Chapeter 10 of edited volume "Game Research Methods" (eds. P. Lankoski & S. Bjork): Quantitative methods and analyses for the study of players and their behaviour

I choose this chapter because I was interested to see what type of quantitative data can be collected to study player's behaviour. Well turns out there is no such thing and they end up turning non-measurable constructs into operational definitions that are measurable. most of the examples were survey based, basically asking user about it.

Part 1, a short summary of the chapter:

The chapter starts with explaining why empirical studies are good: they try to minimise human judgment errors, subjectivity and they rely on data to drive conclusions.

Then comes the qualitative vs quantitative methods. It does a good job of illustrating the strength and use of each method. Each of these methods are neither good nor bad. They have their own specific place and purpose in the research world. Qualitative methods can be used to study a larger scope, new area, vague or ill-defined phenomena or simply as a first step in getting started with a research to define what might interest you for further studies; or vice versa it can be used as second step to get additional information on a previous quantitative study.  Quantitative methods are used when you are interested in a specific object of inquiry.

(as stated in this chapter)The main problem when using quantitative measures is to capture things that are not originally numeric. For example player's enjoyment of a game. Here "psychometrics" comes to rescue.
How psychometrics work:
1. Finding a "construct", which by definition is something unmeasurable such as engagement.
2. coming up with "operational definitions" , their aim is to measure effect of a construct in the form of operational definitions.

Chapter goes on with 3 main issues to consider with quantitative methods:
1. Reliability: How we ensure that operational definitions measure the same thing each time we use them? This is the consistency of the tool(scale, questions...) different practical methods of ensuring reliability is proposed in the chapter(ex: cronbach's alpha ..)
2. Construct validity : Do operational definitions actually represent the construct we are interested in?
Basically there is no 100% sure answer to this question. This validity can come from expert examinations, evidence form response process(here comes qualitative method)
3.Research design: Or how we test hypothesis so that they are reasonable tests of theory?
then it went to too much detail on different research designs and statistic analysis.
Since this text is getting too long ... I will stop here.

Part 2, application of the method on a game of choice:

I personally like experimental design. Therefore when possible, I try to design my study as experiment. However, it is generally impossible to design a perfectly experimental study without jeopardising other factors like natural occurrences of situations(we are forcing groups to one condition). And I am not even sure the thing I here came up with; is experiment.

Since I don't really play digital games myself, I just google searched for most popular games of 2016 and "pokemon go" showed up. In the case of pokemon go I would be interested to see if this game actually made people more active and if so how long lasting this effect is. Lets make the research question more concert: "Does pokemon go make people walk more during 2 month of test period and how the walking habits change after 2 month of not using pokemon go"
to answer this question
1)I would recruit a group of people according to my specific research interest(lets say uni student of age 25-30) and
2)would ask them to install a step counting app. record their walking habit in it as well as a diary report on how much they think they walked each day.
3)after a month of pre-pokemon go use, I will ask them to install pokemon go and play with it for 2 months , keeping the same reports as before.(not sure if I ask them to spend X amount of time with app)
4)after two month I will ask them to uninstall pokemon go and go back to their life without it, still keep the report for one more month.
Note:
Lets be real, this project is improbable at its current settings:
at this point the project took 6 month to just collect the data. My test group is better to be big enough because many people just fail to follow instructions and some might fall in love with pokemon go and refuse to stop using it. and lets be honest, who is going to follow these instructions for 6 months? maybe if I pay them , they would do! but then who is going to fund such a long and relatively boring study.
Perhaps more realistic measures will be a one week pre-pokemon, 2 month use and one week after 2 month. which is still not very easy to accomplish.