The results are in

You heard right, I have results.

I’ve managed to overcome Stata – I wouldn’t say were close friends yet, but we definitely are on better terms compared to a week ago. The data crunching is done, for now anyway. Sure there’s more to explore, but I’m ready to present my preliminary findings to the world – starting at the University of Wollongong.

Good timing, as the presentation is tomorrow! So I guess I’ll better get back to work and tell you all about it when it’s done – I don’t want to give anything away now and ruin everyone’s fun tomorrow.

Stay reading to find out more soon…

Red again!?!

Red… I’ve been seeing so much red! And all I wanted is some black

images (7)

Ok, this is not as serious as it sounds – I am not bankrupt.

Luckily (I guess), the red is not printed on my bank statement but rather appears whenever I enter commands in Stata. This is very frustrating but, I suppose, a way to learn… a very slow, painful way.

This week I’ve started conducting some serious analyses of my data set, HILDA. So far I have only opened it, checked I have a large enough pool of observations for each variable and run a few very simple, fun regressions. Now it’s time… time to delve into the more difficult domains.

First hurdle: merge that data. HILDA consists of 10 waves I need to combine into one big data set in order to analyse it. Sounds like it should be easy? Wrong – at least in my case. My experience with Stata so far has been restricted to solving specific questions. Merging data has never been a task, as data sets usually come nicely prepared for classroom exercises.

I have consulted one of the lovely academics at university, and she was so kind to spend a few minutes explaining the process. I now have the conceptual knowledge to reshape my data set, but that posed only half of the challenge (maybe not even half). Stata will only accept the correct commands! It’s very picky like that.

y-u-no-meme-generator-stata-y-u-no-do-what-i-tell-you-to-4010c5

Researching commands on the internet has hardly been helpful, as I do not speak fluent programmer (yet). It has, however, contributed to my overall understanding of the desired outcome and the theoretical way to achieve it.

Now Stata needs to comply. My first merging attempts are turning fertile – I created at least some black letters and numbers yesterday and am pretty certain I can work it out.

The next step will be to estimate my fixed effects model. This I am more confident about, as I gained some knowledge on fixed effects models earlier in the semester as part of my advanced econometrics course. Go uni skills!

I will also need to graph satisfaction trajectories, which will illustrate the change in happiness over time caused by an event. Again, theoretically I have an idea how this can be done. I guess, I’ll have to wait and see whether Stata thinks I’m right…

Two weeks to go – will I be fine? I’m saying yes. There is much to accomplish, but I’m tough. I can do it! Once these results are out I can compile my final report and work on my presentation. Also, hello to the Associate Dean of Research, who will be attending our presentations; I promise I’ll try to make it interesting! And thanks to Martin for making it just this little bit more scary…

Memes by www.quickmeme.com, diylol.com

Lit’s talk lit reviews

Hey there, it’s me, Lena… I’m back! You’ve probably been wondering what I was up to, weren’t you? I am afraid to tell you, it was nothing that justifies not writing for two weeks or so.

So, what has occupied me? Well, there was that whole frantic week of getting the research proposal together and delivering a convincing presentation (plus exams and illness, mind you). So that’s been conquered but, since I am officially a German/Australian and not the other way around – I’m not a ‘Slashie-Award‘ laureate yet, in case you’re wondering – that was just the beginning to my maddening schedule.

Exams, a flu, a case of man flu, a business conference, a UNSW meeting, another exam, a 3-day PASS conference, a trip to Moss Vale, a trip to Sydney and an application for Club of the Year later, here we are… I finally have time to tell you more about this thing called hedonic adaptation.

So, my research proposal and presentation went well. People seemed not to object too much, which is good to hear (or maybe they just gave up on me). Since then, I have to confess, not much work has been done on my project. But hey, you just heard about my schedule, right? I have not exactly been flush with spare time… The plan is to get back into it as soon as my last exam (today!!!) is out of the way.

In the meantime, though, let’s talk literature review. I have just written a preliminary version of mine and I am glad it’s done, at least for now. Literature reviews are tedious, let me tell you. So why do we even need them?

3sfr1b

Well, I guess the big argument in favour of a lit review (so it is lovingly known to its friends) is justification for your research. You want to let people know what others have done and then indicate where they went wrong, or what they have missed – that’s where you, knight in shining armour, step in. You explain why your research is relevant and why we need you to conduct it – you may even want to attract funding at this stage, which makes the whole lit review process even more important.

Evil tongues, however, may argue that a lit review is a waste of time. If a researcher really wanted to, she could easily fool everyone and conduct a study that had been done before. Careful choice is the key – as long as you don’t pick a well-known, oft-quoted paper, you probably succeed pretending it is your own idea. If you’re smart about it and don’t reference any articles mentioning your chosen study, and check it doesn’t appear as the first result when searching Google Scholar, you probably will be fine. Don’t take this as advice or as a confession, of course I’m not condoning such action. Just saying it could be done…

In my case, writing a lit review was aiding to inform myself about my chosen research area – what had been done, which data sets had been used previously, was my idea feasible and, equally important, was it still novel? My lit review actually helped me acquire an extensive background on hedonic adaptation, which, I think, is very valuable. You really want to know what you’re talking about when telling others about your research – I’m definitely not a fan of being exposed in my lack of knowledge, especially when it comes to my chosen topic.

images (5)

Having said that, I guess it wasn’t necessary to write a lit review – it was merely important for me to know my research area and be familiar with previous studies. So what’s the point in writing it? To signal to others that I can do this? To justify my research? But then we end up back at the first argument against literature reviews…

Another reason for writing a lit review (one I benefited from a lot) is to get an idea of how to conduct your study – that is, how to set it up, which program to use, which model to use, etc. I was lucky enough to stumble upon a paper which almost exactly addresses my research questions, but uses different data (the SOEP, which is basically the German big sister of HILDA) and different variables.

This study, by Uglanova et all, uses an entity fixed effects model to measure hedonic adaptation, looking at marriage, child birth, unemployment, widowhood and divorce – all those non-random variables I was telling you about earlier. Just as I plan to, this study measures the initial effect of those events on life satisfaction and then estimates the trajectory of each, modeling the return to baseline satisfaction. This paper basically told me all I need to know – which model I should use, which control variables I need, etc..The only aspect it can’t assist me with is the actual creation of the model – but I’m sure I can work that out.

For all of you who do not wish to read my lit review, here’s my reason for conducting this research. As I said in earlier posts, my criticism on Uglanova’s paper is the non-randomness of the variables used to estimate a change in satisfaction. Since they are anticipated, the effect of an event is diminished – our estimated coefficients are biased. I therefore will only regard random events in my study, foregoing this source of error and hopefully contributing a small piece to the happiness-research-world.

So, to answer the question posed earlier: Is a lit review worth writing? All in all, I believe that mine has brought me closer to my area of interest. Sure, it was tough at times and I really did not feel like reading that one-millionth article on happiness research, but I do think I gained a lot of knowledge from it. And hey, I’m sure these newly developed lit review-writing-skills will come in handy during my Honours year!

1download

Images by 
alluglyboysdo.wordpress.com
www.quickmeme.com
memegenerator.net

The Evolution of Lena

The work on my research proposal has catapulted me from being a lower life-form of research to an advanced version of myself. Today, I want to share my evolution with you.

Here you can see a naive, small-town-research Lena, who has just been informed that she will have to conduct a study – contributing to existing literature – within the next 13 weeks or so:

Initial FormShe has an idea about the broad areas which interest her, and she has started thinking about concrete topics. She is tossing up between using existing data, like the HILDA survey, or conducting her own experiment, which would be fun but a lot of work for this short amount of time. More research into different ideas is needed to arrive at a topic she likes and is confident about. Her research skills need refinement, however – this will be the first project she will conduct without any guidance as to what to study or which literature to look at.

The next few weeks of my life were quite stressful and challenging. Not only did I have to decide on a topic I liked, I also had to identify a niche not studied before. Daunting and kind of scary – how was I to discover a topic not researched thoroughly before, when there are many highly skilled academics doing exactly this, and probably doing it better?

My interest in behavioural economics led me to look into concepts on hyperbolic discounting and, yes, hedonic adaptation. After having talked to one of my lecturers, it seemed clear that I could use HILDA data to conduct a study on hedonic adaptation – now I just needed to verify others had not done so before. Of course they had, however, no one used random effects and HILDA before.

My research started to take shape – I had a topic – and I felt like I was evolving. After reading many, many journal articles and familiarising myself with HILDA, I finally reached the next step of research existence!

Current EvolutionLena’s understanding of her chosen topic has definitely grown stronger – especially after having written the preliminary literature review. She also decided on a methodology, which she is very excited about. Overall, her research has become much more focused – she rid herself of all the other ideas whirring around in her head and concentrated fully on hedonic adaptation.

This evolution has come at a price – sleepless nights and longing gazes out of windows on weekends while sitting in front of a computer seem to be the norm at the moment. To add to that, I have recently been attacked by the flu and recovery is slow. On the plus side: look at me now! I feel I have gained a lot of human capital within the last few weeks, thanks to this research project.

Where will this take me, I can hear you ask. Well, I guess we’ll have to wait and see. The next step is to focus on my methodology and actually estimate a fixed effects model using HILDA data. Quite the challenge, as this will be the first time I am doing anything like this. Extrapolating from my recent experience with being thrown into the deep research end, I am sure it will be challenging, but highly rewarding.

Want to catch a glimpse at the final form of a Lena. Here it is:

Final Form

Stay reading to discover what her powers will be…




Pictures by: Deviantart and Nintendo

Bias confirmed

While writing the ‘expected outcomes’ section of my research proposal, I thought:

Hey, wow, this is what I will find? Awesome! I’m excited – bring on the results!

Hang on… what? I don’t know I’ll find this – that’s why it’s called ‘expected outcomes’, not ‘certain outcomes’!

Then the alarm bells started ringing in my head and speakers were screaming: CONFIRMATION BIAS – IT’S A BAD CASE! NO ONE IS TO GO NEAR HER!

You’ve probably heard about this thing called confirmation bias – even though I’m sure you don’t suffer from it. Defined as the tendency to favour information that confirms our beliefs or hypotheses, understanding it is crucial to all researchers out there. I mean, no one starts a project thinking: “Oh, I’m really not sure at all what I want to prove with my research – I’m just pottering about trying to find some sort of a result”. No, it is more like: “I want to prove this, and will succeed!”. Ok, maybe that’s a bit harsh. I’d like to think that if a desired result just isn’t there, a researcher would abandon her study ship. Right?

Why is confirmation bias a bad thing? Well, it distorts reality. Say many clever people are working on solving a problem – maybe one that is too small to even bother with, or worse, one of which the solution is the antithesis of the proposition to be proven. Due to confirmation bias, our clever people ignore the evidence against their theory and instead plough on until they finally get results in their favour. Now they propose new policies based on their findings to let all people enjoy the benefits of this amazing, new (but unfortunately incorrect) theory. I’m guessing you get the point.

So, what is to be done about this threat to research?

To have a goal in mind is important, I believe. Pottering about is not the right way to conduct a study – if it wasn’t for the very specific and focused research of Ts’ai Lun and Nikola Tesla you would be reading this blog from a slab of slate by candle light. No, it has to be solved in another way.

Think… Ah yes, think! And then re-think!

Maybe that’s a step towards ridding ourselves of confirmation bias – reevaluating our results. And I am talking about the actual results – not what you hope to find. Look at the data and conduct your analysis as though you’ve forgotten what’s written in your ‘expected outcomes’ section.

Easier said than done, I know. But if everyone works a little towards this unbiased, utopian research world, maybe – just maybe – we can stop furrowing our brows, shaking our heads and running to our desks to replicate a study’s results… and instead feel more confident that the researcher has factored in her confirmation bias. This one definitely has…

Visiting Dr Freud

This morning in the therapists’s office:

Freud: Ah, Lena, good morning. Wie geht es dir?

Lena: Hello doctor! I’m ok, everything seems to be just fine. I don’t trust it…

Freud: Well, take a seat on the couch, get comfortable and tell me about it. You are talking about your research project, I assume?

Make yourself comfortable

Lena: Er, sure… I mean, exactly.

Freud: So, what progress have you made?

Lena: Well, I arrived at specific research questions – questions not answered before, at least not using HILDA. I was scared to discover someone else – probably smarter, prettier and more successful than I –  had conducted a similar study before, risking my project’s relevance. Since that’s not the case, it will be mine forever and I will never let it go!

Freud: Tell me, what is your relationship to your father?

Lena: What has that got to do with it?

Freud: Oh, nothing much – just a side-project I’m working on… forget about it.

Lena: Right, ok. Well, going back to my research: it seems so far things have been running smoothly. I decided on a topic quickly and was lucky to discover no one had conducted my proposed study before, but I fear I’m arriving at the real challenge now.

Freud: Which is?

Lena: Well, the actual project… I’ve acquired background knowledge, I have HILDA, I know what I want to measure and prove… there’s nothing left but to finally start! That is scary – I’ve never done a research project like this.

Freud: How are you going to attack it?

Lena: In my usual German way – Augen zu und durch… I’ll draw on those teutonic skills of determination, workophilia and efficiency. From experience, most problems can be solved via this approach – and should be. Only problem: there are just 24 hours in a day… with mid-semester exams, conferences and other life-changing events coming up, that is not much time.

Freud: Are you worried?

Lena: A little – my sleeping pattern is definitely affected by this. The time management problem should be solvable – I may have to make some lists and calendars and forfeit weekends, but it’s do-able.

Freud: Well, sounds like you’ve got about a 50/50 chance there! We’ll leave it at that for now – don’t stay away for so long again. You certainly need these sessions…