used to cover in class. 

I think that's it for now for Statistics videos, hopefully viewers will dine them useful.




Wednesday, November 11, 2020

How To... Determine Cronbach's Alpha in Excel and SPSS #Statistics

Cronbach's Alpha is a test statistic to measure internal consistency and reliability. I never covered it in any of my statistics classes as the datasets I used did not suit this as a test. It seems to be very suitable for data like survey responses and to determine if questions on the survey correlate with each other strongly enough so that you can conclude they all measure the same thing (the subject of the survey). If, for example, you  were conducting a survey on attitudes to Medical Insurance, you would want to make sure that your survey questions measure the same thing (attitude). You would expect respondents who give a high rating to the statement "I like the quality of my Medical Insurance", to also rate the statement "My Medical Insurance is good value for money" highly. 

Excel does not have an option to calculate Cronbach's Alpha, so a workaround is necessary. This is a very easy test to do in SPSS - here's how to do both:


Friday, November 06, 2020

Two More YouTube Videos on Effect Size

I never really concentrated on Effect Size while teaching statistics at the National College of Ireland - I just showed a formula for calculating it for a two-sample test, didn't really explain it that well, and moved on. I should have placed more emphasis on it, but now I have posted videos for one-sample, two-samples, and ANOVA tests.

When you find a significant difference in Statistics, another way of putting it is that you have found an "effect". But statistical tests do not tell you if this is a small, medium, or large effect. I have concluded this short series of four videos on Effect Size with the final two about ANOVA tests - this is an easy test to do and is worth following up when conducting the ANOVA test.


Tuesday, November 03, 2020

How To... Calculate Effect Size #YouTube #Statistics

Following on from yesterday's publication of a video showing how to calculate Effect Size for a one-sample test, today I have added a new video showing how to calculate Effect Size for a two-sample test.

When you find a difference (in other words, an "effect") with a statistics test, it is often important to know how meaningful that difference or effect is. For example, if you were conducting a test to see if a drug had an effect on the treatment of a disease, you would like to find a meaningful difference rather than a trivial one.

This series of "How To By Hand" videos are being made at home with my GoPro Hero 7. I am using it like a document reader and it gives good quality, but is awkward to set up. I have to use my old SCUBA diving lead weights to anchor the GoPro stand on my desk. Once recorded, I use the GoPro Quik App to transfer it to my computer for upload to YouTube. As usual, I do not edit the videos - if I make a mistake, I simply start over.

Here is today's video:

Monday, November 02, 2020

How To... Calculate Cohen's d for Effect Size #YouYube #Statistics

On my first "working" day of retirement I have made a new statistics video - it is over six months since I last published a video on YouTube. This is one of a few statistics videos that I wanted to do to plug a few gaps in the series that I have already produced. Statistics videos are now my most popular, so I'm hoping that this one will prove popular too.

Cohen's d measures effect size. This tells us how meaningful a difference is when we find a significant difference as a result of a statistics test. Usually we report a difference as p < 0.05, but this does not tell us if the difference is small, medium, or large - so it is important to calculate the effect size.

Here's how it's done for a one sample z test (it can also be used for a one sample t test):

Sunday, November 01, 2020

Kevin Barry and "The Mons" #100years

Image source: Wikipedia.

One hundred years ago today, Kevin Barry was hanged in Mountjoy Prison for the murder of a young English soldier. He was the first of 24 republicans to be executed by the British during Ireland's War of Independence. He was accompanied to the gallows by two priests: Canon John Waters, and a Fr McMahon who gave him all the spiritual comfort that a condemned person would want.

There is a story in our family, which I got from my Uncle Pat, that my paternal grand-uncle Monsignor Charles Francis Hurley (my second name is also Francis in his honour) also attended the execution and that it greatly affected him for the rest of his life. Unfortunately, he is not mentioned on any of the documents in the Kevin Barry Papers archive in UCD, nor does a Google search find any connection between him and Barry.

Affectionately know as "The Mons" in our family, he attended Cistercian College Roscrea (1908 - 1912) was ordained as a priest on the 28th February, 1920 in the Irish College in Rome. One of his first postings was to St Patrick's College in Carlow - at the time it was operating principally as a seminary for the priesthood. So he was a very young priest when Barry was hanged nine months later. Kevin Barry grew up in nearby Tombeagh and went to school in Hacketstown, Co. Carlow. I don't know if The Mons knew the Barrys personally or how he was connected other than he was working in Carlow where Barry was from in 1920. It is also thought by some in my family that at the time of the execution that he had moved to the Holy Cross College on Clonliffe Road in Dublin - this is where Canon Waters was based. It is possible that at the execution he assisted Frs Waters and McMahon, carrying holy water and oils for anointment. Clonliffe College is about a mile from Mountjoy Prison. Perhaps he was outside the prison where hundreds of people were praying for Barry on the morning of his execution - we might never know what exact part my grand-uncle played on that fateful day exactly one hundred years ago.

PS - if any family are reading this and can add more to this story, please get in touch!