Predicting people's mood after watching a movie. Need help:)

[Quicky]

Ah, I was remembering my mood. Naughty, probably. I'll wait till I watch something on my laptop.

Ah, yes... that's a bit of a trade-off I have to make unfortunately. Remembering how you felt is inherently very biased, so from a psychological standpoint it is better to limit that as much as possible (as I do now with the time constraints). If you go to this Wikipedia-page you'll see dozens of different kinds of cognitive biases related to memory.

It's annoying though, because I realize that quite a few people (you're not the first) take the time and effort to help me, and then are basically told that they did it wrong . I've contemplated lowering the time limit, but in all honesty I would be cheating myself in order to gather data that is unreliable.

As for the project, yesterday I wrote a first post on the Facebook page about the reasons for doing this project.

A quote:

Why am I doing this research, you might wonder. Why do I want to know if mood after watching a movie is predictable?

There are two reasons:
1) I want to improve and enrich people's movie watching experience.
2) I find translating subjective concepts into numbers that you can calculate with immensely fascinating, especially when people suspect it can't be done. All the more motivation for me to try it.

The latter reason is just a personality trait of mine, but I'm going to explain the former reason.

[...]

Go to the Facebook page to read the rest .

[ShogunRua]

I don't understand the point of this project.

For starters, the biggest regulator of mood after watching a film would simply be its quality. I'm annoyed after watching a bad film, and happy after seeing a good one, regardless of what genre either is in.

Thus, the biggest variable is also the easiest and most commonly measured already. The stuff you're looking at is all secondary and tertiary.

But okay, let's say that someone with pre-mood X has to decide between films A and B which they both rate the same. Firstly, how does one accurately gauge a mood? It's usually more complicated than simply "angry", "sad", or "happy". In fact, most humans don't attain such extremes of emotion on a regular basis.

Secondly, and more importantly, EVEN IF you draw some tenuous connection between the genre of some movie and what pre-mood works best with it for a GENERAL POPULATION, it's completely meaningless for the INDIVIDUAL.

This is a very important statistical point, and it directly contradicts what you wrote in point 1 of your post above.

For instance, the example you use on your site is that a person with a frustrating day doesn't want to watch "Schindler's List". Actually, when I am feeling down, I often like to watch depressing films, which make me happier about the problems plaguing my own life.

Thus, for me, a general observation like yours, even if it's true, would be entirely meaningless. And not just me, but every other individual.

Thirdly, your project ignores variety in terms of movie selection. I'm sure that something like "comedy" probably works best for the greatest variety of pre-moods. So what?

After watching 3-4 comedies in a row, someone will want a change of pace.

So basically, I'm wondering again what the point of this is? It doesn't appear to do what you claims it does. (allow people to make more efficient movie-watching choices)

[Quicky]

I don't understand the point of this project.

You have some valid points, but I think you don't realize is that this project is just a tiny piece in a much larger puzzle. You cannot solve the larger puzzle without investigating the individual pieces, especially not in a graduation project like mine.

For starters, the biggest regulator of mood after watching a film would simply be its quality. I'm annoyed after watching a bad film, and happy after seeing a good one, regardless of what genre either is in.

That is your opinion, not a scientific fact. There is very little scientific literature about the effect of movies on people's mood, but in the limited data I have collected already thus far, it is quite clear that the correlation between mood and genre seems stronger than the correlation between mood and rating. I don't think you can deny that most people will feel very different after watching Black Swan than after The Artist, even though they might love both films (as I do myself).

Thus, the biggest variable is also the easiest and most commonly measured already. The stuff you're looking at is all secondary and tertiary.

Again, that is your opinion. Obviously I recognize that appreciation for a movie has a big influence. If you hated a boring comedy you just watched, you won't feel happy. That is the reason why I ask people in my research both an expected rating before the movie and a rating after the movie, to have information about how much they liked the movie and how much they like it more than they expected. You might like a film, but have expected much more which could mean it would have a stronger influence on your mood than a film you liked so-so, but of which you expected it to be so-so.

But okay, let's say that someone with pre-mood X has to decide between films A and B which they both rate the same. Firstly, how does one accurately gauge a mood? It's usually more complicated than simply "angry", "sad", or "happy". In fact, most humans don't attain such extremes of emotion on a regular basis.

I don't know if you've had a look at my web application, but at no point will it ask you to label your mood. It definitely *is* more complicated than simply "angry", "sad", or "happy". That is why I'm using the AffectButton, a scientifically validated tool to measure emotion/mood, which allows people to represent their mood in a dynamic smiley rather than discrete labels.

Secondly, and more importantly, EVEN IF you draw some tenuous connection between the genre of some movie and what pre-mood works best with it for a GENERAL POPULATION, it's completely meaningless for the INDIVIDUAL. This is a very important statistical point, and it directly contradicts what you wrote in point 1 of your post above.

For instance, the example you use on your site is that a person with a frustrating day doesn't want to watch "Schindler's List". Actually, when I am feeling down, I often like to watch depressing films, which make me happier about the problems plaguing my own life.

Thus, for me, a general observation like yours, even if it's true, would be entirely meaningless. And not just me, but every other individual.

This is your best point and I mostly agree. Mood and influences on mood will definitely vary a lot from individual to individual. However, I'm convinced that there are also general trends, which I hope to uncover with this research. The reason I'm looking for general trends, rather than individual influences is simple: limited resources. I'm doing a graduation project which I'm paying out of my own pocket and in which I have to rely on the goodwill and effort of people like you. I don't work for a large research institution or a big entertainment website like Netflix or IMDb. In other words, I don't have the resources to gather the thousands and thousands of datapoints that you would need to uncover and model individual influences on mood. In my application I do ask for some general demographic information which might make it possible to separate trends in for example men vs. women, but I don't have the means to gather enough data to model individual trends.

You may wonder now: "Why do it at all then?". Again, the answer is quite simple. I do it anyway, because I believe that if I can show that it is possible to uncover trends in influence of mood on movies in a general population with very limited amounts of data, and despite all the external influences on mood that you have no control over, then it is very likely that if you had access to thousands or tens of thousands of data points, you could uncover individual trends or trends in small groups of people (similar to groups of people here on Criticker with very low TCI's). And that is what I would eventually want to uncover. Who knows, maybe it might convince a company like Netflix who has already invested millions in their recommendation engine to explore mood-based recommendations as well.

Thirdly, your project ignores variety in terms of movie selection. I'm sure that something like "comedy" probably works best for the greatest variety of pre-moods. So what?
After watching 3-4 comedies in a row, someone will want a change of pace.

Sure, but how often will people watch 3-4 comedies in a row in a short time span? I'm aiming for the general population here, not for a movie buff who watches two movies per day and is checking off the TSPDT list.

So basically, I'm wondering again what the point of this is? It doesn't appear to do what you claims it does. (allow people to make more efficient movie-watching choices)

At no point do I claim that my project on itself will allow people to make more efficient movie-watching choices, and if I did, I apologize, as that is not what the project in itself will do. I do think it is a valuable first step in a much larger project to investigate the role mood plays in movie-watching and how it can be used to improve people's movie watching experience.

[ShogunRua]

I don't understand the point of this project.

You have some valid points, but I think you don't realize is that this project is just a tiny piece in a much larger puzzle. You cannot solve the larger puzzle without investigating the individual pieces, especially not in a graduation project like mine.

I don't see how this is related to anything I wrote above.

For starters, the biggest regulator of mood after watching a film would simply be its quality. I'm annoyed after watching a bad film, and happy after seeing a good one, regardless of what genre either is in.

That is your opinion, not a scientific fact. There is very little scientific literature about the effect of movies on people's mood,

There are no "scientific facts" at all when it comes to a study like this. However, rating being the biggest predictor of mood is not just "my opinion", it's common sense.

If people rated a film highly, then that typically means they enjoyed it. Ergo, we can expect an improvement in their mood.

If you have some startling data that suggests this is false (that people's moods remain the same or even worsen from films they really like), I would like to see it.

but in the limited data I have collected already thus far, it is quite clear that the correlation between mood and genre seems stronger than the correlation between mood and rating. I don't think you can deny that most people will feel very different after watching Black Swan than after The Artist, even though they might love both films (as I do myself).

You know why this is?

It's because of the narrow and imprecise selection of moods in your project. (I will write more about this below)

Human beings can simultaneously feel several parts happy and several parts melancholy, for instance.

I don't know if you've had a look at my web application, but at no point will it ask you to label your mood. It definitely *is* more complicated than simply "angry", "sad", or "happy". That is why I'm using the AffectButton, a scientifically validated tool to measure emotion/mood, which allows people to represent their mood in a dynamic smiley rather than discrete labels.

"Scientifically validated" by whom, exactly? Using a smilie face to describe my internal emotional state feels absurd, and I wouldn't even know how to do it accurately. What the hell do some of those facial expressions mean?

The only interpretation I can gather is that some are more "angry", others "sad", others "happy", and still others "neutral". It seems like a very convoluted method that is still based on those same discrete labels, and is a piss-poor way of indicating one's emotions.

Seriously; since when has a smilie face picture ever accurately described anyone's mood better than those same discrete labels? (And yes, the latter is imprecise and limited, too)

Apparently, I am not the only person that feels this way, looking through other comments.

This is your best point and I mostly agree. Mood and influences on mood will definitely vary a lot from individual to individual. However, I'm convinced that there are also general trends, which I hope to uncover with this research.

Like I wrote before, general trends in such projects usually don't improve specific recommendations for individuals.

For instance, the statement "60% of boys prefer playing basketball to baseball" does not tell a specific boy which of those he would rather play himself.

The reason I'm looking for general trends, rather than individual influences is simple: limited resources. I'm doing a graduation project which I'm paying out of my own pocket and in which I have to rely on the goodwill and effort of people like you.

Why are you paying out of your own pocket? Does your university not have research grants, fellowships, and aren't you being paid yearly by them?

I don't work for a large research institution or a big entertainment website like Netflix or IMDb. In other words, I don't have the resources to gather the thousands and thousands of datapoints that you would need to uncover and model individual influences on mood.

You kind of missed the point. All the resources in the world wouldn't allow you to do this.

Thirdly, your project ignores variety in terms of movie selection. I'm sure that something like "comedy" probably works best for the greatest variety of pre-moods. So what?
After watching 3-4 comedies in a row, someone will want a change of pace.

Sure, but how often will people watch 3-4 comedies in a row in a short time span? I'm aiming for the general population here, not for a movie buff who watches two movies per day and is checking off the TSPDT list.

What exactly does being a "movie buff" have to do with my comment, exactly?

I can watch 3-4 comedies in a row over the span of 2 whole months, and I might still be burnt out on them. Obviously, fatigue with a certain genre is important, yet you see how convoluted its implementation becomes?

At no point do I claim that my project on itself will allow people to make more efficient movie-watching choices, and if I did, I apologize, as that is not what the project in itself will do. I do think it is a valuable first step in a much larger project to investigate the role mood plays in movie-watching and how it can be used to improve people's movie watching experience.

Oh? I thought

"1) I want to improve and enrich people's movie watching experience."

meant allowing people to make more efficient movie-watching choices, no?

Anyways, I don't believe this project will aid in this, for the reasons mentioned above, as well as several other ones.

As far as revealing general trends, the way the study is set up, the variables that are includes and the ones that aren't, the lack of existing theory/mathematical models that are used, I don't think you can find a phenomenon that

1) Is sufficiently unexpected/surprising.
2) Rigorous enough to hold up in a larger, better-implemented study.

[MustUnsee]

what does each area of the face/mood chart mean? I'm sure I could choose an accurate face on the chart to best fit how I feel.... but only if I understood it a bit better.

[td888]

Hi Quicky,

I am trying to enter a new movie, but the site is not working anymore.